{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Machine Learning for the 20th century:
Artifact Classification From an Aviation Mystery\n", "\n", "
\n", "\n", "
















\n", "Photograph credit: ID [98648556](https://www.dreamstime.com/editorial-photo-historic-lockheed-model-e-electra-airplane-seattle-museum-flight-display-silver-metal-sheeting-makes-dazzling-image98648556) © [Mickem](https://www.dreamstime.com/mickem_info) | [Dreamstime.com](https://www.dreamstime.com)\n", "\n", "### Can a United Kingdom database from 1987 classify a glass cosmetics jar that might have belonged to Amelia Earhart?

\n", "Author: Joe Cerniglia


\n", "Reviewers:
Jennifer Ding
Research Application Manager
\n", "The Alan Turing Institute

Eirini Zormpa
Community Manager of Open Collaboration
The Alan Turing Institute

\n", "Date: October 6, 2022

\n", "\n", "## Background\n", "In the summer of 2010, researchers from The International Group for Historic Aircraft Recovery excavated, in fragments, a small semi-opaque cosmetic jar from the Pacific island of Nikumaroro. They were searching for evidence to support their hypothesis that Amelia Earhart and her navigator Fred Noonan perished there in 1937, after very nearly succeeding in their pathbreaking mission to circumnavigate the globe by air at its widest point, the equator.\n", "\n", "My interest in glass artifacts from the island ultimately led to a personal quest to become a resident glass expert of the team. In partnership with archaeologists Thomas King and Bill Lockhart, and chemist Greg George, we produced a [paper](https://www.academia.edu/40823470/A_Freckle_In_Time_or_a_Fly_in_the_Ointment) that summarized our findings. By then, our work on the jar had already generated some [press](https://www.nbcnews.com/id/wbna47623025).\n", "\n", "As we continued our research in 2013, Greg George and I worked with EAG Laboratories to determine the chemical composition of the Nikumaroro jar and a sibling glass jar, used as a control, that I had purchased on eBay and had named the clear facsimile. While the two jars were similar in most respects, interestingly, the clear facsimile jar was transparent, while the Nikumaroro jar was semi-opaque. Both jars were manufactured by the Hazel-Atlas Glass Company of Wheeling, West Virginia, but EAG Laboratories' [ICP-MS analysis](https://tighar.org/Projects/Earhart/Archives/Research/ResearchPapers/freckleintime/Document_02_Facsimileclearjarreport.pdf) revealed them to have considerably different chemical profiles.\n", "\n", "\n", "## Data \n", "In the social sciences, \"data\" often mean a collection of facts. In industry, however, we often consider data to be a collection of numbers and vectors, and we use computational tools such as statistics and machine learning to aid in their analysis. As part of an eCornell machine learning course, in 2022 I had the opportunity to revisit my work on the Nikumaroro jar as a learning exercise, and to use the data returned by EAG Laboratories to attempt to solve a novel machine learning classification problem I had proposed.\n", "\n", "Thus was my interest in analyzing the Nikumaroro jar and comparing it with its sibling rekindled. Instead of using the jars' data as a collection of facts, I now wanted to use them to study machine learning classification. \n", "\n", "By analyzing the weight percent of the chemical elements of which each jar's glass was made, plus refractive index, machine learning can build a model that will predict a pre-defined class to which it believes a given glass sample belongs. We may then assess the skill of the model by comparing all predictions to the actual correct classes for each sample. The correct class of the Nikumaroro jar and of the clear facsimile is that of a **container.**\n", "\n", "Two sets of measurements for the purposes of machine learning are not of much value, however, because machine learning models require a full range of known examples upon which to 'train' a model to recognize examples it has never seen before. \n", "\n", "EAG Laboratories had been thorough, and for the purposes of our 2013 analysis, their data was ample enough. For the purposes of machine learning, however, the data, while eagerly anticipated by our team, was paltry. Where would I find the data upon which to train a machine learning model? I did not need to search very far for my answer. British forensic scientists ``Ian W. Evett`` and ``Ernest J. Spiehler`` assembled a database of glass samples in 1987 for the purposes of solving crimes, such as break-ins. Their paper was first presented at the 1987 conference of the KBS (Knowledge-Based Systems) in Goverment and is titled:

\"Rule Induction in Forensic Science.\" http://gpbib.cs.ucl.ac.uk/gp-html/evett_1987_rifs.html.

Their database is available at the University of California Machine Learning Repository [here](http://archive.ics.uci.edu/ml/datasets/Glass+Identification).\n", "\n", "The weight percent of eight chemical elements and refractive index comprise the features of their database and each sample is represented by a target variable called Type. The types of glass represented in the sample are: Window Float, Window Non-Float, Vehicle Float, Container, Tableware, and Headlamp. There are no Vehicle Non-Float types represented in the data.\n", "\n", "My two glass samples, as luck would have it, were independently measured for most of the same elements as found in the U.K. database, with a few caveats:\n", "\n", "* Because silicon was measured in my lab data as 'Matrix,' with no actual number stated, I estimated the silicon for both containers based on the average for containers in the database (72.37). \n", "* Because K (potassium) was measured at 980 ppm for the clear facsimile, rather than by wt%, I assigned it a value of 50% of the wt% of the Nikumaroro jar: .5 X .24 = .12. \n", "* Because Fe (iron) is at very low wt% levels in the 1987 database, and the levels in my data are listed at very low parts per million (ppm), I assigned to the Nikumaroro jar and to the clear facsimile a wt% of Fe of .02 and .01, respectively.\n", "* Refractive index was not measured for either the Nikumaroro jar or the clear facsimile. Since the clear facsimile is completely transparent, I assigned it the minimum refractive index of containers from the 1987 dataset. Since the Nikumaroro jar is semi-opaque, I assigned it the maximum refractive index of containers from the 1987 dataset. \n", "\n", "These educated guessed were necessary to ensure the ability of the machine learning algorithm to compare the data from EAG Laboratories with the training dataset from the U.K.*\n", "\n", "#### *Note: In a later section of this story, we will use a technique to evaluate whether or not the decision to supply a few of the values for the jars makes a significant difference in our machine learning algorithm's ability to predict the class of the two jar samples.\n", "\n", "\n", "## Research Agenda\n", "There are three main research questions I wish to answer in my machine learning classification problem. I follow each of these questions with a rationale that explains the derived benefits in answering each question.

\n", "**1) What can Python's Matplotlib and Seaborn graphing capabilities reveal about the similarities and/or dissimilarities between the Nikumaroro jar, clear facsimile, and the glass samples in the 1987 database?**\n", "\n", "Rationale: Before tackling the machine learning problem, we need to learn whether and how the Nikumaroro jar and the clear facsimile, which precede the samples in the U.K. database by many decades, differ from the samples in that U.K. database. If the difference is too great, it may not be possible to use machine learning to classify the Nikumaroro jar and the clear facsimile at all.\n", "\n", "
**2) What do the correlations between elements for the different types of glass in the 1987 database reveal about late 20th century glassmaking, as compared with early 20th century glassmaking?**\n", "\n", "Rationale: The Nikumaroro jar and the clear facsimile precede the samples in the U.K. database by many decades. During these many decades, the recipes used in the glassmaking batches may have changed considerably. It is always fascinating to observe historical changes made manifest in data, but beyond this historical interest, the graphs that may be built from these correlations may also provide visual explanations for some difficulties a machine learning model may be having with classifying the Nikumaroro jar and the clear facsimile.\n", "\n", "
**3) Using machine learning to train a model on the 1987 database, could that model be used to classify the type (container) of one or both of the older samples unseen by the model?**\n", "\n", "Rationale: A successful classification of the Nikumaroro jar may suggest that it is not quite so unusual as we had supposed in our earlier research. The U.K. database was never built to classify highly unusual or historical glass samples. It was built to classify ordinary glass samples retrieved from break-ins that occurred in the 1980s. Successful classification would strengthen the argument, often expressed but without evidence, that the jar is not a rare item likely to be from the 1930s, but of far more recent provenance, too recent to have possibly been brought to the island in 1937 by Amelia Earhart. \n", "\n", "The outcome of this classification problem has the potential to disverify or perhaps weaken a part of our [hypothesis](https://www.academia.edu/9015080/Amelia_Earhart_on_Nikumaroro_A_Summary_of_the_Evidence) that Earhart and her navigator, Fred Noonan, perished on Nikumaroro. Efforts to disverify are an important element of scientific research, for as [Richard Feynman once said](https://faculty.mtsac.edu/cbriggs/Biol-1%20Readings%20Cargo%20Cult%20Science.pdf), \"The first principle is that you must not fool yourself — and you are the easiest person to fool.\" \n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### First, before we start to analyze these questions, we will read in the data file and create a simple report.\n", "\n", "Our first task is to read in a copy of the 1987 glass.data file that has been uploaded to Zenodo.org. Then, we can assign the names of the variables in this file to a variable called names. Notice I am carefully avoiding the variable name 'Type' because it is a reserved word in Python. (This was obviously not true in 1987 when the database was built, because Python did not exist then.) Using that reserved word will co-opt the type function, thus disabling my ability to check the types of specific variables. Instead, I will use the variable name 'GlassType' to contain the various types of glass.\n", "\n", "I also include a docstring explaining what these variables mean and their general significance to glassmaking." ] }, { "cell_type": "code", "execution_count": 78, "metadata": {}, "outputs": [], "source": [ "\"\"\"\n", "ID: an integer that identifies the 1987 sample number. This column should always be droppped in any analysis.\n", "Ref_ix: Refractive index. There are many scientific definitions of refractive index, but in basic terms,\n", "refractive index refers to how much the path of light is refracted when it enters the glass. The higher the \n", "refractive index, the less transparent the glass is.\n", "Na: Sodium oxide reduces the temperature at which silica melts to around 1500-1650 Celsius. A lower temperature\n", "equates to a more efficient and less costly manufacturing process.\n", "Mg: Magnesium oxide is used as a reducing agent in glassmaking. It reduces sulfates, which are used to fine the\n", "glass, and other impurities, such as iron.\n", "Al: Aluminum oxide is added to increase durability and heat resistance and to reduce crystallization.\n", "Si: Silicon oxide. The main constituent (matrix) of glass is silica sand.\n", "K: Potassium oxide is used for strengthening and fining the glass.\n", "Ca: Calcium oxide is considered a stabilizer. It is used for increasing the chemical durability of glass, and \n", "also to lower the melting point of the silica. Calcium oxide can also be used to raise refractive index.\n", "Ba: Barium oxide. Barium was, and is, employed in glass manufacture to add luster or brilliance to glass, \n", "and also to raise the refractive index.[1] Museum glass, salt shakers, shot glasses, and microscopic slides \n", "are often high in barium. The Nikumaroro jar is also high in barium.\n", "Fe: Iron oxide can be used as a colorant or it can exist by chance as an impurity.\n", "GlassType: This is the integer assigned in 1987 to signify the type of glass of the sample. This column should\n", "only appear in the validation dataset; it should be dropped from the training dataset.\n", "\"\"\"\n", "import urllib.request\n", "with urllib.request.urlopen(\"https://zenodo.org/record/6913745/files/glass.data\") as f:\n", " html=f.read().decode('utf-8')\n", " with open('glass.csv', 'w') as out:\n", " out.write(html)\n", "\n", "filename='glass.csv'\n", "\n", "names=['ID','Ref_ix','Na','Mg','Al','Si','K','Ca','Ba','Fe','GlassType']\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The three questions of the research agenda (listed above) are quite different. It will be convenient to modify the dataset in different ways for each problem. Therefore, we will read in the file twice to create two Pandas dataframes of the glass data, thus allowing us to have separate data pipelines to keep the machine learning question separate from the other two. The pandas library in Python provides a convenient function, read_csv, for this purpose. One of the columns is imported as an object, which I will coax to be an integer for more reliable data processing." ] }, { "cell_type": "code", "execution_count": 79, "metadata": {}, "outputs": [], "source": [ "from pandas import read_csv\n", "dataset = read_csv(filename, header=None, sep=',',names=names)\n", "dataset_ml = read_csv(\n", " filename, header=None, sep=',',names=names)\n", "# Convert clunky object formats to ints\n", "dataset['GlassType'] = dataset['GlassType'].astype('string')\n", "dataset['GlassType'] = dataset['GlassType'].astype('int')\n", "dataset_ml['GlassType'] = dataset['GlassType'].astype('string')\n", "dataset_ml['GlassType'] = dataset['GlassType'].astype('int')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Next, we will provide the list of variables in the file with descriptions of their counts and data types. I could use the info() method for this purpose, but an anonymous user on [StackOverflow](https://stackoverflow.com/questions/64067424/how-to-convert-df-info-into-data-frame-df-info) has provided a function, which I have adopted, to give this information a more attractive display. It would be useful as well to have a simple statistical report of this dataset, which is provided by the describe method. Notice that the columns ID and GlassType, as supplied from the 1987 study, are integers, but they are also categorical variables. I will therefore drop these columns from my simple statistical report because they would be meaningless in terms of the measures of central tendency (mean, max, etc.). However, because GlassType will be very important to the rest of this code, I will take a copy of the dataset prior to making this report, to ensure I do not drop this variable from the original Pandas dataframe." ] }, { "cell_type": "code", "execution_count": 80, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
ColumnNon-Null CountDtype
0ID214int64
1Ref_ix214float64
2Na214float64
3Mg214float64
4Al214float64
5Si214float64
6K214float64
7Ca214float64
8Ba214float64
9Fe214float64
10GlassType214int64
\n", "
" ], "text/plain": [ " Column Non-Null Count Dtype\n", "0 ID 214 int64\n", "1 Ref_ix 214 float64\n", "2 Na 214 float64\n", "3 Mg 214 float64\n", "4 Al 214 float64\n", "5 Si 214 float64\n", "6 K 214 float64\n", "7 Ca 214 float64\n", "8 Ba 214 float64\n", "9 Fe 214 float64\n", "10 GlassType 214 int64" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Ref_ixNaMgAlSiKCaBaFe
count214.000000214.000000214.000000214.000000214.000000214.000000214.000000214.000000214.000000
mean1.51836513.4078502.6845331.44490772.6509350.4970568.9569630.1750470.057009
std0.0030370.8166041.4424080.4992700.7745460.6521921.4231530.4972190.097439
min1.51115010.7300000.0000000.29000069.8100000.0000005.4300000.0000000.000000
25%1.51652212.9075002.1150001.19000072.2800000.1225008.2400000.0000000.000000
50%1.51768013.3000003.4800001.36000072.7900000.5550008.6000000.0000000.000000
75%1.51915713.8250003.6000001.63000073.0875000.6100009.1725000.0000000.100000
max1.53393017.3800004.4900003.50000075.4100006.21000016.1900003.1500000.510000
\n", "
" ], "text/plain": [ " Ref_ix Na Mg Al Si K \\\n", "count 214.000000 214.000000 214.000000 214.000000 214.000000 214.000000 \n", "mean 1.518365 13.407850 2.684533 1.444907 72.650935 0.497056 \n", "std 0.003037 0.816604 1.442408 0.499270 0.774546 0.652192 \n", "min 1.511150 10.730000 0.000000 0.290000 69.810000 0.000000 \n", "25% 1.516522 12.907500 2.115000 1.190000 72.280000 0.122500 \n", "50% 1.517680 13.300000 3.480000 1.360000 72.790000 0.555000 \n", "75% 1.519157 13.825000 3.600000 1.630000 73.087500 0.610000 \n", "max 1.533930 17.380000 4.490000 3.500000 75.410000 6.210000 \n", "\n", " Ca Ba Fe \n", "count 214.000000 214.000000 214.000000 \n", "mean 8.956963 0.175047 0.057009 \n", "std 1.423153 0.497219 0.097439 \n", "min 5.430000 0.000000 0.000000 \n", "25% 8.240000 0.000000 0.000000 \n", "50% 8.600000 0.000000 0.000000 \n", "75% 9.172500 0.000000 0.100000 \n", "max 16.190000 3.150000 0.510000 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "def infoOut(data,details=False):\n", " dfInfo = data.columns.to_frame(name='Column')\n", " dfInfo['Non-Null Count'] = data.notna().sum()\n", " dfInfo['Dtype'] = data.dtypes\n", " dfInfo.reset_index(drop=True,inplace=True)\n", " if details:\n", " rangeIndex = (dfInfo['Non-Null Count'].min(),dfInfo['Non-Null Count'].min())\n", " totalColumns = dfInfo['Column'].count()\n", " dtypesCount = dfInfo['Dtype'].value_counts()\n", " totalMemory = dfInfo.memory_usage().sum()\n", " return dfInfo, rangeIndex, totalColumns, dtypesCount, totalMemory\n", " else:\n", " return dfInfo\n", "display(infoOut(dataset))\n", "temp=dataset.copy()\n", "display(temp.iloc[:,1:-1].describe())\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "There are no nulls in the dataset. The column types are all numeric. All 214 samples of the database have been included." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Create a dictionary of glass types.\n", "We need a way to translate the numerical glass types found in the 1987 database to their English equivalents. This is accomplished by creating a Python dictionary.\n", "\n", "Note we could have omitted Vehicle Non-Float from this dictionary, since there are no examples in the data of this type; however, for the sake of clarity we will retain it.\n", "\n", "We include also an exhibit that lists the counts for the various glass types." ] }, { "cell_type": "code", "execution_count": 81, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Counts
GlassType
Window Non-Float76
Window Float70
Headlamp29
Vehicle Float17
Container13
Tableware9
\n", "
" ], "text/plain": [ " Counts\n", "GlassType \n", "Window Non-Float 76\n", "Window Float 70\n", "Headlamp 29\n", "Vehicle Float 17\n", "Container 13\n", "Tableware 9" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ " dict = {1: 'Window Float',\n", " 2: 'Window Non-Float', \n", " 3: 'Vehicle Float',\n", " 4: 'Vehicle Non-Float',\n", " 5: 'Container',\n", " 6: 'Tableware',\n", " 7: 'Headlamp'}\n", "\n", "vc=temp.replace({\"GlassType\": dict},inplace=True)\n", "vc=temp['GlassType'].value_counts().rename_axis('unique_values').reset_index(name='Counts')\n", "vc.set_index('unique_values',inplace=True)\n", "vc.rename_axis(['GlassType'],inplace=True)\n", "display(vc)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### What do the types of glass represent in everyday life?\n", "The terms \"float\" and \"non-float\" are glass industry terms. They describe the process by which window glass is made. Prior to 1959, all windows were made with non-float processes. These commonly involved grinding and polishing sheets of plate glass, which were cast on an iron surface. The result was not the perfectly smooth and clear glass we know today but rather was somewhat more translucent and sometimes ['wavy'](https://brennancorp.com/blog/why-is-the-glass-in-my-windows-wavy/). Today, this kind of window glass is prized for its artistic appearance and historical significance, if not for its lack of clarity. Float glass is made by floating molten glass on a bed of molten tin under controlled environmental conditions in a bath chamber. An excellent video describing this process is [here:](https://www.youtube.com/watch?v=pMKHs_FF0bw&t=200s)\n", "\n", "If a glass sample is Window Float, in all probability it was manufactured no earlier than 1959. Window Non-Float samples are probably older than 1959. \n", "\n", "Vehicle Float glass describes the type of glass that has been used for vehicle windshields since 1959.\n", "\n", "Headlamps are vehicle lamps, which illuminate the road ahead.\n", "\n", "Container glass describes commercial product containers, whether for bottles or jars.\n", "\n", "Tableware glass describes glassware for the table or any glass table item used in dining or the serving of food." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Create some utility functions for plotting graphs from the 1987 dataset." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Before using machine learning on the data, there is much we can learn about the two jars' likeness to glass samples in the 1987 database, simply by observing counts based on specific criteria. By comparing each jar with its \"nearest neighbors,\" based on measurements for three elements in the periodic table, we may gain a sense of which type of glass in the database each jar most closely resembles. \n", "\n", "We will therefore write functions to create three graphical reports, one for each of the three elements magnesium, calcium, and barium. Each of the three reports will display three things:
\n", "1) The complete range of counts for each glass type. (This graph will be repeated for each report.)
\n", "2) The range of counts for all glass types within 0.15 above and below the elemental measurement of the clear facsimile.
\n", "3) The range of counts for all glass types within 0.15 above and below the elemental measurement of the Nikumaroro jar.\n", "\n", "The key function, numgroups, requires an element from the periodic table so that it can look up parameters in the included dictionary. This dictionary contains the measurements obtained from the lab for each element for the clear facsimile and for the Nikumaroro jar. Low and High parameter values for each element are also required, even when the complete range of counts is to be displayed. These will be passed to the function as hard-coded values that are either
\n", "> a. exactly 0.15 above and below the measurement for each jar; or
\n", "> b. the lowest and highest measured values in the database for each element.\n", "\n", "Last, the jar_spec parameter is simply the type of jar to be analyzed, clear facsimile or Nikumaroro jar. The jar_spec is used to create the title for each graph.\n", "\n", "If the graph's input parameters result in a graph with a single type of glass, no graph is created. (Bar graphs with but a single bar are, after all, neither very attractive nor useful!) Instead, a message is displayed that mirrors the text that accompanies the other graphs.\n", "\n", "If the number of glass types returned by the criteria is equal to zero, we cause an error to be thrown with int('d'), which attempts to convert the letter 'd' to an integer, rather than waiting for the error that would have been thrown naturally by the plot statement. If we had waited for the plot statement to throw the error, useless information about the size of the graph (which cannot be drawn in this case) is displayed. In the case of this error, a simple message is printed to inform us that no relevant glass samples were found in the database.\n", "\n", "The other function, ticks_restrict_to_integer, is a utility function that controls the number of tick marks on the y-axis of each graph. The y-axis represents the count. Because the scale of counts varies with each graph, it was necessary to make the y-axes more uniform with one another by having this function." ] }, { "cell_type": "code", "execution_count": 82, "metadata": {}, "outputs": [], "source": [ "from matplotlib.ticker import MultipleLocator\n", "def ticks_restrict_to_integer(axis):\n", " \"\"\"Restrict the ticks on the given axis to be at least integer;\n", " that is, no half ticks at 1.5 for example.\n", " \"\"\"\n", " major_tick_locs = axis.get_majorticklocs()\n", " if len(major_tick_locs) < 2 or major_tick_locs[1] - major_tick_locs[0] < 1:\n", " axis.set_major_locator(MultipleLocator(1)) \n", "\n", "def numgroups(element,low,high,jar_spec):\n", " \"\"\"\n", " non-fruitful function. Outputs a graph if the number of types of\n", " glass within the DataFrame is > 1. Otherwise, the function prints\n", " a message, for aesthetic reasons.\n", "\n", " parameters:\n", " element: an element in the periodic table\n", " low: A minimum value for the measured element, which acts as the \n", " minimum value to allow into the dataset sample\n", " high: A maximum value for the measured element, which acts as the \n", " maximum value to allow into the dataset sample\n", " jar_spec: A value of either 'clear facsimile' or 'Nikumaroro jar', used\n", " in the title of the graph. \n", " \n", " However, when the function is called to obtain the complete distri-\n", " bution of counts, the value of jar_spec is set equal to None. \n", " \n", " Preconditions:\n", " element is a string with value of: 'Mg' or 'Ca' or 'Ba'\n", " \n", " Low cannnot be greater than high.\n", " \"\"\"\n", " jar_spec='full range of ' + element + ' measurements in the database' if jar_spec==None else jar_spec\n", " rangedict = {'Ba':['Barium',[.37,.74]],\n", " 'Mg':['Magnesium',[2.4,4.3]],\n", " 'Ca':['Calcium',[3.6,8.5]]}\n", " facsimile_ref=rangedict[element][1][0]\n", " artifact_ref=rangedict[element][1][1]\n", " element_full_name=rangedict[element][0]\n", " sampl = dataset[(dataset[element] >= low) & (\n", " dataset[element] <= high)]\n", " number_groups=sampl['GlassType'].value_counts().shape[0]\n", " titletext1=\"Count of Glass Types in the 1987 database \\n with a range of measured values of \" + \\\n", " element_full_name + \" \\n between \" + str(low) + \" and \" + str(high) + \" wt%. \\n \" + \\\n", " \"This is the range of counts for the \" + jar_spec.upper()\n", "\n", " if 'clear' in jar_spec or 'Niku' in jar_spec:\n", " #Tolerance provides information in the report about the range of values considered in the graph.\n", " tolerance=round((high-low)/2,2)\n", " titletext2=\" that fall within a tolerance of +-\" + str(tolerance) + ' of its measurement.'\n", " else:\n", " tolerance=0\n", " titletext2=\".\"\n", " \n", " if low>high:\n", " print('Low value must be smaller than the high value. Try again.')\n", " return\n", " \n", " if number_groups==1:\n", " print(titletext1 + titletext2)\n", " print('There is only ONE sample Type.')\n", " print(sampl['GlassType'].values[0],'=',sampl.shape[0])\n", " print('reference: Clear facsimile jar =',facsimile_ref,'wt%'\n", " '\\n Nikumaroro jar =',artifact_ref,'wt%')\n", " else:\n", " print('reference: Clear facsimile jar =',facsimile_ref,'wt%'\n", " '\\n Nikumaroro jar =',artifact_ref,'wt%')\n", " try:\n", " vc=sampl['GlassType'].value_counts()\n", " int('d') if len(vc)==0 else int('1')\n", " \n", " vc.plot(\n", " kind='bar', figsize=(2.4*number_groups, 6), rot=0, cmap='Spectral');\n", " plt.xlabel(\"Glass Type\", labelpad=14,fontsize=16,rotation=0)\n", " plt.ylabel(\"Count of Type\", labelpad=70, \n", " fontsize=16,rotation=0)\n", " plt.xticks(fontsize=14)\n", " plt.yticks(fontsize=14)\n", " plt.figtext(0.5, 1.0, titletext1+titletext2, wrap=True, horizontalalignment='center', fontsize=16) \n", " ax = plt.subplot()\n", " ticks_restrict_to_integer(ax.yaxis)\n", " plt.show()\n", " except Exception as e: \n", " #print(e)\n", " print('For the ' + jar_spec + ', no glass samples in the U.K. data are in the range of ' + str(low) +\n", " \" and \" + str(high) + ' wt% for ' + element_full_name + '.')\n", " print()\n", " return" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### What types of glass in the 1987 database are most similar to the Nikumaroro jar and clear facsimile in terms of magnesium, barium or calcium content?\n", "To review what was stated above, examining Mg, Ba, and Ca individually in the U.K. database, we can perform the following steps to produce a report: \n", "1. Obtain the complete distribution of counts in the 1987 database by restricting the element's values to the range between its minimum and maximum wt% values. \n", "2. See the distribution of counts in the 1987 database that results from setting the range of measured wt% for the element to a tolerance 0.15 wt% above and below its measurement for the ```clear facsimile```. \n", "3. See the distribution of counts in the 1987 database that results from setting the range of measured wt% for the element to a tolerance 0.15 wt% above and below its measurement for the ```Nikumaroro jar```.\n", "4. Repeat steps 1 to 3 for the next element until all elements have been reported." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Report for Magnesium" ] }, { "cell_type": "code", "execution_count": 83, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 2.4 wt%\n", " Nikumaroro jar = 4.3 wt%\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 2.4 wt%\n", " Nikumaroro jar = 4.3 wt%\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 2.4 wt%\n", " Nikumaroro jar = 4.3 wt%\n", "For the Nikumaroro jar, no glass samples in the U.K. data are in the range of 4.15 and 4.45 wt% for Magnesium.\n", "\n" ] } ], "source": [ "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "import warnings\n", "import random\n", "\n", "dataset.replace({\"GlassType\": dict},inplace=True)\n", "element='Mg'\n", "Mgdict={'clear facsimile':[2.25, 2.55],'Nikumaroro jar':[4.15,4.45]}\n", "low=0\n", "high=dataset[element].max()\n", "\n", "#Run report for all glass types\n", "numgroups(element,low,high,None)\n", "\n", "#Run reports for the clear facsimile and Nikumaroro jar\n", "for jar_type in Mgdict:\n", " low=Mgdict[jar_type][0]\n", " high=Mgdict[jar_type][1]\n", " numgroups(element,low,high,jar_type)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Magnesium Analysis\n", "Magnesium:
\n", "The clear facsimile jar has 2.4 wt% Mg. There are only two glass types, tableware and non-float windows, with values between 2.25 and 2.55 wt% Mg. Magnesium, if it were the only feature, would seem to predict the clear facsimile jar to be in the window or the tableware family, a close cousin to the container family.
The Nikumaroro jar has 4.3 wt% Mg. Checking back to the report we created with the describe() method, this value is well above the 75th percentile. We know from the literature on glassmaking that any Mg measurement from a modern glass sample that is above 3.5 wt% is likely to be a window, and not a container.[2]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Report for Calcium" ] }, { "cell_type": "code", "execution_count": 84, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 3.6 wt%\n", " Nikumaroro jar = 8.5 wt%\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 3.6 wt%\n", " Nikumaroro jar = 8.5 wt%\n", "For the clear facsimile, no glass samples in the U.K. data are in the range of 3.45 and 3.75 wt% for Calcium.\n", "\n", "reference: Clear facsimile jar = 3.6 wt%\n", " Nikumaroro jar = 8.5 wt%\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "dataset.replace({\"GlassType\": dict},inplace=True)\n", "element='Ca'\n", "Cadict={'clear facsimile':[3.45, 3.75],'Nikumaroro jar':[8.35,8.65]}\n", "low=0\n", "high=dataset[element].max()\n", "\n", "#Run report for all glass types\n", "numgroups(element,low,high,None)\n", "\n", "#Run reports for the clear facsimile and Nikumaroro jar\n", "for jar_type in Cadict:\n", " low=Cadict[jar_type][0]\n", " high=Cadict[jar_type][1]\n", " numgroups(element,low,high,jar_type)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Calcium Analysis" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Calcium:
\n", "The clear facsimile jar has 3.6 wt% Ca. We can see from the report we created with the describe() method that this is off-scale low, a value less than all the samples in the database. No glass samples are in the range of 3.45% and 3.75% weight calcium.
The Nikumaroro jar has 8.5 wt% Ca. This is close to the mean in the database (8.96) for all glass types, but no containers are displayed on the graph within +- .15 of the calcium value measured on the Nikumaroro jar." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Report for Barium" ] }, { "cell_type": "code", "execution_count": 85, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 0.37 wt%\n", " Nikumaroro jar = 0.74 wt%\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 0.37 wt%\n", " Nikumaroro jar = 0.74 wt%\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAyEAAAH4CAYAAABdQOI9AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAACSN0lEQVR4nOzdd1xP++MH8FelRUYZ17q2PmiXSqUpQlxZ18ye3ewtbna2aCBCCBnZ49rjopTt2rKyCUml0vn90e9zvn3aheNyX8/Hw+Ohsz7vz/tzPudzXuf9fp+jJAiCACIiIiIiIokof+8CEBERERHRfwtDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEJGkeEM+yg/3ka+L9UlE/0YMIUQ/scOHD6Nv376wsrKCiYkJ3NzcEBoaitTU1O9Snjt37qBnz55fvJ34+Hj0798fhoaGsLGxQUpKSq7L3r9/H1OnTkXTpk1haGgICwsL9OnTB8eOHVNYLjw8HDKZDHFxcV9cvqKIjY2FTCbL99+/VWRkJGQyGa5evVrkbaSkpGDGjBk4cuSIOM3JyQnTpk37GkXMVUJCAhwdHXHgwIFs89auXSvuO25ubjh48GC2ZbZu3QpXV1eYmJjgt99+w759+8R58nrJ7d+TJ08KXM6i7KNF+c59jc+SiCg/xb53AYjo25g6dSo2bdoENzc3dOnSBcWLF8e5c+cwZ84cREREwNfXFyoqKpKW6cCBA1/lxGbnzp04efIkZs+ejerVq0NNTS3H5Y4dO4aRI0eievXq6NevH2rWrIn4+Hjs27cPgwYNwoQJE9CrV68vLs/XUKFCBYSFhYl/R0VFYf78+fD390f58uW/Y8kKRk9PD2FhYahdu3aRt/Hy5UusW7cODRs2/Ioly1tCQgI8PDzw9OnTbPOWLVuGRYsWoWvXrnB2dsbly5cxcuRI+Pj4oHXr1gCAvXv3wsvLC7169YK9vT2OHTuGESNGQFNTE46OjmK9ZPbp0ycMHToUenp6qFSp0jd9f1/rO0dE9LUxhBD9hHbs2IENGzZg2rRp6NSpkzjd2toaurq6GDFiBHbv3g03N7fvV8gv8P79e2hqaqJt27a5LvPixQuMHTsWhoaGWLFihUJQcXZ2xi+//IIFCxagVatWKFeunBTFzpOamhqMjY3Fv1+/fg0AqF+/PqpWrfqdSlVwWlpaCuX/EZw7dw7e3t548+ZNtnmfP3/GypUr4erqCm9vbwCAjY0NkpKSMGfOHLRs2RIqKirYvn07GjZsiAkTJgDI+I5dvXoVmzZtgqOjY471MnPmTCgpKWH+/PlQVmaHBCL6b+LRj+gnFBwcDJlMphBA5Fq2bIk+ffpAW1tbnBYbG4thw4bB2toaJiYmGDx4MB48eCDO9/Pzg4mJicJ2bty4AZlMhsjISADA+PHjMXToUISEhMDR0RGGhoZwd3fHvXv3xG34+/sjMTERMpkM4eHhuZb/0KFDaN++PYyNjWFvbw9fX1+xC5m7uzv8/PyQlJQEmUwGPz+/HLexYcMGfPjwAVOmTMmxpWTQoEGwsbHB27dvc1xfEASEhISgdevWMDAwgImJCXr37o1bt26Jy7x69QrDhg2DpaUljIyM0LVrV5w7d67A8wvr7du30NfXR3BwsML0CxcuQCaT4datW/Dz80OrVq2wc+dOODo6wsTEBAMHDkRsbKzCOteuXUPPnj1hZGSERo0aYfr06UhKSipy2bN24XF3d4ePjw8WLVoEGxsbGBkZwcPDAy9evMhx/djYWDRp0gQAMGzYMLi7u4vzkpOTMWXKFFhYWMDMzAzjxo1DQkKCwvpr165Fs2bNoK+vD1dXV4UuUbn5448/oKuri5UrV2ab9+bNG3z48AGNGzdWmG5mZoZXr17h5s2bADK6kGlpaSksU6ZMGbx//z7H17x79y5CQ0MxfPhw6Ojo5Fm+HTt2wMXFBYaGhujfvz/evXunMD+/fTS371xMTAyGDh2KRo0aQV9fH05OTggICMg2duSff/5Bu3btYGBggHbt2uH06dMK869cuYL+/fujYcOG0NfXh4uLCzZt2qSwzMqVK9G0aVMYGBjA2dkZAQEBSE9PF+e/efMGY8eOhYWFBUxMTDBo0CA8fvw4z3ohop8DQwjRT+bly5e4ffs27O3tc11m3Lhx4vznz5+jY8eOePjwIby9veHj44PY2Fh07do11xPG3Jw5cwY7duyAl5cX5s2bh4cPH2L8+PEAgI4dO6JDhw7Q0NBAWFgYHBwcctxGWFgYPD09YWBgAH9/f3Tv3h2rVq0SrzR7e3srbKdjx445bufYsWOoV68eatasmeP80qVLY9myZahbt26O81etWoX58+ejQ4cOCA4OxuTJk3H37l2xHAAwYcIEPHr0CD4+PggMDISmpiYGDhwonizmN7+wtLW1YW9vj7179ypM3717N+rXry+OGXny5Anmzp0LT09P+Pj4ICYmBr169RLHzty9exfdu3eHkpISfH19MXr0aOzbtw/Dhw8v8HsriG3btuHy5cuYNWsWpkyZgsjISPj4+OS4bIUKFeDv7w8AGDlypNj6AADbt2/H+/fv4evriyFDhmD37t3isgDg7+8vtk4sW7YM1tbWGDlyJPbv359n+UJDQ7F48eIcw0DZsmWhpqaWrZuWPMzJx3J07doVp0+fxv79+/Hhwwfs27cPJ0+ehKura46vuWjRItSoUQO///57nmXbv38/xo0bBxsbGwQEBODXX3/FwoULFZbJbx/N6Tv38eNH9OjRA+/evcOcOXOwfPlyWFpaYsmSJdnGSc2aNQvOzs7w9/dHuXLlMHDgQPGiwtOnT9GjRw8UL14cixcvRkBAAGrWrAlvb28xoO3btw+LFy9Gr169EBwcjI4dO8LPzw+bN28GkBEue/TogfPnz2PSpEmYO3cuXr9+je7du+ca4ojo58HuWEQ/mefPnwMAKleuXKDl16xZg+TkZKxatUo8GbOwsICzszNWr14thoiC+PjxI5YvX44KFSoAyOgSNXPmTLx9+xYVK1ZExYoVoaysnGu3nfT0dPj6+sLV1RVTpkwBADRu3BglS5aEt7c3+vXrh3r16uW7HSDjJNHGxibb9LS0NIW/lZWVc+wS8+zZM3h4eIiDei0sLBAfHw8fHx98/PgRJUqUQHR0NAYPHgwnJycAQN26dbF69WokJSWhTJky+c4virZt2+KPP/5ATEwMatWqhbS0NOzfvx8DBw4Ul0lMTMTixYthZ2cHAKhVqxZ+++037N27F23btkVgYCDKli2LoKAgsZWoRo0a6NatG6KiomBubv5Vyq6iooLly5dDXV0dAHDz5k3xBDQrNTU11K9fHwBQvXp11KlTR5xXs2ZNLFy4EEpKSrC2tkZERITYAhcfH4+goCD069dPDFGNGzfGx48fsWDBArRo0SLX8unq6uZZdldXV6xevRr16tWDpaUlrl27JrZCJSYmAsjo2ufm5qYQ4Dp27KjQkiP3+PFjHD16FNOmTcu3G1ZQUBBsbW3x559/AgBsbW3x9OlThaCQ3z6a03fu2rVrqFatGnx9fcXvu5WVFQ4fPoyoqCjx8waAPn36wMPDQ1ymadOmCA4OxqxZs3Dnzh0YGxtj/vz5UFVVBQAYGRnB0tIS0dHRqFevHqKiolClShV07doVSkpKsLCwQLFixcTjw44dO3D//n3s3r1bHEtkZWUFR0dHrFu3Dp6ennnWERH92BhCiH4y8sHmmbs85CUqKgqWlpYKV4N1dHRgZWVV6K5DlStXFk8wAKBixYoAgKSkJIXuX7m5d+8e4uLi0Lx5c4XprVq1gre3t3hyUxDp6elQUlJSmHbp0qVsXdS6desmnuhlNmnSJABAXFwcYmJiEBMTg6NHjwLI6IJTokQJmJiYYMmSJbh16xbs7e1hb2+PcePGidvIb35R2Nvbo0yZMti7dy+GDBmCU6dOIT4+XuHKe8mSJcUAAmScbP/666+Ijo5G27ZtERkZiSZNmkBZWVkMZcbGxtDS0sLZs2dhbm7+Vcouk8nEAAJk7A+Zu3wVlJGRkcJnWbVqVdy5cwdAxmf66dMnODg4KARMOzs7bNu2DY8fP8avv/5a6NcEgIkTJyIxMRF//PGHWH5PT09MmjQJmpqaADK6IR46dAijRo2CsbExrly5goCAAJQsWTJbfW3ZsgWlSpVCmzZt8nzdpKQk3LhxAxMnTlSY7uLiohBCCrKPZqWvr48NGzYgNTUVd+/exYMHD3D9+nWkpaVlu8uci4uL+H81NTU0btwYV65cAQBxn/j06RNu3ryJBw8eiF3x5NsxMTHBhg0b0L59ezRv3hwODg7o27evuM3IyEhUr14d1atXFz87DQ0NmJmZISIigiGE6CfHEEL0k5HfbefZs2e5LvPy5UuUK1cOysrKiI+PF69AZ1a2bFncvXu3UK8tPzGTk1/tLWggknfBKFu2rMJ0LS0tqKurZxsHkJfKlStn60qjq6uLrVu3in8PHjw41/Xv3buHyZMn4/z589DU1ES9evXEkzp53/lFixYhICAA+/fvx969e6Gqqop27dph0qRJUFNTy3d+UaiqqqJVq1bYs2eP2DXJyspKIfzldDctHR0dsX7fvXuHsLCwbHdtAjLGghTkvRVE1v1BSUmpSM+syGs78u5hnTt3znHdV69eFTmElCpVCkuWLMG7d+/w5s0bVK9eHRcuXACQ0Z3v2bNn2L17N8aMGYN+/foByGiNKFGiBKZOnYouXbqgWrVq4vYOHz4MZ2fnfOsvPj4egiBkC+5Zb6BQkH00J8uWLcPKlSvx4cMHVKlSBSYmJihWrFi2dbJ+D3V0dPDy5UsAGQP3Z8+ejbCwMKSmpqJatWriXc3k2/ntt9/w+fNnhIaGYuHChViwYAHq1auHhQsXonbt2nj37h1iYmKgp6eXrYw1atTIs46I6MfHEEL0k9HR0UGDBg1w6tQpjB49OsdlevfujXLlyiEkJASlS5cW78SU2evXr8VuN0pKStmCxMePH7962eWvl/VuRfHx8fj06VOhujA5ODhg1apVePHiBX755RcAQPHixWFgYCAuk9vJYHp6OgYPHowyZcpg9+7dqFOnDpSVlREaGoq///5bobxeXl7w8vLCjRs3sGvXLqxevRpVq1bFgAED8p1fVG3atMH69etx9epVHDt2DFOnTlWYn9O4jTdv3ohjRrS0tNCkSRN06dIl23LyE99vVfavrWTJkgCAgIAA8XPOLLcxQQVx5MgRVKhQAQYGBuK+d+vWLSgpKUEmk4k3b8jaLdDMzAyCIODevXtiCHn69Cnu3btXoNak0qVLQ0lJKdv3IPPnWtB9NKsdO3bA19cX3t7eaNWqlVh/VlZW2ZaNj49XCLeZjwlLly7F5s2bMWfOHNjb26N48eJISkpSCPlARvfBtm3b4s2bNzh69CgCAgLg6emJ/fv3o2TJkqhXrx5mzJiR7bWLGtKJ6MfBgelEP6GePXvi5s2b2LJlS7Z5O3fuxN27d8XnHJiZmSEyMlLhAWhxcXE4e/YsTE1NAWSctCYnJyM+Pl5c5vz584UuV3794GvWrAltbe1sD42T3+lIXp6CcHd3h5aWFiZOnIjk5ORs81+8eJFry0pcXBwePnyI33//Hbq6umK5T506pbCMg4MDDh06BCDjVrrjxo1D5cqV8ezZs3znfwlDQ0PUrl0bs2fPBpAxLiFr+eXdZoCME+fHjx+jUaNGADI+85iYGOjr68PAwAAGBgaoVKkSFixYgDt37nzTsuelKM+tMTIygqqqKt68eSO+FwMDA9y5cwcBAQFfVJ7Q0FCFbaSkpGDLli0wMTGBtrY2qlWrBmVlZbF1RO7y5csAoHBrZfnnYWhomO/ramhowNjYGIcPH1aYfuLECfH/BdlHgezfuYsXL6JixYro0qWLGED++ecfxMXFZWsJybyt5ORknDx5EhYWFgAyusHp6+ujRYsWKF68uMLy8u14eXlh6NChADJaVeQD5eX7kKmpKWJjY1GlShXxc9PX18eaNWtw/PjxfOuJiH5sbAkh+gm1adMGx48fx59//okrV66gSZMmUFJSwt9//42NGzeiRYsWaN++PQCgV69e2L59uzgIVRAELF26FGpqauKAV1tbW/j4+MDLywvdunXDzZs3sWHDhkKXq1SpUkhKSsLhw4dhaGiocJUVyDgJ9fT0xPTp01G6dGk0adJEvO1s8+bN8xxInFXFihWxaNEijBgxQnxgo66uLpKTk3HmzBls3boVampqsLW1zbZuuXLlULlyZYSEhIjd1nbs2CGeGCUlJaFKlSqoXr06ZsyYgY8fP6JSpUo4fvw4njx5gqZNm0JHRyfP+V/Kzc0NCxYsgJubm3gSKKekpIThw4eLLWGLFi1C/fr10axZMwCAh4cHOnfujGHDhqF9+/ZISUlBYGAgnj17hgYNGnzzsudGflJ85swZ1KhRo0Djf3R0dODu7o7Zs2fj/fv3MDQ0xM2bN7Fo0SI0adIk2+1zC6NLly7w9PTEsmXLYGRkhJCQEMTExGDNmjUAMk6su3TpIt6py8jICNevX4efnx9cXFwU7rx2584daGtrF2hsFAAMGTIE/fr1w4QJE9CyZUtEREQohJKC7KNA9u+cgYEBNm3aBH9/f1hYWODevXsICAiAkpJStrAuv6lAlSpVsGrVKiQlJaF///4AAAMDA6xYsQLr16+Hrq4url69mm075ubmGDduHBYuXAhra2s8f/4cGzduFPehDh06YN26dejTp4/YchgWFoaDBw/it99+K+SnRUQ/HIGIfkqfP38WNm7cKHTs2FGwsLAQjI2NhbZt2wobN24UUlNTFZa9ffu20L9/f8HY2FgwMzMTBg8eLMTExCgss3nzZsHR0VHQ19cXunXrJly6dEnQ1dUVIiIiBEEQhHHjxgmurq4K6xw6dEjQ1dUVHj9+LAiCILx580Zo166doKenJwQFBeVa9i1btggtWrQQ9PT0BEdHR8HX11dISUkR5y9ZskQwNjYuUD08ffpUmDNnjtC8eXPByMhIMDExETp27CgsX75cePfunbjctm3bBF1dXeHNmzeCIAjC1atXhU6dOglGRkaCjY2NMGjQIOHMmTOCrq6usGfPHkEQBOH169fC2LFjBRsbG0FPT09o1aqVsHv3bnGb+c3PS9a6y+rGjRuCrq6ucPr0aYXp8rrZtm2bYGNjI5iZmQmjR48W4uLiFJaLjo4WunfvLhgaGgrm5ubCwIEDhdu3bxe57BEREYKurq5w5coVQRAEoXv37sKAAQMUllm9erWgq6ub5/ueN2+eYGRkJLRq1UoQBEFwdHQUpk6dqrDMjBkzBEdHR/Hvz58/C0FBQYKzs7O4zyxYsED49OlTnq8l9/jxY0FXV1fYv39/tnkbN24UmjZtKhgbGwudOnUSzp49qzA/LS1NCAwMFL8bLi4uQmBgYLbX9vb2Fpo2bVqg8sgdOnRIaNWqlaCvry906dJFWL9+faH30azfuc+fPwtz584VbGxsBCMjI6Fly5bCqlWrBC8vL6F58+aCIPzvszxw4IDg6uoq6OvrC506dRKuXr0qli0xMVHw8vISLC0tBWNjY8HNzU0IDw8X+vTpI/Tp00dcbt26dULz5s0FAwMDwdraWpg6daqQkJAgzn/27JkwYsQIwdzcXDAyMhI6dOggHD16tFD1REQ/JiVBKMIoQSIi+q5WrFiB0NBQHD16VKHLjZ+fH1atWoWLFy9+x9IRERHljd2xiIh+IH/99ReuXLmC0NBQDBkyJN9xNkRERP9G/PUiIvqBPHz4EOvXr4eTkxN69OjxvYtDRERUJOyORUREREREkmJLCBERERERSYohhIjyxQZTIiIi+poYQohIgUwmQ3BwMICMJyaPGjUK//zzT47zqeh8fHzQsGFDmJqaFunBj/8lsbGxkMlk2R5i+bX9SPv26dOn0bRpUxgYGGD69OnZ5svrLPM/AwMDuLi4ICgo6KtcWIiMjIRMJsPVq1e/eFtE9N/Du2MRkYKwsDBUrlwZAHDjxg3s2bMHvXr1+r6F+sncunULa9asQc+ePdG0aVPUr1//exeJfjALFiyAhoYGVqxYgUqVKuW63MiRI2FpaQkg4wGGly5dwuLFiwEAAwYM+KIy6OnpISwsDLVr1/6i7RDRfxNDCBEpMDY2/t5F+Om9f/8eANCqVSsYGhp+59LQj+jdu3ewt7dHo0aN8lyuevXqCt9pKysrPHr0CJs2bfriEKKlpcXjBREVGbtjEf3EDh8+DJlMhtjYWHHazJkzIZPJ8PjxY3HatGnT0KFDBwD/65ISGRkp3gK2Q4cOGD9+vLj8u3fvMHLkSJiYmMDS0hKzZs1CampqnmXZtWsX2rdvDyMjIxgZGaFz586IiorKdXl5d5KQkBA4OTnBxsYGFy5cgCAICAkJQevWrWFgYAATExP07t0bt27dEtd1d3eHj48PFi1aBBsbGxgZGcHDwwMvXrwQl0lPT4e/vz/s7OxgZGSEIUOGYM2aNZDJZArl2LNnj/hazs7OWLduXZ7vEwBu3ryJfv36wcLCAhYWFhgzZgxev34NIONhgu7u7gCAjh07iv/Pavz48Rg6dCiCg4NhZ2cHY2NjDB06FAkJCfD394e1tTUsLS0xY8YMpKeni+u9efMGY8eOhYWFBUxMTDBo0CCFzxoATp06he7du8PExAQGBgZo06YNDh48KM7//Pkz5s6dCwcHB+jr66Nly5bYuHGjON/Pzw8mJiYK27xx4wZkMhkiIyPF8nt4eGDUqFEwNTXFiBEjAACJiYmYPn06rK2tYWhoCHd3d1y/fl1hW5cvX0aXLl1gZGSE1q1bZ5ufU125uLhkm96uXTuMHTsWAJCQkIAZM2bA0dER+vr6aNSoEcaNG4f4+PgctxkeHg6ZTIa4uDhxWnx8PGQyGcLDw8VpDx8+hIeHB0xMTNCwYUOMGTNGYZ3ExER4eXmhcePGMDQ0RNu2bRXqOicfP37EnDlz4OTkBENDQ3To0AF///03gP99L548eYINGzZk+34XRMmSJbNNy+/7mdPnmbU7lru7OwYOHKiw3azfKScnJwQFBWHSpEkwMzODpaUllixZgg8fPmD06NEwMTGBo6OjQh0T0c+JIYToJ9aoUSOoqqoiIiJCnHbu3DkAUBiHcPr0adjZ2Smsq6enhz///BNAxvgFDw8Pcd7KlSuhra2NwMBAdO7cGSEhIdi0aVOu5Thw4ADGjh0LBwcHBAUFwcfHB/Hx8RgxYgRSUlLyfA+LFy/G6NGjMWbMGOjr62PVqlWYP38+OnTogODgYEyePBl3797FhAkTFNbbtm0bLl++jFmzZmHKlCmIjIyEj4+PON/X1xfLli1D165dsWTJEgAZXVwy2759O0aNGgVzc3MsXboUbm5u8PHxwcqVK3Mt740bN9CpUyekpqZi9uzZmDhxIqKjo9G9e3ckJiaiY8eOCvXq7e2d67b+/vtvHDp0CNOnT8eYMWNw6NAhtG/fHpcvX8bs2bPRtm1brFu3Dvv27QMAJCcno0ePHjh//jwmTZqEuXPn4vXr1+jevbvY+nLlyhUMGDAAdevWRWBgIBYtWgRNTU2MGjVKPHkODg7Gtm3bMHz4cAQHB8PW1hZTpkzBqVOn8vyssjpx4gQ+ffqEgIAAdOrUCYIgYPDgwdi7dy+GDx+OxYsXQ01NDe7u7nj06BGAjJPsXr16QV1dHUuWLEH79u2zfbZZtWrVCg8ePMDNmzfFaY8fP8Y///wDV1dXAMCoUaNw9OhRjBo1CsHBwejTpw/27NmDwMDAQr2nzF6/fo2uXbvi6dOnmDt3LqZOnYpLly6hb9++4n49Z84cREREwMvLC8uXL0ft2rUxbNgw3Lt3L8dtpqeno1+/fggPD8eAAQPg5+eHypUrY8CAATh16hQqVKiAsLAwlC9fHi4uLggLC0OFChVyLWN6ejrS0tKQlpaGxMREnD59Gjt37kSXLl3EZQr6/cz6eRbVsmXL8PnzZ/j7+6NFixYICAhAhw4dUL58eSxevBi1atXCn3/+iadPnxb5NYjo34/dsYh+YlpaWjAxMUFkZCQ6dOiA9+/f4/bt22jQoAGio6Ph5uaGJ0+e4MGDB7C3t8+2bp06dQAAdevWRbVq1cR51tbWmDx5MoCM7h1Hjx5FZGRkrlf1Hz16hG7dumHIkCHiNFVVVXh6euLBgwfQ1dXN9T24ubmhZcuW4t/Pnj2Dh4cHevbsCQCwsLBAfHw8fHx88PHjR5QoUQIAoKKiguXLl0NdXR1ARuvE5s2bAWRcFV+9ejUGDhyIQYMGAQDs7OzQpk0bsUUlPT0dCxcuROvWrcXQ0LhxYygpKSEwMBBdu3ZF8eLFs5U3MDAQOjo6WLFiBdTU1AAA+vr6aN26NbZt2wZ3d3eFepX/PyeJiYlYsmSJeJK5a9cu3L17F9u2bYOWlhbs7Oywf/9+XL58Ga1atcKOHTtw//597N69W+ynb2VlBUdHR6xbtw6enp64c+cOmjZtqhB+KleujLZt2+Ly5ctwdHREdHQ09PX14ebmBgCwtLSEhoYGNDU1cy1rTtLS0jBt2jTo6OgAyGiBiYiIwOrVq2FtbQ0AsLW1haurK5YuXQofHx+sW7cOampqWLp0KTQ1NWFvbw9BEDB79uxcX8fKygrlypXDgQMHUK9ePQDA/v37oa2tDRsbG3z69AmpqamYMmWKGLYtLS1x8eJFMZQXRUhICD59+oRVq1aJ79HQ0BAuLi7Yt28f3NzcEB0dDRsbG7Ro0QIAYGZmhnLlyiEtLS3HbR4/fhwXLlzAypUrYWtrCwCwt7dHp06dsGjRIoSHh8PY2BhqamooV65cvt2h5C1QmRkbG6Nbt27i3wX9fmb9POWtXoX1yy+/YNasWVBSUoKJiQnCwsLwyy+/YNy4cQCAGjVqoGnTprh+/bo4Po2Ifj4MIUQ/OVtbW6xfvx4AEBUVhQoVKsDV1RXbtm0DkHG1XVtbGwYGBgXeZtauOFWqVMm1WwvwvwGw8fHxiImJwf3793H06FEAyLclJOug10mTJgEA4uLiEBMTg5iYGIVtyUOITCYTAwgAVKxYEUlJSQAyuvukpKTA2dlZnK+kpIRmzZqJIeT+/ft4+fIlHBwcFE4Y7ezssGTJEly5ciXH/vhRUVFo1aqVGEAAoE6dOpDJZIiKiso1qOWkUqVKCle5y5Yti8+fP0NLS0ucVqZMGXz48AFAxklh9erVUb16dbHMGhoaMDMzQ0REBDw9PdG+fXu0b98eiYmJuHfvHh48eCC2lMk/CxMTE/j6+sLd3R3Ozs5wcnLK8WQ2Pzo6OuIJq7x8mpqaMDc3V6jTxo0bi5/hhQsXYG5urhB4mjVrlmcIUVFRQYsWLXDgwAEMHz4cQEYIad68OYoVK4ZixYph1apVADJaWh48eIA7d+7g3r17CvtIYUVGRsLY2BilSpUS30+lSpVQu3ZtnD17Fm5ubjAxMcHmzZvx8uVLODo6wsHBQaFrY1ZRUVEoUaKEGEDkWrZsiVmzZiEhIUHh88/P6NGjxf3006dPuHnzJvz8/NCzZ09s2LABqqqqBf5+Zv08i8rQ0BBKSkoAMvbPEiVKQF9fX5xfpkwZsTxE9PNiCCH6ydnZ2WHBggW4f/8+IiMj0bBhQ5iZmWHevHmIi4vD6dOnYWtrC2XlgvfOzHpFXFlZOc9bfr569QpeXl44efIkVFVVUbduXVSpUgVA/s8gKVu2rMLf9+7dw+TJk3H+/HloamqiXr16YvDIvK2sZVRSUhLnv337FgCynVCVK1dO/P+7d+8AZHTjGTVqVI7vKSfx8fHZyix/HwkJCTmukxv5+8osr9aId+/eISYmBnp6etnm1ahRA0BG68qff/6J/fv3AwBq1qwpth7I62fAgAHQ1NTE1q1bMWvWLMyaNQsWFhaYP38+fvnllwKXP2s9vHv3DklJSQonnHKqqqoAMupPXh658uXL5/tarVq1wrp163Dr1i1oamri+vXr8PLyEucfOXIEPj4+ePz4MbS1taGvrw8NDQ2F8TSF9e7dO1y+fDnH+paXedKkSahQoQJ27tyJY8eOQVlZGU2bNsWsWbNyDBPx8fEK+6GcfNrHjx8LFUJ+/fVXhQsMDRs2hLa2NkaOHIkjR46gefPmBf5+5rRfF0Vh92si+jkxhBD95OrVq4cKFSogMjIS0dHR+P3336Gvrw9NTU2cO3cOERERYteqb2XUqFF48eIFwsLCoKenh2LFiuHEiRP5DtDNKj09HYMHD0aZMmWwe/du1KlTB8rKyggNDRUH7haEvHUhLi5O4aQ684Bi+eDdP//8M8c7WFWtWjXHbZcuXRpv3rzJNv3169ff/FamJUuWRL169TBjxoxs8+QtM9OnT8fp06cRFBQEc3NzqKmp4e7du9i9e7e4rIqKCnr16oVevXrh6dOnOHz4MPz8/ODl5YWVK1dCSUkp28n7x48fC1S+smXLYvny5bkuU6ZMmWz1Jw+NeTE2Nsavv/6KgwcPQk1NDZUqVYKZmRkA4MGDBxg2bBjatm2L9evXo2LFigCQ59gM+ZX6zCfhiYmJCsvIu8QNHTo02/ryE20NDQ0MHToUQ4cORUxMDP766y8EBgZi3rx5mDp1arb1SpcuLd7EIDN56JW3EnwJ+UBx+Ticr/X9BJBtv8haZ0REchyYTvQfYGtri2PHjuHWrVswNzeHqqoqjI2NsWbNGnz48AGNGzfOcT0VFZWv8vqXLl1Cy5YtYWRkhGLFMq59yAc5F+ahaXFxcXj48CF+//136Orqiq03hR0wXb9+fZQoUQJHjhxRmC7vggIAtWrVQpkyZfDixQsYGBiI/969e4fFixfn2qphZmaGI0eOKHRjuXfvHm7fvg1TU9NClbOwTE1NERsbiypVqojl1dfXx5o1a3D8+HEAGZ+Fra0tbGxsxGCS9bPo06ePOIi/cuXK6NGjB5ydnfHs2TMAGSffycnJCt1lCvLARTMzM8TFxaF48eIKdbp7927s2rULQMZYjcjISIVtnzx5skDv39XVFcePH8fBgwfRsmVLMUhcv34dqampGDBggBhAEhMTcf78+Vz3P3lrw8uXL8Vp0dHR2d5PTEyM+CBAAwMD6Orqwt/fH+fPn8fnz5/RqlUrrFmzBkDGPjV48GAYGxuLdZlTHX38+DHbPr1//37o6el9UfcxOfndrOTjvL7W91NLS0uhvoCC7RdE9N/ElhCi/wBbW1sMHz4c2tra4tX4hg0birda1dbWznE9eWvAiRMnULx48SJfyTcwMMD27dshk8lQunRpHDp0SLzla3JycoG3U65cOVSuXBkhISEoV64clJWVsWPHDvEEWz7mIz8lS5ZEz549sXz5cqipqaF+/frYuXMn/vnnH/HEtVixYhgyZIg4FsHKygqxsbFYsGABatSokWtLyKBBg9C5c2f0798fvXr1wocPH+Dr64sqVaqIA72/lQ4dOmDdunXo06cPBgwYgDJlyiAsLAwHDx7Eb7/9BiDjszh69Ci2b9+OSpUqISIiQnxKuPyzMDMzw9KlS1G+fHkYGBjg3r17OHDggHgzAFtbW/j4+MDLywvdunXDzZs3sWHDhnzL5+joCAMDAwwYMACenp6oVKkSDh48iNDQULFVoGfPnggLC0P//v0xaNAgPH/+HP7+/gV6/61btxZbWTI/Rbx+/fpQUVHBvHnz0KVLF7x9+xarVq3C69evFcbuZGZpaQl1dXXMnDkTgwcPxtOnT7F06VKF5Xv37o2dO3eiX79+6NGjB1RVVbFq1SpcunQJw4cPh4qKCgwNDREQEAB1dXXUqlULly9fxvnz53NsBQEABwcHGBkZYcyYMRgxYgQqVaqE8PBwXL58GcuWLStQPWT28OFDXLp0CUBGoLh9+zYWLlyIGjVqwMnJCcDX+37a2dlhypQp8PPzg7m5Of766y9cu3at0GUmov8GtoQQ/QfY2NhARUUFDRs2FE+yLSwsACDbrXkzq1u3Ltq0aYPly5dj3rx5RX59Hx8f1K5dGxMmTMCIESNw7949rFu3DsWLFxdPkArKz88PJUqUwPDhwzFx4kQkJSVh9erVAFCobXl6eqJ3794ICQmBp6cnUlNTs93xqnv37pgyZQqOHj2K/v37Y/HixWjevDmWL18u1mNW+vr6CAkJQVpaGoYNG4aZM2eiYcOG2LhxY6H68heFlpYWQkNDUatWLUyZMgUeHh54+vQpAgMDxbufjR8/HtbW1pg1axaGDBmCiIgI+Pv7o0aNGrh48SKAjCA1cOBAbNy4EX379kVQUBB69uwJT09PABk3C5gxYwb++ecf9O/fH4cPHxZvc5wXFRUVBAcHw8bGBvPmzcOAAQMQFRUFHx8fdO7cGUDGuIP169dDU1MTw4cPx5o1a3I9Yc+qTp060NXVRY0aNdCgQQNxes2aNTFnzhzcunULAwYMwPz586Gvrw9vb288e/ZM4fkxcqVKlYKvry/i4uIwcOBAbNiwAXPnzlXYPypXrowNGzZAU1NTDA3p6elYvXo16tevDyBjTEibNm2wbNky9O3bF9u2bcO4cePQsWPHXOto5cqVaNasGRYtWoQhQ4bg+fPnCAoKgoODQ4HqIbOFCxeiU6dO6NSpE7p16wY/Pz80b94ca9euFQPV1/p+duzYET179sT69esxePBgJCQkYOLEiYUuMxH9NygJhWlrJSL6CaSkpGDfvn1o3LixwiDgUaNGISYmBtu3b/+OpSMiIvr5sTsWEf3nqKmpITAwEFu2bEG/fv2gqamJs2fPYt++fTkO6iYiIqKviy0hRPSfdP/+fcyfPx/nz59HYmIiatasiV69eqFt27bfu2hEREQ/PYYQIiIiIiKSFAemE9F/Eq+/0L8B90Mi+q9iCCGiH4b8lsJf6s6dO+LtZn9Et2/fRs+ePWFiYgIHBwcEBQXlezL77t07TJkyBY6OjjA1NUWnTp1w9uxZhWWePXuGUaNGoXHjxjA3N0evXr3wzz//fMu3kicnJydMmzYtz2Wio6PRsWNHGBkZoVmzZti6dWu+2w0ODoZMJsv279ixY+Iyd+/exaBBg9CoUSM0atQIHh4e4sP9vobnz5+jb9++Cg9iPHDgAJycnGBubo4pU6bg8+fPCut06NABO3fu/GplICL6njgwnYj+cw4cOCA+sO1H8+bNG/Tu3Rt169aFr68v/vnnH/j6+kJFRQV9+/bNcR1BEDB06FA8ePAAw4cPR4UKFRAeHo4+ffpgw4YNMDExQXJyMvr06QMlJSVMnDgRJUqUwJo1a9C9e3fs2rULv/76q8TvNH/37t1Dv3794OjoiCFDhuD06dPw8vKClpYWmjdvnut6t27dQsOGDTFmzBiF6bVq1QKQUcfu7u6oXr06Zs6cCUEQEBAQgG7dumHv3r0oVarUF5f9zJkz+Pvvv8W/379/j/Hjx6Nv375o0KABJk2aBD09PfFWvgcPHsSnT5/QunXrL35tIqJ/A4YQIqIfSGhoKNLS0rB06VJoamrC3t4eKSkpCAoKEh+Yl9XVq1cRGRmJNWvWwMrKCgBgbW2NO3fuYM2aNTAxMcGxY8cQExODgwcPonr16gAyniXj6OiIjRs3YuzYsZK+z4IICgpClSpVsHDhQigpKcHOzg5xcXEICAjIN4TY2trC2Ng4x/nbt2/Hp0+fsHz5cpQuXRoAYGRkBHt7e+zevRvdunX76u/l4cOHSEpKwoABA6Curo49e/bg+vXrAID09HQsWbIEw4cPh7IyOzAQ0c+BRzMi+uHs2LEDjo6OMDIywsCBA/Hw4UOF+deuXUPPnj1hZGSERo0aYfr06eLT1P38/ODv74/ExETIZDKsWrUK9evXR3h4uLj+4cOHIZPJsG3bNnHagQMHoKenhw8fPgAATp8+jY4dO8LQ0BB2dnZYvHhxtu4ze/bsQevWrWFgYABnZ2esW7dOYb5MJkN4eDhGjBgBExMTWFpaYubMmUhLS8v1vZ85cwZWVlbQ1NQUpzk7O+Pdu3e5tu4oKyujY8eOMDU1VZhWvXp1xMbGAsh4OF+PHj3EAAIAmpqaqFSpkrhMThISEjBjxgw4OjpCX18fjRo1wrhx4xAfH1+o9/nq1SsMHToUZmZmsLW1xY4dO3J9zcx14eDgoPDgSGdnZ9y+fTvHBxACQFpaGmJiYiCTyXLdbuXKldGnTx8xgABA+fLloaWllWNdpKenw9LSEn5+fuK0GzduQCaTKTzE8erVq5DJZAgNDcWECRMAAFZWVvDz80OlSpWgrKyMU6dO4dWrV7h+/TqqVKkCANi1axeKFy8OZ2fnfOuEiOhHwRBCRD+UpKQkzJ8/H0OHDsXcuXPx4MED9OnTB6mpqQAy+vJ3794dSkpK8PX1xejRo7Fv3z4MHz4cQMZTnTt06AANDQ2EhYXBzc0NBgYGiIiIEF8jMjISQMZ4A7nTp0/DxMQEJUuWxNmzZ9G/f39UrVoV/v7+6Nu3L1avXq3wjJHt27dj1KhRMDc3x9KlS+Hm5gYfHx+sXLlS4f3MmjULOjo6CAwMRLdu3bB27Vps3rw51/f/4MEDhaAAQOwq9eDBgxzX0dfXx4wZM6Curi5OS0hIQFRUlNgFycbGBl5eXgrrPX78GHfu3BGXycmoUaNw9OhRjBo1CsHBwejTpw/27NmDwMDAAr/Pz58/o2/fvrh27RqmT5+O8ePHY8mSJbkGCQBITEzEy5cvC10XMTExSElJwalTp+Do6Ag9PT106tQJly9fFpdp2bKl+HR4ufPnz+P9+/c51oWysjKsra0LtA9VqVIFLVq0wODBgwEAK1euRMeOHVG+fHl4eHjA09MTjRs3RpkyZdClSxekpqbC398fI0aMyLUuiIh+ROyORUQ/FEEQMG/ePLFbUa1atdC6dWvs3bsXbm5uCAwMRNmyZREUFAQ1NTUAQI0aNdCtWzdERUXB3NwcFStWhLKystgdx87ODlu2bBFf49y5c2jQoAHOnz8vTjt9+jQ6d+4MAPD19YWRkREWLVokrl+6dGlMmDABffv2ReXKlbFw4UK0bt0af/75JwCgcePGUFJSQmBgILp27YrixYsDAExMTDB58mQAGVfFjx07hpMnT6Jr1645vv+EhASUKFFCYZr874SEhALX49SpU5GQkIDevXvnOD8lJQVeXl5QU1NDly5dclzm06dPSE1NxZQpU2BnZwcAsLS0xMWLF3Hu3DmFZfN6n8ePH8etW7cQFhYmfiY1atRAu3btci2//L0Wti5u3boFAHj9+jVmzJiB5ORkrFixAj179sS2bdtQu3btbOt8+PAB3t7eqFSpElq1apXjdu3s7DB58mQkJSVBU1NT3IeuXLmC1NRUqKqq4vTp07C1tYWOjg6qVasGANDT04OOjg4AYMiQIejatSs+fvyIX3/9FUpKSggNDUXVqlVhYWGB2bNn4/jx46hfvz4mT54srkdE9CNiSwgR/VBKliwpBhAAqFu3Ln799VexK1JkZCRsbGygrKyMtLQ0pKWlwdjYGFpaWtnuBiVna2uL58+f48GDB3j//j1u376Nvn374uHDh3j16hXu37+PJ0+ewN7eHklJSbhy5QocHR3F7aelpcHOzg7p6emIjIzE/fv38fLlSzg4OGRb5uPHj7hy5Yr42kZGRgpl+eWXX5CYmFikuinIeAFBEDB16lTs2rUL48ePR4MGDbItk5KSguHDhyM6Ohpz587FL7/8kuO21NXVsWrVKtjZ2SE2NhZ///03Vq9ejXv37oktU3J5vc8LFy6gdOnSCmM09PT0xO5Iub0PAApdsTJPz60uGjVqhGXLlmH58uWwsbFBkyZNsHLlShQvXhzBwcHZlo+Pj0e/fv0QGxuLxYsXK3SDy8zW1hZpaWm4cOECBEHA+fPn0bdvXyQlJeH69etITEzExYsXYW9vn+t7AoCyZcuiWrVqUFJSQnJyMpYtW4YRI0YgNDQUp0+fhp+fH5SVlTFlypQ8t0NE9G/HlhAi+qGULVs22zQdHR28fPkSQMataMPCwhAWFpZtuVevXuW4TQMDA2hrayMyMhJly5ZFuXLl0Lx5c0yePBnnz5/H69evUbFiRchkMrx48QLp6elYsGABFixYkONrvHv3DkBGV6VRo0blWY6sJ7XKysp53m5XS0sLHz9+VJgm/1tLSyvX9YCMcDF27Fjs378fo0aNgru7e7ZlPnz4AA8PD1y4cAGzZ8/OdxzCkSNH4OPjg8ePH0NbWxv6+vrQ0NBAenq6wnJ5vc/4+Hhoa2tn23b58uVzfV35e81aF/JgU7JkyRzXK1++PBwdHbNty8TEBDdv3lSY/vz5c/Tv3x+xsbFYunRptiCVWbly5VC/fn1ERkZCR0cHCQkJcHJyQo0aNRAdHY24uDgoKSmhUaNGuW4jq3Xr1sHAwABGRkaYO3cufvvtN9StWxc9e/ZE586d8fnzZ6ioqBR4e0RE/yYMIUT0Q8k84Fnu9evX0NXVBZBxQtmkSZMcuxDldKILZJwQ29jYIDIyEuXLl0fDhg1RrFgxmJiYIDo6Gk+ePBG7G8m7+wwePBhNmjTJtq0KFSrg/fv3AIA///wThoaG2ZapWrVqAd9tdjVq1Mg2OPrx48cAkOfYjeTkZAwaNAiRkZGYMmVKjvUTFxeHXr164cGDB1iyZEmO7y+zBw8eYNiwYWjbti3Wr1+PihUrAgCGDRuGe/fuFfg9lSlTBm/evMk2XR7mclKiRAmUL19efO9y8r9r1KiR43pRUVF4+fIlXF1dFaYnJycr7B8PHz5Ez549kZSUhNWrV+d6J63MbG1tERkZiXLlykFPTw/FixeHhYUFoqOj8ezZM5ibm4vd8PKTkJCAVatWISQkBEDGbYPLlCkDIOMmAp8/f8bbt29Rrly5Am2PiOjfht2xiOiHEhcXp/AAvX/++QexsbGwsLAAAJiZmSEmJgb6+vowMDCAgYEBKlWqhAULFuDOnTsAcu6qY2tri3PnzuH8+fNo2LAhAKBhw4aIiIjAuXPnxG40WlpaqFevHh4/fixu38DAAKqqqli4cCGeP3+OWrVqoUyZMnjx4oXCMu/evcPixYsLNXYjq0aNGuHMmTMKXbYOHz6MMmXKoF69ermuN3r0aERFRWHBggU5BpDU1FQMHDgQjx8/RnBwcL4BBACuX7+O1NRUDBgwQAwgiYmJOH/+fKGeBG5paYkPHz4odJe7f/9+vg8HlI8tyXxXssOHD0NXVzfXk/OzZ89i3LhxCq1Rr169woULF8R9KD4+Hn369EFaWhpCQ0MLFECAjHEh165dw6lTpxT2ofPnz+P06dNikAXy7zoXHByMxo0bi+G6bNmyYplfvXoFFRUVMZQQEf2I2BJCRD8UNTU1jBw5EqNHj0Zqairmz5+PevXqwcXFBQDg4eGBzp07Y9iwYWjfvj1SUlIQGBiIZ8+eieMfSpUqhaSkJBw+fBiGhoaoUKECbG1tMX78eLx69QqzZs0CAJibm2Px4sVQVVVVGIcydOhQ/PHHH9DS0kLTpk3x9u1b+Pr6QllZGbq6uihWrBiGDBmC2bNnA8g4WY6NjcWCBQtQo0aNL2oJ6dq1K9avX48BAwagb9++uHnzJoKCgjBq1ChxIH5CQgLu3r2LatWqQUdHB4cOHcKhQ4fg5uaGypUr49KlS+L2NDQ0UK9ePYSGhuLKlSvo378/VFVVFZYpVapUjq0s9evXh4qKCubNm4cuXbrg7du3WLVqFV6/fi2WpSBsbGxgbm6OMWPGYPTo0ShevDh8fX1zfOZJZn379kWHDh0wbNgwdOzYEWfPnsWuXbvg6+srLhMXF4dHjx6hTp060NLSQqdOnRAaGoqBAwfijz/+QEpKCgICAqCtrY3u3bsDAJYsWYLY2FhMnDgRCQkJCnVRrly5XD8/Y2NjaGpqKtxYwMLCAu/fv8f79+8VxoPIH3h46NAh2NjYKGwzLi4OGzZsUHj6u729PTZu3IgGDRpg3bp1sLOzQ7FixXJ8j0REPwSBiOgHsWTJEsHFxUXYuHGjYGNjIxgbGwvDhg0TXr9+rbBcdHS00L17d8HQ0FAwNzcXBg4cKNy+fVuc/+bNG6Fdu3aCnp6eEBQUJE5v27atYGFhIaSnpwuCIAifPn0SDAwMhJ49e2Yry5EjR4R27doJ+vr6QqNGjYSRI0cKT58+VVhmy5Ytgqurq6CnpyfY2NgI3t7ewrt378T5urq6wsqVKxXWGTx4sNC9e/c86+HKlStCp06dBH19fcHBwUFYvny5wvyIiAhBV1dX2LZtmyAIgjBu3DhBV1c3x3+urq6CIAhC9+7dc11mwIABuZZl165dQrNmzQR9fX3B0dFR8Pb2FjZs2CDUq1dPeP78eYHfZ3x8vDB+/HihYcOGQqNGjYSgoCChY8eOwtSpU/Osi5MnTwq//faboK+vLzRr1kx8z3Lbtm0TdHV1hYiICHHanTt3hAEDBgjm5uaCqampMGTIEOHJkyfifEdHx1zrIr/yDBkyRKhXr57w/v17he05OzsrLJeYmCj06dNH0NPTy7ZNHx8f4c8//1SYlpycLIwZM0YwNTUV3N3dxbrN7T0SEf3bKQlCIdrMiYiIiIiIvhDHhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIiktR/PoRwXP5/z7/9M/+3l48os2+5v/4M34Uf7T38aOWlnwv3vx9TUT+3QoWQ8PBwyGSyPP+5u7sDAJycnDBt2rQ8tyeTyRAcHFykgn+N9Q8fPgxvb2/xbz8/P5iYmBR5e/9FqampGD16NIyNjWFubo4nT5587yLlafPmzQrPECiKGzduoHXr1tDX18egQYO+TsH+X3R0NIYOHSr+Lf/OxcXFfdXXkRMEAdu3b0e3bt1gYWGBhg0bolOnTti1a5fCcgUph5+fX57Hhr59+2ZbJy0tDVZWVtDT08PLly+zzY+MjMxxW8bGxnB1dcXKlSvzfH+5rS//t3HjRoXlZ86cCZlMlud2C1pnQMbzOhYsWIBmzZpBX18flpaW6N+/PyIiIhSWy1q/8r/19PTEp69nNWHCBMhkMoXjrLu7OwYOHJjr3zlxcnLKs4727t2b67rf6hj6/Plz9O3bF2/fvs1zOR8fHzRs2BCmpqY4f/58vtuNjY2FTCbDgQMHvmp55TL/7qWkpGDGjBk4cuRIjvNz86W/a0VV0DrPTP79unr16jcs2Y/h9OnTaNq0KQwMDDB9+vSvvv1nz57hjz/+gJmZGaytrTF37lykpKQUeP2QkBC0atUq2/QDBw7k+L1fv379F5X36dOn6Ny5MwwMDNCmTZscl8m8r8fHx2PUqFEKD6KlH4O/vz82bNhQpHUL9bBCBwcHhIWFiX+HhIQgKioK/v7+4rTCPCgpLCwMlStXLkwRvur6ISEhKF68eJHXJ+DUqVPYvXs3Ro0aBRMTE1SqVOl7FylPy5Ytg4ODwxdtIzAwEG/fvsWyZcvwyy+/fJ2C/b+tW7fi/v37X3WbuUlLS8OQIUPw999/o3PnzujXrx9UVFRw8uRJjBs3DlevXoWXl1ehtqmhoYGQkJAc55UsWTLbtBMnTiAlJQXlypXD9u3bcz1h9vHxUXhYXlxcHLZu3Yp58+ZBU1MT3bp1y7NcWdeX+/XXX8X/p6WlYe/evahbty62bt2Kfv36ZVu+MHUmCAL69euHly9fYuDAgahRowbi4+Oxbds29O7dG4GBgXB0dMyz3J8/f8bRo0fRtm1bhempqakKJ7dfysXFBX369MlxXvXq1XNd71sdQ8+cOYO///47z2Vu3bqFNWvWoGfPnmjatCnq16//1ctRWP7+/uJDCF++fIl169aJT04vqC/9XSuqgtQ55W7BggXQ0NDAihUrvvrvYEpKCvr06QMNDQ3MnTsXz549w/z585GcnIw///wz3/UPHTqEefPmoUaNGtnm3bp1C9WrV8fcuXMVpn/JA1UBYO3atbhx4wYWLVqEihUr5rhM5n39xo0b2LNnD3r16vVFr0vS8/Pzw9ixY4u0bqFCiI6ODnR0dMS/9+7dCzU1NRgbGxfpxYu63tdan76c/Cpthw4dFPaNn9m7d+/QoEEDNG7c+HsX5YssW7YMx44dw4oVK2BraytOt7OzQ4UKFbBgwQK4uLgU6iRKWVm5UN/LnTt3olGjRqhUqRK2bt2KAQMGQElJKdtydevWhYGBgcI0e3t7ODs7i60Seclp/axOnTqFt2/fYsGCBejVqxeio6OzvffC1FlUVBQuXryIzZs3w8jISFy2SZMm6NSpEwICAvINISYmJjh48GC2EHL27FkoKSl9tRBcrly5H+54Kj/2tGrVCoaGht+5NBkaNGjwxdv40T4HyvDu3TvY29ujUaNG+S4bGRmJHj164MiRIwU62d+9ezcePXqEI0eOiCf06urqmDJlCjw8PFCuXLkc10tISEBAQABWr14thuOsbt26BT09va++371//x5Vq1aFs7NzrstwX6dvOiYkOTkZU6ZMgYWFBczMzDBu3DgkJCSI8zM3xX3+/Blz586Fg4MD9PX10bJly2xdJbL6kvXd3d1x7tw5HD9+HDKZDLGxseK8ffv2wcXFBQYGBmjXrh0uXLigsO61a9fQs2dPGBkZoVGjRpg+fTqSkpJyfS15k/WmTZvQuHFj2NvbIzY2FqmpqViyZAlcXFygr68Pc3NzeHp64tmzZ+K6Tk5OWLFiBby9vWFhYQFTU9Ns9fjp0yfMmDEDVlZWMDU1hZeXFxYuXAgnJyeFcqxdu1bsFuLq6op9+/blWb8AEBUVhW7dusHU1BTW1taYNm0aPn78CAAYP348xo8fDwCwsrIS/5+TgwcPol27djAyMoKTkxOWLVum0Ifw0KFDaN++PYyNjWFvbw9fX1+kpqYq1EPWbgwzZ85UeI8ymQzh4eEYMWIETExMYGlpiZkzZyItLU3cxpMnTxAaGgqZTAYASExMhJeXFxo3bgxDQ0O0bdsWBw8ezPV9yGQynDt3DidOnIBMJkNkZGS+9QRk7G+TJ09G3759YWpqijlz5mTb9vjx47F9+3bcuXNHYdsAEBERgTZt2sDAwACurq7ZroI/fPgQHh4eMDExQcOGDTFmzJg8u06lpqZi3bp1cHR0VDiZluvRowe6desGZeVvd4iIj4/HsWPHYGtri9atW+PRo0cK7zk/Kioq0NDQ+Grl2bFjBwwNDWFlZYVatWphy5YtCvMLW2dv3rwBAKSnpyssp6ysjBEjRqBdu3b5lsnFxQWnT59W2JeAjC4UTZs2RbFihbqO9FV9yTF0165daN++PYyMjGBkZITOnTsjKioKQEZXtAkTJgDIOK74+flle20/Pz+x62/Hjh3F/798+RITJkxA48aNoaenh8aNG2PmzJmF6rYid+PGjWzfwzVr1kAmkyl0pwsODoa1tTUEQRCPU7GxsWjSpAkAYNiwYWL5gML9Lvr5+aFdu3bYs2ePWJ/t27fPVp9ZFbYecqvzjx8/Ys6cOXBycoKhoSE6dOiQb2tJfr+PuR0Lr1y5gv79+6Nhw4bQ19eHi4sLNm3apFBGS0tLnD17Fm3atBF/57MeC2/evIl+/fqJx+IJEybg3bt34vzCHivzqwd5F78nT55gw4YN2b4LX8OZM2fQoEEDhRYFZ2dnpKWl4ezZs7mut3XrVuzevRvz58/Pdj4gd+vWLfH3sKAEQcDmzZvRunVrGBoaolmzZlizZo0438nJCeHh4bh79674u5wT+b4uD2VAxgVN+bnE5cuX0a1bN5iYmMDCwgJDhw7Ns8u3/PuyY8cONG3aFIaGhujVqxdevnyJTZs2wcHBAWZmZhg9erTCPpmYmIjp06fD2toahoaGcHd3x/Xr1xW2nd/+CQArV64Uu+Q5OzsjICBAPP7n1KU5Pj5eoX7k5Z81axYaNmyIzp07A8hogV+8eDEcHBzEY2rmz11+jhkREYGOHTvC0NAQrVq1QnR0NKKjo+Hm5gYjIyN07doVDx8+VChzXueF8n376NGj6Nu3L4yMjGBra4ulS5cqfIYAMHfu3Fz3sbx80xCyfft2vH//Hr6+vhgyZAh2796t0HUrs+DgYGzbtg3Dhw9HcHAwbG1tMWXKFJw6dapAr1XY9b29vdGgQQOYmpoiLCwMFSpUAAAkJSVh0aJFGDp0KBYvXoykpCQMGTJEPJG9e/cuunfvDiUlJfj6+mL06NHYt28fhg8fnm8ZAwMDMW3aNIwYMQJVq1aFj48P1q9fj/79+2PVqlUYPnw4zp49i1mzZimst3z5csTHx2PhwoUYPnw49u7dq7ATTJw4EeHh4fD09MSCBQvw6NEjrF69WmEb/v7+mDNnDlq2bIlly5bB2toaI0eOxP79+3Mt74kTJ9CjRw+UL18eixYtwpAhQ7B3714MHDgQ6enp8PDwwODBgwFkfPk8PDxy3M5ff/2FIUOGQCaTwd/fHz169IC/vz9WrFgBIKNJ1tPTEwYGBvD390f37t2xatUq8UexMGbNmgUdHR0EBgaiW7duWLt2LTZv3izWQfny5eHi4iJ2K5wzZw4iIiLg5eWF5cuXo3bt2hg2bBju3buX4/bDwsIU9hs9Pb1860kuPDwcVatWxZIlS9CiRYts2/bw8IC9vT1+/fVXcdtyM2fOhLu7OwIDA1GyZEmMGDFCPMl9/fo1unbtiqdPn2Lu3LmYOnUqLl26hL59++Z60nHt2jXxyl1ONDQ08Oeff8LU1LQAta4oLS0tx39ZB67t3bsXgiCgefPmMDIyQo0aNbKd+Mulp6eL20lJScHz58+xYMECxMTE4Lfffsu3TJnXl//L/Nl8+PABx44dQ+vWrQEAbdq0wYEDB/DhwwdxmcLWmbm5OYoXLw5PT0/4+fnh8uXL4nHE2toaXbt2zbfcTk5OSE9Px4kTJ8RpaWlpOHLkCJo3b57v+gUlCEKOn9nnz59zXaeox9ADBw5g7NixcHBwQFBQEHx8fBAfH48RI0YgJSUFDg4OCseVjh07Znvtjh07it1QfHx84O3tjfT0dPTr1w/Xr1+Ht7c3Vq5ciTZt2mDt2rUK3YgLqn79+ihfvrxC4Dh37hyAjLFbcqdPn4atra1CC16FChXE37qRI0cqjJspzO8iADx48ABLliwR96NPnz5h2LBhYn1mVZR6yKnO5dsJDw/HgAED4Ofnh8qVK2PAgAG5/q4W9Pcx67Hw6dOn6NGjB4oXL47FixcjICAANWvWhLe3N27evCmu9/HjR0ycOBHdunXD8uXLoa2tjREjRogh48mTJ+jatSsSEhIwd+5cTJo0CadPn8aoUaMAFO1YmV89VKhQAWFhYQq/LfLvQmaZv2PyY0/m41JeA3sfPHiAatWqKUzT1taGlpYWHjx4kOt6TZo0weHDh3McCyKvzydPnuD69etwcXGBnp4eWrdurXC8ycnChQsxZcoUODk5ITAwEM2bN8fcuXOxaNEiABm/tZl/y/LrAq2np6fwffbw8EBSUhIGDBiAX375BYGBgZg+fTquX7+OkSNH5rmt+/fvY8WKFRg7dixmzJiBy5cvw93dHdu2bYO3tzcGDhyIPXv2YO3atQAyPpfBgwdj7969GD58OBYvXgw1NTW4u7vj0aNHAFCg/XPfvn1YvHgxevXqheDgYHTs2BF+fn7i+UdB3bp1C1evXoWfn5845nTy5MlYvXo1evTogYCAANSqVQv9+/fPdjFizJgx6NChA/z9/ZGeno7hw4dj4sSJ6NWrF2bNmoV79+4pXMwt6HnhhAkTYGRkhGXLlsHR0RG+vr7iPiI/pri7u+d5HMvNN72MVrNmTSxcuBBKSkqwtrZGRERErlc6o6Ojoa+vDzc3NwCApaUlNDQ0oKmpWaDXKuz6derUgZaWFooXL67QJCgIAubNmydOk/cBv3v3LurVq4fAwECULVsWQUFBUFNTAwDUqFED3bp1Q1RUFMzNzXMtY8+ePRWSYlxcHMaOHYsOHToAACwsLHD//n3s3r1bYb2KFSuK9di4cWOcO3cOJ0+exJgxY3D//n3s2bMHPj4+4pXVRo0aiVfhgIy0HRQUhH79+ok/Bo0bN8bHjx+xYMGCHE+IAWDx4sUwNDRUGMhdtWpV9OvXD8ePH4eTk5N4YNTT08u1O9bSpUvRqFEj+Pj4AABsbW3x6tUrXLhwAenp6fD19YWrqyumTJkilq1kyZLw9vZGv379UK9evVzrNCsTExNMnjwZQMYVvWPHjuHkyZPo2rUrGjRoADU1NYWuJ9HR0bCxsRHrwMzMDOXKlcv1B97Y2DjbflOQegKAEiVKYNKkSVBVVc1x29WqVYOOjg6ePn2arZl64sSJcHV1BZDRLbJdu3a4dOkSmjRpgpCQEHz69AmrVq0SPwNDQ0O4uLhg37594ncis+fPnwPAV+97npiYqBCeMluxYgXs7OzEv3fu3AkHBweUKVMGQMaJ/7Jly/D+/XuULl1aYd3ff/892/aqVq0KLy8vhavMuclp/W7duok/fPv27UN6ejpatmwplsXX1xe7d+8Ww0Jh66xcuXJYunQpJkyYAH9/f/j7+6N48eJo1KgRunXrVqDufFpaWrCyssKhQ4fEssmvgBWk20dBbdiwIceBheXKlcPp06dzXKeox9BHjx6hW7duGDJkiLiOqqoqPD098eDBA+jq6uZ7XKlYsSLq1KkDIKOrXZ06dfDs2TOULl0aXl5e4jHDysoKp06dQlRUVIH2k6xsbW3F3yxBEHD+/Hk0aNBAHAT/6dMnREdHY/bs2QrrqampiWNUqlevLpYVKNzvIpBxorhmzRqxy9nnz5/h4eGBmzdvQl9fP9vyL168KHQ96OjoZKvzo0eP4sKFC1i5cqXY8mdvb49OnTph0aJFObYGFvT3Meux8MSJEzA2Nsb8+fPFaUZGRrC0tER0dLT4PlJTUzFmzBjxu1C2bFm0adMGkZGRcHFxQUhICFRUVLBy5UpxfKq6ujrmzp2Lt2/fFulYefz48TzrITw8HMbGxtl+W7Lavn17tgtrTZs2Ff+/du1aWFpa5rhuQkICSpQokW16iRIlFFrRsso85i0nt27dgiAIiI2Nxfjx46GiooINGzZg0KBBWL16dY7HmLdv32L16tXo27cvRowYASDjN1sQBAQHB6Nnz55o0KBBrr9lOdHS0lL4PlerVg1XrlzBu3fv4O7uLt48QltbGxEREUhPT8+1lT4xMRGzZs0Su8AeP34ce/fuxdGjR1GlShU4Ojri+PHjuHz5MgDg77//RkREBFavXg1ra2sAGd97V1dXLF26FD4+Prhz506++2dUVBSqVKmCrl27QklJCRYWFihWrFiOgTQvaWlpmDhxoth9+N69ewgPD8eMGTPECzJ2dnZ49eoVfH19xTAFZASBTp06AcgITt7e3pgzZ464X9+5c0e84UBhzgtbtGgh3jDH0tISf/31F06ePAl7e3vx861UqVKRuqN+0xBiZGSkcHWoatWquHPnTo7LmpiYwNfXF+7u7nB2doaTk5O4gxfEl64vp6KiotC/uEqVKgAgXhGNjIxEkyZNoKysLJ6oyk9Mz549m2cIyfxDBEA8aX3x4gViYmIQExODCxcuZLsiY2BgoFCPFStWxI0bNwBA7MKQud+lpqYm7O3txR+2S5cu4dOnT3BwcFA4ubazs8O2bdvw+PHjbAerjx8/4vr16xg3bpzCdFtbW5QuXRpRUVEFanpLTk7GjRs3sh18x4wZAyDjSxEXF5ftqm6rVq3g7e2t8ANUEJn73gPAL7/8gsTExFyXNzExwebNm/Hy5Us4OjrCwcEhz25lWRWmnqpVq5ZrAMlP5jv45LRPGhsbo1SpUuLnW6lSJdSuXRtnz57N8YdVRUUFwNe/HaKGhkaud1WpWbOm+P+HDx/i4sWLmDNnDuLj4wFkXPVfvHgxdu3ale1Eac6cOahduzY+ffqEtWvXIjIyUmw+Lwj5+pmVLVtW/P+OHTtgaWmJYsWKIT4+HiVKlICJiQm2bt0qhpCi1FmjRo1w+PBhREZG4uTJk4iIiMCxY8dw9OhRDBw4MN+regDQrFkzzJo1CykpKVBTU8OBAwfg7Oz8VbtitWjRIse7lxVlf83vGDpgwAAAGT+CMTExuH//Po4ePQoAReo2JVepUiWsW7cO6enpePDgAR48eICbN2/izZs3RQ7btra22L17NxITE/Hw4UN8/PgRPXr0wLRp0/D582dERUUhLS0NNjY2Bd5mYX4XAaBYsWIKYUPeJSe3LsBfqx6ioqJQokSJbGGjZcuWmDVrVo4nvwX9fcx6LLS3t4e9vT0+ffqEmzdv4sGDB+Idt7LuE5lParPWxcWLF2Fubq5wg5wmTZqIF+WKcqwsSD0U5IY8jo6O2Lp1KwDgn3/+gbe3N5YuXYry5csDyDg+CoKQrfVR/j3PaaycIAhf1GW2Tp06CAoKgpmZmfgebGxs0KZNG/HiYVaXL19Gampqtt9sV1dXBAUF4fLly/mOdSuIWrVqoUyZMhg0aBBcXV1hb28PKysrWFhY5LmekpKSwvelbNmy0NHREY9DAFCmTBmF309NTU2Ym5srnB81btxYPC4VZP80MTHBhg0b0L59ezRv3hwODg45HlMLIvNvlbz11c7OTqF89vb2WLhwocL3I/NxVz5OKHNdZH7fBTkvlO9zmb9zysrKqFChQp7nVYXxTUNI1lYIJSWlXH/ABwwYAE1NTWzduhWzZs3CrFmzYGFhgfnz5xdo8OWXri+nrq6u8KWW/1/efPru3TuEhYXl2Kz96tWrPLed9YrehQsXMGXKFNy6dQslS5ZE/fr1oa6unm29vOrx7du3UFVVzTboLPNANXlTtbx/YU7lzhpCPnz4AEEQFE7UMr+PvK6+ZCYfPJrTdvKar6WlBXV19QK/jlzWulJWVs7zpHHSpEmoUKECdu7ciWPHjkFZWRlNmzbFrFmzCvTDUph6yq0OCiLz2Iec9snLly/n2AIh/4HLSn4y8vTp01xf88WLF4Ue+KysrJzvAHAg46QfQLbwBgBbtmzJFkJq164tbtfU1BS9evXCH3/8gU2bNhWoP3Pm9bN6/Pix2Kyd00WE69evo0GDBkWuMxUVFVhbW4uBKTY2FhMnTkRQUBA6dOiQrZtFVs7OzpgyZQr+/vtv2NnZ4fDhw5g/f36e6xSWjo5OgT63gsjvGPrq1St4eXnh5MmTUFVVRd26dcUThC8NxVu2bIGvry9ev36N8uXLw8jICOrq6kXero2NDdLT03HhwgXcvXsX+vr6sLGxQWJiIq5fv47Tp0/D2Ng4W8tdXgrzuwhktKrkVZ85+Rr1EB8fn+OAZ/m0rOOUgIL/PmY9Fn7+/BmzZ89GWFgYUlNTUa1aNfGmEFnLnNex8P3793letCrKsbIg9VCQ3wptbW1oa2sDgHgCp6urqzAwPfPYHLlbt25BS0srx/pOTEws1B1JsypVqlS27qXy49XOnTtzXEf+m521TuSfaWF/s3OjpaWF9evXIyAgANu3b0doaChKlSqFESNG5NmVVVNTU7xglHlabt69e4ekpKQcWxXlQbkg++dvv/2Gz58/IzQ0FAsXLsSCBQtQr149LFy4MNsFsLwUL15c4Y6D8vO3zL0IMst8S+2cWstyGzdZkPNCeStO1m3kd15VGN9vVGMWKioq6NWrF3r16oWnT5/i8OHD8PPzg5eXV77PAvga6xeUlpYWmjRpgi5dumSbJz/AFMSHDx8waNAgmJqaws/PT7wN5ty5cxX6wOanQoUKSE1NRXx8vEIQyTz4SX5r1ICAgBxPKjNfoc68jpKSkjjuILPXr1+LXWjyI/9SZB349/z5czx8+FAMZllfJz4+Hp8+fVJ4naw/ul8jiWtoaGDo0KEYOnQoYmJi8NdffyEwMBDz5s3D1KlT813/a9XTl9DS0oKdnZ3C80XkcjooARCby0+dOpXjvpySkoLWrVvD2dk52xilLyUIAnbt2gUnJ6dst2M8e/Ysli5diitXruR6xyNlZWXMmDEDrq6umDhxIrZs2fJFVwN37NgBDQ0NLFu2TGE7nz9/xqBBg7BlyxZx/ENh6kzebz8gIEBhuapVq2LixIlo06YN7t+/n28I0dbWhrm5OQ4dOiT+GFhZWRX5/X5vo0aNwosXL8RxT8WKFcOJEyfyvCFEQZw7dw6TJ0+Gh4cHunfvLh5b5N1di6J06dIwNDREZGQk7t+/j4YNG6JChQqoUaMGzp8/j9OnT4tdg/4tvlY9lC5dGq9fv842XR4mcjq2FfX3cenSpdi8eTPmzJkDe3t7FC9eHElJSWLLQUFpaWll+61JSUnB2bNnYWJiUqRjZVHqoagyt5ZkVqNGjWyD3d++fYuEhIQcf78L6vr16/jnn3+yjbtKTk7O9fOSv9/Xr18rnE/I6+hr1kfdunXh6+uLlJQUnD9/HiEhIZg6dSr09PSy9XooqpIlS6Js2bJYvnx5rssUdP9s27Yt2rZtizdv3uDo0aMICAiAp6cn9u/fL7YqZD55L8g5jPwcY+PGjTm2fmtra+c5Liiv7QJ5nxdmvqHDt/KveWJ6nz59xDEDlStXRo8ePeDs7Kxwp6ivvX5RTlzMzMwQExMDfX19GBgYwMDAAJUqVcKCBQvybFLPKiYmBu/fv0fPnj3FAJKeno4zZ84UKmGamppCWVlZbDYEMg66mQcOGhkZQVVVFW/evBHLbGBggDt37mQ7QZIrUaIE6tevLz7US+7UqVP48OFDgQcsa2lpQVdXF8ePH1eYvm7dOowePRq1atWCtrZ2tteR36FB/jpaWloKD7NLT0/HxYsXC1SGzLKeZLZq1Uq8q0etWrUwePBgGBsbF3i/+1r1lFP5Ckq+T8pkMvGz1dXVhb+/f64PcFNWVka3bt1w9OhRnDlzJtv8lStX4v379+JA7a8pOjoasbGx+P3332Fpaanwr3fv3lBVVc335KNatWro06cPrl27luudVwpq165dYlN/5rJYW1vDwcEBe/bsQXJycqHrrGrVqjhx4kSOx4UHDx5AWVk5x2eX5KRZs2Y4evToN+mK9SWKsr9eunQJLVu2hJGRkfg+5Mcr+bGvqNtVUlLC4MGDxRPvFy9e4Pbt21901c7W1hbnzp3DhQsXxJayhg0bYv/+/bh9+3auNyrIejVWKkWth6x1bmZmho8fP2YbhL5//37o6enl2Gpf1N/HS5cuQV9fHy1atBCvAmfdJwrC1NQUUVFRCq0GZ8+exYABA/DmzZsiHSuLUg9Fpa2trfAbLW+dbNSoEa5duyaOSwMyHhSqqqqaZxfw/Ny4cQOTJk1SuBNUcnIyTp48mWu3JwMDA6iqqub4m12sWLEi3y476/fl5MmTsLKyQlxcHNTU1GBlZSWO98yrNbqwzMzMEBcXh+LFiyvU++7du8UH0BZk//Ty8hLDbdmyZdGxY0d06NBBPJeQt1hlPo/JfIOLvMonCAI+fvyoUL6zZ89izZo1Rf4tKMp5YW6+5CLgv+OXDBkVLe8faWBggHv37uHAgQPo2bPnN1u/VKlSuHHjBiIjIwucqj08PNC5c2cMGzYM7du3R0pKCgIDA/Hs2bNCDcqpVasWSpQogcDAQKSnpyM5ORkbNmzAzZs3xeb5nPqAZlW9enW0bt0aM2bMQGJiIqpUqYK1a9fi1atXYvcRHR0duLu7Y/bs2Xj//j0MDQ1x8+ZNLFq0CE2aNMm1OXfIkCHw8PDA8OHD0a5dOzx79gwLFy6EiYlJrk2DOfnjjz8wbNgwTJ48Gc2bN8ft27exdu1ajB07FioqKvD09MT06dNRunRpNGnSBLdu3YKfnx+aN28OXV1dABlNkatXr8a6detQp04dbNq0CW/evMn16lVuSpUqhX/++QdRUVFo2LAhDA0NERAQAHV1ddSqVQuXL1/G+fPnC9QK8rXrSV6+58+f4/Tp0zk2D+ekd+/e2LlzJ/r164cePXpAVVUVq1atwqVLl/K8a5v8yd0DBw5E9+7dYW1tjZSUFBw6dAg7duxAnz59sl1xDwsLy9a0nfle8Onp6bh06VKOr6ekpAQjIyPs2LEDJUuWzLEffenSpWFnZ4c9e/bkOzanf//+YreTFi1aFHpfADJ+BB49epTr+LHffvsNf/31Fw4cOAA3N7dC1Vnfvn1x8OBBdOvWDT169ICpqSmUlJRw/vx5rFq1Ct27d8934Khcs2bNMH36dISHhyvcGa8gYmNjFW6fKWdvby9eRX39+nWun1vZsmVzLWdRjqEGBgbYvn07ZDIZSpcujUOHDom3U09OTha3C2TcutvGxqZAz1IwMDBAeno6Zs2ahebNm+PZs2dYunQpUlJS8ryFen5sbW2xZMkSKCsrK9z5bOvWrahQoUKuD0mUX2k8c+YMatSoUaixbV+iqPWQtc4dHBxgZGSEMWPGYMSIEahUqRLCw8Nx+fJlLFu2LMdtFPX30cDAACtWrMD69euhq6uLq1evIiAgAEpKSuI+URA9e/YUH3rap08fJCYmYv78+WjWrBlq1qxZpGNlUeohP5aWlrh161aBl2/VqhWWLl2Kfv36YdiwYXj58iXmzZuH33//XexGlpKSguvXr6NixYq5Phwwq+bNmyMoKAjDhg3DiBEjoK6ujuDgYCQmJop3S8tKfj4RHBwMFRUVmJubIyoqCsHBwejdu3ehuiZmJv++nDhxAsWLF4ehoSEEQYCnpyf69+8PVVVVhISEoFSpUrkO4C8KR0dHGBgYYMCAAfD09ESlSpVw8OBBhIaGiucBBdk/zc3NMW7cOCxcuBDW1tZ4/vw5Nm7cKN58wNLSEurq6pg5cyYGDx6Mp0+fYunSpeINHHJTv359uLi4YMyYMfD09ETt2rVx7tw5cX8oagAoyHlhQVtCSpUqhfPnz6Nhw4YwMjJCQkIC7t69K95sJy//mhAyaNAgpKenY+PGjfD19UW5cuXQs2dPeHp6frP1e/XqhREjRqBfv365PuU5K319fYSEhMDX1xdDhw6Furo6TE1NMXfu3EL1ny9ZsiT8/Pwwd+5cDB48GNra2mjYsCEWL16MoUOH4vLlywV+kM+UKVOgoaEBX19fpKWloVWrVmjevDnu3r0rLjNmzBjo6Ohg8+bNWLJkCSpUqJBv/Tg5OSEgIAABAQHw8PBAmTJl0KpVK4wYMaJQV/maN28OX19fBAYGYvv27ahcuTLGjRuH7t27AwC6d+8ODQ0NrFq1Clu2bEGFChXQu3dvhVv+Dho0CK9evcKiRYtQrFgx/Pbbbxg4cGCug6BzM3DgQPGuW3/99RcmTZqE4sWLY9myZXjz5g2qVKmCcePG5Xhb0Nx8rXoCgE6dOuHYsWMYOHBgtifY5qZy5crYsGED5s2bhzFjxkBJSQl6enpYvXp1nk+Rlv/grFu3Dnv27MHmzZuhrKyM2rVrY+HChTneNS3zHcDkGjduLIaQ5ORk8e4cWamoqODixYv466+/4OTklOvBt3Xr1jhy5Aj279+f58mnlpYWhg4dCm9vbyxfvrxAg7yz2rVrFzQ0NHK9mm1nZ4fSpUtj69atcHNzK1Sdyb9vQUFB2Lt3r9gttHbt2hg/fnyh9rFy5crB1NQUd+/eLXRXrLt374qtxJmVL19eDCF//fUX/vrrrxzX79ChA2bOnJnjvKIcQ318fDB16lRMmDAB6urqkMlkWLduHfr3749Lly7BwsICVlZWaNy4MaZPn47ff/+9QE+FtrKywoQJExASEoJt27ahYsWKaNGiBYoVK4aQkJAiD3o3MDCAjo4OKlasKJ4oya8+53WRQUtLC/3798f69etx8eLFbHc9/FYKUg85ffdyqvOVK1di/vz5WLRoEZKSklC/fn0EBQXl+r6L+vs4YMAAvHr1Cv7+/vj06RNq1KiByZMnY8+ePYVq8f7111+xfv16zJ07FyNGjEDJkiXRvHlz8SJDUY6V8rttFaYevjZNTU2sXr0a06ZNw+jRo1GyZEl06dJF4Zj38uVLdOrUCZ6engp3nstLiRIlsGbNGsybN0+8kGlmZob169fn+dT3MWPGQFtbG2FhYVi5ciWqVKmCsWPHFviicU7q1q2LNm3aYPny5bh27RqWLVuGlStXYsGCBRg7dixSU1NhaGiI1atXf9UHI6uoqCA4OBjz58/HvHnzkJCQgOrVqyvccbQg+6ebmxsSEhIQGhqKNWvWoGTJknBxcRFvD12qVCn4+vpi/vz5GDhwIOrWrYu5c+fijz/+yLeM8+fPx+LFixEUFCSep4waNarIA9/linJemBNPT0/4+voiOjoaZ86cwT///IMePXoo1GFulISvfXscklRcXBxOnz4NR0dHhRaNzp07o1y5ckW6bzMRERER0bf0r2kJoaLR0NDA1KlTceDAAXTu3BnFihXD/v37cenSpWwPLCQiIiIi+jdgS8hP4MqVK1i0aBGuXbuG1NRUyGQyDB48ON+nlBIRERERfQ8MIUREREREJKl/zS16iYiIiIjov4EhhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaSKfe8CEP2ILC0tUaVKle9dDCIi+kE9efIEkZGR37sYRN8NQwhREVSpUgXh4eHfuxhERPSDateu3fcuAtF3xe5YREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFE/y8lJQWtWrXCmTNnvndRiIiIiH5qDCFEAD59+oSRI0fizp0737soRERERD89hhD6z7t79y5+//13PHr06HsXhYiIiOg/gSGE/vOio6NhY2ODsLCw710UIiIiov+EYt+7AETfW+fOnb93EYiIiIj+U9gSQvQDS0tO+d5F+NdgXfwP6+J/WBcZWA//w7og+ndgSwjRD6yYhhpWKzf53sX4V+idfuR7F+Ffg/vF/3C/yMB94n+4TxD9O7AlhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpDgwnSiTW7dufe8iEBEREf302BJCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphpAfnCAI37sIRERERESFUux7F+BHcvjwYWzcuBHXr19HcnIyqlevjo4dO+L333+Hqqqq5OW5c+cOpk+fjrVr137RduLj4zFq1ChERkaiZMmSOHbsGNTU1MT57u7uOHfuXJ7b8PHxQbt27b6oHERERET038AQUkBTp07Fpk2b4Obmhi5duqB48eI4d+4c5syZg4iICPj6+kJFRUXSMh04cABXr1794u3s3LkTJ0+exOzZs1G9enWFAAIA3t7eSEhIEP/u3bs3WrZsiY4dO4rTqlWr9sXlICIiIqL/BoaQAtixYwc2bNiAadOmoVOnTuJ0a2tr6OrqYsSIEdi9ezfc3Ny+XyG/wPv376GpqYm2bdvmOL9OnToKf6uoqKBixYowNjaWoHRERERE9LPhmJACCA4OhkwmUwggci1btkSfPn2gra0tTouNjcWwYcNgbW0NExMTDB48GA8ePBDn+/n5wcTERGE7N27cgEwmQ2RkJABg/PjxGDp0KEJCQuDo6AhDQ0O4u7vj3r174jb8/f2RmJgImUyG8PDwXMt/6NAhtG/fHsbGxrC3t4evry9SU1MBZHS18vPzQ1JSEmQyGfz8/IpUR0OGDEGrVq2yTXdxccGcOXMQGxsLmUyG/fv3o3v37jA0NETLli2xb98+heUTExMxffp0WFtbi+/5+vXrRSoTEREREf07MYTk4+XLl7h9+zbs7e1zXWbcuHHi/OfPn6Njx454+PAhvL294ePjg9jYWHTt2hUvXrwo1GufOXMGO3bsgJeXF+bNm4eHDx9i/PjxAICOHTuiQ4cO0NDQQFhYGBwcHHLcRlhYGDw9PWFgYAB/f390794dq1atwoQJEwBkdLXKvJ3MXawKo02bNrhz5w5u3bolTrty5QoePHiANm3aiNMmT56MevXqwd/fH3p6ehg5ciT+/vtvABmD7AcPHoy9e/di+PDhWLx4MdTU1ODu7o5Hjx4VqVxERERE9O/D7lj5eP78OQCgcuXKBVp+zZo1SE5OxqpVq6CjowMAsLCwgLOzM1avXi2GiIL4+PEjli9fjgoVKgAAXrx4gZkzZ+Lt27eoWLEiKlasCGVl5Vy7RaWnp8PX1xeurq6YMmUKAKBx48YoWbIkvL290a9fP9SrVy/f7RSEvb09dHR0sGfPHshkMgDA7t27oauri3r16iE2NhYAYGtri0mTJgEA7OzscP/+fSxfvhyNGzfG33//jYiICKxevRrW1tbi8q6urli6dCl8fHyKXD4iIiIi+vdgS0g+5IPN09PTC7R8VFQULC0txQACADo6OrCyssr3DlNZVa5cWQwgAFCxYkUAQFJSUoHWv3fvHuLi4tC8eXOF6fJuU9HR0YUqT15UVVXh6uqKvXv3AgA+f/6Mffv2KbSCAICrq6vC305OTrh48SLS09MRGRkJTU1NmJubIy0tDWlpaQAyglNERMRXKysRERERfV9sCclHpUqVAADPnj3LdZmXL1+iXLlyUFZWRnx8POrXr59tmbJly+Lu3buFem1NTU2Fv5WVMzJjQQPR+/fvxdfOTEtLC+rq6gp3vPoa3NzcsG7dOly8eBEJCQmIi4tD69atFZYpX768wt86OjpITU1FYmIi3r17h6SkJOjr62fb9ve4BTIRERERfRsMIfnQ0dFBgwYNcOrUKYwePTrHZXr37o1y5cohJCQEpUuXxuvXr7Mt8/r1a5QpUwYAoKSklC1IfPz48auXXf56b968UZgeHx+PT58+ifO/Fn19fdStWxd//fUXPn78iEaNGuGXX35RWObdu3cKf7958wbq6uooUaIESpYsibJly2L58uVftVxERERE9O/C7lgF0LNnT9y8eRNbtmzJNm/nzp24e/eueMXfzMwMkZGRiIuLE5eJi4vD2bNnYWpqCiCjJSI5ORnx8fHiMufPny90ueQtI7mpWbMmtLW1ceDAAYXp8jtSycvzNf322284cuQIjh8/nq0rFgAcO3ZM4e8jR47AwsICSkpKMDMzQ1xcHIoXLw4DAwPx3+7du7Fr166vXlYiIiIi+j7YElIAbdq0wfHjx/Hnn3/iypUraNKkCZSUlPD3339j48aNaNGiBdq3bw8A6NWrF7Zv344+ffrAw8MDgiBg6dKlUFNTQ8+ePQFkDLb28fGBl5cXunXrhps3b2LDhg2FLlepUqWQlJSEw4cPw9DQUGH8CJAxnsXT0xPTp09H6dKl0aRJE9y6dQt+fn5o3rw5dHV1v7xysmjTpg0WLVoEdXV1NG3aNNv8LVu2QEdHByYmJtixYwdu3bqF9evXAwAcHR1hYGCAAQMGwNPTE5UqVcLBgwcRGhqKqVOnfvWyEhEREdH3wRBSAEpKSli4cCE2b96M8PBwHDx4ECkpKahZsyYmTZqEDh06QElJCUDGGJLQ0FDMmzcP48aNg4qKCiwsLLBo0SJxYHnt2rUxY8YMLF26FP3794eRkRGWLFmC33//vVDlcnV1xY4dOzB8+HAMGzYM/fv3z7ZM9+7doaGhgVWrVmHLli2oUKECevfuDQ8Pjy+vmBz88ssvkMlkqFOnDkqUKJFt/vDhw3Ho0CGsXLkSurq6WLlypfjMFBUVFQQHB2P+/PmYN28eEhISUL16dfj4+KBdu3bfpLxEREREJD0lQRCE710I+nm8fPkSDg4OWLlypXibXSDjAY5NmjTB4sWLs92t60fUrl27PB8QKaXVyk2+dxH+FXqnH/neRfhX4X6RgfvF/3CfyPBv2Sf+Tb8jRN8DW0Loq3j06BF27dqFw4cPo3bt2rCysvreRSIiIiKifykOTKevQhAEhISEIDk5GfPmzRO7pxERERERZcWWEPoqqlevjqioqFznV61aFbdu3ZKwRERERET0b8WWECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYQQEREREZGkGEKIiIiIiEhSDCFERERERCQphhAiIiIiIpIUQwgREREREUmKIYSIiIiIiCTFEEJERERERJJiCCEiIiIiIkkxhBARERERkaQYQoiIiIiISFIMIUREREREJCmGECIiIiIikhRDCBERERERSYohhIiIiIiIJMUQQkREREREkmIIISIiIiIiSTGEEBERERGRpBhCiIiIiIhIUgwhREREREQkKYYQIiIiIiKSFEMIERERERFJiiGEiIiIiIgkxRBCRERERESSYgghIiIiIiJJMYTQf15KSgomT54Mc3Nz2NjYYMWKFd+7SEREREQ/tWLfuwBE39vcuXNx8eJFrF69Gs+fP8fYsWNRuXJluLq6fu+iEREREf2U2BJC/2mJiYnYvHkzJk6cCH19fTg7O6Nfv35Yv3799y4aERER0U+LIYT+027evImUlBSYmZmJ08zMzHD16lWkpaV9x5IRERER/bwYQug/7dWrVyhdujTU1dXFaeXKlUNqairi4uK+Y8mIiIiIfl4cE0L/aUlJSVBTU1OYJv87JSUl1/WePHmCdu3afdOyFZhb6e9dgn+F3f+Wz+PfgvsFAO4XCrhPAPj37BNPnjz53kUg+q4YQug/TV1dPVvYkP+tqamZ63qRkZHftFxEREREPzN2x6L/tF9++QXx8fEKQeTVq1dQU1ND6dK8akhERET0LTCE0H9a/fr1oaqqiosXL4rTzp8/Dz09PRQrxoZCIiIiom+BIYT+0zQ1NeHm5oapU6fiypUrOHLkCFatWoUePXp876IRERER/bSUBEEQvnchiL6npKQkTJkyBQcPHkSJEiXQp08f9OnT53sXi4iIiOinxRBCRERERESSYncsom+of//+GDt2rMK0EydOQCaTYebMmQrTN2/eDEtLSwiCACcnJ2zZsqVIr7llyxY4OTkVucwFERkZCZlMhg0bNmSbN378eIwePfqbvbaTkxNkMlm2f61atfrqr//48WMcP378q2zr3yA+Ph5z5sxBkyZNYGRkBBcXFwQFBSE1NfWLty0IAjZu3Ij09PQCLf8l+zh9mdzq/syZM5DJZN/kNe3s7BAeHg4AcHd3x6JFi77J6xDRj4Mjb4m+IXNzc2zbtk1hWkREBCpUqICIiAiF6ZcuXULDhg2hpKSErVu3onjx4lIWtUgWLVoEFxcXlC1bVtLXHT9+vBg65L7FjQQmTpwIU1NTODg4fPVtS+3du3fo1KkTypYtixkzZqBq1aq4fv06ZsyYgdu3b2P+/PlftP2oqChMmTIFHTt2hLJy/te3fpR9nIiIvg22hBB9Qw0bNsTDhw8RHx8vTouMjETfvn1x584dhaeyX7p0CRYWFgAAHR0daGhoSF7ewipZsiTmzJkj+etqaWmhfPnyCv+0tbUlL8ePZP78+VBVVcXq1athZWWFX3/9FS4uLliwYAF2796Ny5cvf9H2C9uz90fZx4mI6NtgCCH6hgwMDKCuro6rV68CyOgOc/PmTbRu3RrVqlUTW0M+fPiAmJgYWFpaAlDsLuHu7o6AgAD07dsXRkZGaN26NU6cOCG+xosXL9CvXz8YGxujXbt2iI2NVSjDvXv30LdvX5iamqJx48bw8/NDeno6rl+/jvr16+Pdu3cAgPfv36NevXpYs2aNuO6gQYOwdOnSXN+fl5cXdu7ciaioqFyXuXjxIrp06QJjY2M4OTkhNDRUnDd+/HjMmDEDI0eOhLGxMVxcXMQuG1/LsWPH0LZtWxgaGqJFixbYv3+/OC8hIQFeXl6wsrKCvr4+XFxc8Ndff4llO3fuHJYtWwZ3d/evWiappaSkYO/evejWrRvU1dUV5llYWCAkJAS6urp4//49Jk+eDGtra5iammLUqFHi/hEZGQk7OzuEhYXBzs4OlpaWGDNmDJKTkxEbGyveUU5PTw+RkZFITU3FnDlzYGdnBz09PTg6Oip03yvMPv7hwweMGzcOZmZmsLGxweTJk5GQkKBQrmnTpsHMzAx+fn7fsir/M/Kqc+B/3ysDAwOYmZlh+PDhCvM3bdoEe3t7mJmZYfny5bm+TkH2k7CwMLRv3x6Ghobo27cvnjx5Ak9PTxgZGcHNzQ337t0DAISHh+P333/HokWLYGpqCnt7e2zatOkb1A4RfQ0MIUTfkKqqKoyMjMSrzOfOnUPNmjVRtmxZWFhYiCHk0qVLKF26dK79sYOCguDq6orw8HDUrFkTXl5e+Pz5MwBg2LBhSE9Px5YtW9CvXz+sXbtWXC8uLg5du3ZFhQoVsGXLFkyZMgWhoaFYtWoV6tevDx0dHURHRwOAGCQuXLgAAEhLS8O5c+dgZ2eX6/uzt7eHs7Mzpk6dmuO4gnv37qFnz54wNzfH9u3bMWTIEMybN08hCGzatAn169dHeHg4GjdujClTpognvl/q7NmzGDJkCNq0aYOdO3eiU6dOGD16NK5cuQIA8PHxwb1797Bq1Srs2bMH5ubmmDx5MlJSUuDl5QUTExP07Nnzhz+xffToERITE2FgYJDj/EaNGkFTUxOenp64ceMGli1bhjVr1uD+/fsKY5revHmDffv2ISgoCDNnzsTBgwcRHh6OSpUqiXV08uRJmJiYYMWKFTh69CiWLFmCAwcOoG3btpgxYwZevHiRYxny2scnTpyIt2/fIjQ0FMuXL8f9+/cxYcIEcd0XL14gISEB27dvR9u2bb9Wtf2n5VXnjx8/xpAhQ9C5c2fs378fixcvRkREBDZu3AgAOHXqFGbOnIkRI0Zg06ZNuHTpUq6fe0H2kyVLlmDEiBEIDQ3FtWvX0LZtW9ja2mLLli1QVlaGr6+vuOz169dx7do1bNq0CUOHDsWMGTMUAi0R/XtwTAjRN2Zubi6e9EZERIitHZaWlvD39wcAXL58WRwPkhM7Ozu0a9cOAODh4YE2bdrgxYsX+PjxIy5evIgjR46gatWqqFu3Lq5evSpezd+zZw/U1dUxbdo0qKqqonbt2nj16hUWL16Mfv36wdraGpGRkXB2dkZUVBTs7OzEEHLp0iVoaGigQYMGeb6/SZMmoWXLllizZg369++vMG/z5s2QyWQYOXIkAKBmzZq4d+8eVq5ciRYtWgAAdHV1xfVGjBiB9evX486dOzA3N8/1NadNm4ZZs2YpTDt8+HC2sSmhoaFwdnZGr169xNe/fPkyVq5ciSVLlsDMzAw9evQQw1+fPn2wZcsWvHjxAr/++itUVVWhqamJMmXK5FkH/3by7oAlS5bMdZmbN/+vvTsPqqr+/zj+xMQN/IKgEkFCYZCKioiYCiJZiaLpOFimVhouOJA6paKmlY4bJkKRNqWRjqZiVppLGZpNI26jouKCyEWDsppSIk1S4fL7g+H8urKKeo16PWbujOd8zvI5n/O5eN7ns9xMDh48yPbt2/Hy8gJKu3D17duXs2fPAqWB6YwZM/Dx8eHRRx8lODiYjIwMhg0bhoODAwDOzs7Ur18fb29v5s2bh5+fH1DaqrZ06VLOnTuHi4tLufNXVseLiopITU1l//79xn2Ii4vj8ccf56effjL2Hz16NK1atbq9gvoPqeg7VBb05ebmVlnmxcXFvPbaazz77LMAuLu70717d7Kzs4HSyTHCw8MZNGgQAPPmzSMkJKTCfNSkngwcOJCgoCCgtOUuPz/fOPfTTz9dbpB9XFwczZs3x9vbm4MHD5KSklLp+UXk3lEQInKXBQQEGF0CDhw4QExMDFD6n+n58+e5ePEi6enpVbY4PPjgg8a/7e3tgdIHwuzsbOzt7XF3dzfSfX19jSDEZDLRtm1bbG1tjfROnTqRn5/PpUuXCAoKIjk5GShtCXnllVeIiooiNzeXvXv3EhwcXGlgVMbV1ZXx48ezbNmycoPFTSYTHTt2tFjXqVMniy5ZlV3boUOHLIKacePGERUVBUBMTAxhYWEWx60oUDCZTDzzzDPlzr9hwwYABg0axM6dO/nkk0/Iycnh5MmTADWe4amuKBsvU1BQUOk2OTk52NnZGQEIwMMPP4yDgwMmk8k4xt8f9O3t7SkqKqrweE888QRpaWksXLiQnJwcTp06BVRetpXVA5PJRElJCaGhoeX2OX/+vDEI3s3NrdJrk/Iq+g6lp6cTGxtbbZl369aNBg0a8N5773H27FnOnj1LdnY24eHhQOn3bsiQIcY+Tk5Old6fmtSTv9eNhg0b8sADD1gsX79+3WLb5s2bG8u+vr6sWbOmRmUiItalIETkLvPz86OgoICTJ0+SnZ1tvOF3cXHB09OTw4cPk5GRwZQpUyo9xt+DiDJlA4FvHhD891mibu7/D///n7vZbCYoKIjp06eTl5eHyWSiS5cutG/fniNHjrB3715GjBhRo2scNWoUmzdvZt68ecYDZFXnL3vjWtW1+fr6smnTJmNd2Zt2KH2o8fDwqDZf1Z1/6tSpHDlyhIEDB/Lcc8/RokUL4w3rv0mrVq1wdHQkIyODDh06lEufNGlSpW+Ki4uLLR4Ib75flQ1IT0hIMPryDxw4kDfeeKPKqaMrqwfFxcU0adLEoi6UadGihTHeqqJ7LZWr6Dv0448/AlRb5pmZmTz33HOEhobSuXNnRo4cyapVqyy2u7leVHR/oWb15OaZ76qafe3mbYuLi2s0W5uIWJ++mSJ3WePGjWnXrh3r1q3jkUcewcnJyUjr2rUrO3bswMbGplbz83t7e/Pnn3+Sk5NjrCt7kwjg5eXFqVOnLMZrpKen4+joiJOTE87Ozvj4+PDBBx/Qtm1bGjZsSEBAALt37+bEiRP06NGjRvmwtbXljTfeIDU1lYMHD1qc/+ZZl9LT03nooYeqPWajRo3w8PAwPrXpElXV+a9cucLWrVuJj49n4sSJPPnkk0ZLwb/tN1zvu+8+wsPDWbNmjcVbYyjtIvjll1/i5ubGn3/+aQzyBcjOzubKlSs1ul83t5itX7+emTNnMmXKFMLDwyksLARuvWwfeughrl69SnFxsVEXoHQ8z98HQsudU12Zb968GX9/f5YsWcLw4cPp0KED33//vXFvy7qFlrly5Qp5eXkVnutO1ZMyeXl5FvXixIkTd+23T0Tk9igIEbGCgIAAtm3bZowHKRMYGMiuXbuqHA9SFS8vLx577DFmzJhBZmYmO3fuNAaHAvTv3x+z2czrr7+OyWRi165dJCUlMXToUOPtYFBQEJ9//jkBAQFGXnfs2EHbtm1vadrbwMBAnn76aeNtKsCwYcPIyspiyZIlnDt3jk2bNrF27doat7DcrpEjR5KamsrKlSs5f/48K1euJDU11ZglqnHjxnz99df88MMP7Nmzhzlz5gAYD+p2dnbk5uZy8eJFq+T3boqJieHatWuMGjWK/fv3k5uby+eff86kSZMYPHgwgYGBhIaGEhsby/Hjxzl+/LgxO1KbNm2qPX7Zb36cOnWKa9eu4ejoyO7du8nLy+Pw4cPGAPebg6DqeHl5ERwczNSpUzl27BiZmZnExsZy8eJFWrZseesFIdWqrswdHR3Jysri2LFjnD9/noULF5KRkWG87Bg+fDhff/0169evx2QyMXPmTK5du1bhue5UPSlTWFho/L3bsGEDX331FcOHD69dQYjIXaUgRMQKunTpwtWrVysMQgoLC43fB6mNxMREmjdvztChQ0lISLCYTtbOzo4VK1aQl5fHoEGDmDNnDi+88AITJ040tgkKCuLGjRtGENK5c2fq1atHcHDwLedl2rRp/O9//zOW77//ft5//3327NnDgAEDWLZsGbGxsRb9xe+m9u3bs3jxYlJSUujfvz+ffvopiYmJ9OjRA1tbW9566y127txJv379mD9/PlFRUbi4uBitSc8++yxpaWnlBtzXRU5OTqxbtw4vLy9iY2Pp378/y5cvZ+zYsUbwtXDhQjw8PBg5ciSRkZE88sgjVU7R/Hfe3t4EBQUxbNgwvvvuO+bPn09WVhbh4eHExsYSFhaGn5+fRUtdTS1atAgPDw9eeuklRowYQcuWLVm2bNktH0dqrqoyf/755/H392fUqFEMHTrUmDL39OnTQOnfuwULFrB8+XIiIiJwcXHB29u7wvPcyXoC0LJlS9zc3IiIiGDFihUsWrSoykkuROTesSn5t/U7EBERkf+czz77jMTERL777rt7nRURqQG1hIiIiIiIiFUpCBEREREREatSdywREREREbEqtYSIiIiIiIhVKQgRERERERGrUhAiIiIiIiJWpSBERERERESsSkGIiIiIiIhYlYIQERERERGxKgUhIiIiIiJiVQpCRERERETEqhSEiIiIiIiIVSkIERERERERq1IQIiIiIiIiVqUgRERERERErEpBiIiIiIiIWFX9e50BERG5PRkZGaxevZpDhw7x66+/Ym9vT6dOnYiMjKRz587GdklJSSQnJ5Oenn5P8nngwAFeeOGFKrdxc3Pjm2++sVKORETkXlEQIiJSh23YsIHZs2fj7+/PhAkTcHNz47fffmPjxo08//zzJCQk0KdPn3udTQDatWtHSkqKsbx9+3ZWrVplsa5Bgwb3ImsiImJlCkJEROqozMxM5syZQ3h4OHFxcdjY2Bhpffv2ZeLEicyePZvQ0NB/xMO9vb09fn5+xvLRo0cBLNaJiMh/g8aEiIjUUStWrKBBgwbMmDHDIgApM2HCBAICAsjPz69w/xs3bvDOO+/Qp08ffH196dKlCzExMfz000/GNjk5OYwePZqAgAD8/f2JjIwkMzOzxum36syZM/j4+PDVV19ZrN+yZQu+vr7k5+czbdo0xo0bR3JyMt27dycgIIBXX32V33//3WKftLQ0hgwZQocOHejZsydvv/02xcXFtc6biIjcOQpCRETqqG+//ZZu3brh6OhYYbqXlxfvvPMOLi4uFaYvWLCANWvWMGbMGJKTk5k0aRL79u1j/vz5xjbR0dEUFxeTkJBAQkIC+fn5jBs3zniYry79Vvn4+NCmTRu2bdtmsX7Lli2EhITQrFkzAA4fPszatWuZNWsWM2fOZO/evYwfP97Yft++fYwZMwZ3d3feffddIiMj+eijj5g7d26t8iUiIneWumOJiNRBBQUFXL58mVatWlmsLykpKRcA3HfffRW2lFy6dImpU6cSEREBQGBgIOfOnWPLli1Gek5ODtHR0QQHBwPg6urK1q1buXr1Kjdu3KgyvWnTprW6tkGDBhEfH8/ly5dp2rQply5dIi0tjYSEBGObK1eusH79elq3bg2Ao6Mj48aN4+DBgwQGBpKYmEjHjh2NfXr27ImDgwPTp08nMjISd3f3WuVNRETuDLWEiIjUQWWBxs3Bxfbt22nXrp3FJzk5ucJjJCYmEhERwS+//MK+ffv4+OOPOXLkCNevXwdKH+w9PT2ZNWsWM2bMYMeOHbi5ufHKK6/QtGnTatNra8CAAZjNZlJTU41rsrOzo1evXsY2Pj4+RgACEBISgq2tLYcOHaKwsJDjx48TGhpKUVGR8enZsydms5kDBw7UOm8iInJnqCVERKQOcnJyokmTJly4cMFifVBQEBs3bjSWy1o5KnLkyBHefPNNzpw5Q9OmTWnTpg0NGzY00uvVq8fKlStJSkpi165dfPrppzRq1IjIyEhefvnlatMran2pCWdnZ4KDg9m2bRuDBw9my5YthIWFWQyub9GihcU+NjY2ODo6UlBQwB9//IHZbCY+Pp74+Phyx//1119rlS8REblzFISIiNRRISEhpKWlUVhYSOPGjQFwcHCgffv21e57+fJloqKi8Pf3JykpCQ8PDwAWLVpkMbDc1dWV+fPnYzabOXr0KJ988glLly6ldevW9OvXr9r02ho4cCCTJ08mKyuLo0ePMnXqVIv0mwehm81m8vPzcXZ2xs7ODoDx48fTu3fvcsdu2bJlrfMlIiJ3hrpjiYjUUWPHjqWwsJA5c+ZUOBA8Ozu70n1zcnIoKCjgxRdfNAIQs9nM3r17KSkpAUqnAA4KCuLkyZPUq1cPf39/5s6dS/369blw4UK16bejd+/eNGnShNmzZ+Pu7m7xo4tlefv555+N5W+//ZaioiK6du2Kvb09jz76KHl5ebRv39742NrasmTJEov9RETk3lBLiIhIHdW2bVvmzp3L66+/ztmzZxkyZAienp788ccf7N69my+++AJXV1e6dOlSbt+HH34YOzs7li1bhtls5q+//mLt2rVkZmZiY2NDSUkJrVu3xs7OjtjYWGJiYnBwcGDTpk3Y2NjQq1cvPD09q0y/HQ0aNKBv376kpKQQHR1dLr2oqIioqChiYmIoKChg8eLF9OrVi44dOwKl0xNHR0djb2/Pk08+SX5+PomJidSrVw9vb+/bypuIiNw+m5KyV14iIlInmUwmVq9eTVpaGr/88guNGjXCx8eHsLAwBg8ebHTVSkpKIjk5mfT0dKD0dzQWLVrEuXPnaNasGQEBATz11FNMmDCBlJQU/Pz8yM3NJS4ujsOHD3P16lV8fHyYNGkSPXr0AKg2vSorV65kwYIFnDlzpsL0nTt3Eh0dzY4dO/D09DTWT5s2jRMnTjBgwAA+/PBDbGxsGDBgAJMnT6ZRo0bGdt988w1Lly4lKysLe3t7unfvzuTJk3F1da1tUYuIyB2iIERERP6RygbNr1u3zmJ9WRCydevWe5QzERG5XeqOJSIi/ygbN27k9OnTbNiwgSVLltzr7IiIyF2gIERERP5RTpw4webNmxkxYgRhYWH3OjsiInIXqDuWiIiIiIhYlaboFRERERERq1IQIiIiIiIiVqUgRERERERErEpBiIiIiIiIWJWCEBERERERsSoFISIiIiIiYlX/B2O4GZ/TnkTCAAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "reference: Clear facsimile jar = 0.37 wt%\n", " Nikumaroro jar = 0.74 wt%\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAyUAAAH4CAYAAABUq0JHAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAACWiklEQVR4nOzdeVhO+f8/8GdFirIku7HrjnapVEqFsQ+SXbJTk0T2kD27aKGGEkJ2so19rEVm7ITs+5KkhUrn90e/+3zd7WE+x/B8XJfr0tnu1/2+z/Y67+UoCYIggIiIiIiISCLKUgdAREREREQ/NyYlREREREQkKSYlREREREQkKSYlREREREQkKSYlREREREQkKSYlRCQJDvxHheE+8m2xPInoe8akhOgncPjwYQwePBiWlpYwMTFBly5dEBERgYyMDEniuX37NlxcXL56O0lJSRg6dCgMDQ1hbW2N9PT0fJe9d+8eZsyYgdatW8PQ0BDm5uYYNGgQjh07prDc9u3bIZPJkJCQ8NXxfYnHjx9DJpMV+u97FRMTA5lMhitXrnzxNtLT0zF79mwcOXJEnObg4ICZM2d+ixDzlZycDHt7exw4cCDXvLVr14r7TpcuXXDw4MFcy2zduhUdOnSAiYkJfvvtN+zbt0+cJy+X/P49efKkyHF+yT76Jcfct/gtiYiKqoTUARDRv2vGjBnYtGkTunTpgt69e6N06dI4d+4c5s+fj+joaPj5+UFFReV/GtOBAwe+yY3Orl27cOLECcybNw+1a9eGqqpqnssdO3YMY8aMQe3atTFkyBDUrVsXSUlJ2LdvH0aMGIFJkyZhwIABXx3Pt1C5cmVERkaKf58/fx6LFi1CQEAAKlWqJGFkRaOnp4fIyEjUr1//i7fx8uVLrFu3Dk2bNv2GkRUsOTkZbm5uePr0aa55K1euxNKlS9GnTx+0atUKly5dwpgxY+Dr64tOnToBAPbu3Qtvb28MGDAALVq0wLFjxzB69Gioq6vD3t5eLJfPffz4ER4eHtDT00O1atX+1e/3rY45IqJ/C5MSoh/Yzp07sWHDBsycORM9e/YUp1tZWUFHRwejR49GVFQUunTpIl2QX+Hdu3dQV1dH165d813mxYsXGD9+PAwNDfHHH38oJC6tWrVClSpVsHjxYnTs2BHa2tr/i7ALpKqqCmNjY/Hv169fAwAaNWqEmjVrShRV0WloaCjE/19w7tw5+Pj44M2bN7nmffr0CatWrUKHDh3g4+MDALC2tkZaWhrmz5+P9u3bQ0VFBTt27EDTpk0xadIkANnH2JUrV7Bp0ybY29vnWS5z5syBkpISFi1aBGVlNlwgop8bz4JEP7DVq1dDJpMpJCRy7du3x6BBg1ChQgVx2uPHjzFq1ChYWVnBxMQErq6uuH//vjjf398fJiYmCtu5ceMGZDIZYmJiAAATJ06Eh4cHwsPDYW9vD0NDQzg7OyM+Pl7cRkBAAFJTUyGTybB9+/Z84z906BC6desGY2NjtGjRAn5+fmKTM2dnZ/j7+yMtLQ0ymQz+/v55bmPDhg14//49pk+fnmdNyogRI2BtbY23b9/mub4gCAgPD0enTp1gYGAAExMTDBw4EHFxceIyr169wqhRo2BhYQEjIyP06dMH586dK/L84nr79i309fWxevVqhel///03ZDIZ4uLi4O/vj44dO2LXrl2wt7eHiYkJhg8fjsePHyusc/XqVbi4uMDIyAjNmjXDrFmzkJaW9sWx52zy4+zsDF9fXyxduhTW1tYwMjKCm5sbXrx4kef6jx8/RsuWLQEAo0aNgrOzszjvw4cPmD59OszNzWFqaooJEyYgOTlZYf21a9fi119/hb6+Pjp06KDQhCo/v//+O3R0dLBq1apc8968eYP379+jefPmCtNNTU3x6tUr3Lx5E0B2kzMNDQ2FZcqXL493797l+Zl37txBREQEPD09oaWlVWB8O3fuRJs2bWBoaIihQ4ciMTFRYX5h+2h+x9zdu3fh4eGBZs2aQV9fHw4ODggMDMzV9+TatWtwdHSEgYEBHB0dcfr0aYX5ly9fxtChQ9G0aVPo6+ujTZs22LRpk8Iyq1atQuvWrWFgYIBWrVohMDAQWVlZ4vw3b95g/PjxMDc3h4mJCUaMGIFHjx4VWC5E9GNhUkL0g3r58iVu3bqFFi1a5LvMhAkTxPnPnz9H9+7d8eDBA/j4+MDX1xePHz9Gnz598r2BzM+ZM2ewc+dOeHt7Y+HChXjw4AEmTpwIAOjevTucnJygpqaGyMhI2NnZ5bmNyMhIuLu7w8DAAAEBAejXrx9CQ0PFJ9E+Pj4K2+nevXue2zl27Bh0dXVRt27dPOeXK1cOK1euRMOGDfOcHxoaikWLFsHJyQmrV6/G1KlTcefOHTEOAJg0aRIePnwIX19fBAUFQV1dHcOHDxdvHgubX1wVKlRAixYtsHfvXoXpUVFRaNSokdjn5MmTJ1iwYAHc3d3h6+uLu3fvYsCAAWLfmzt37qBfv35QUlKCn58fxo4di3379sHT07PI360otm3bhkuXLmHu3LmYPn06YmJi4Ovrm+eylStXRkBAAABgzJgxYu0EAOzYsQPv3r2Dn58fRo4ciaioKHFZAAgICBBrL1auXAkrKyuMGTMG+/fvLzC+iIgILFu2LM/koGLFilBVVc3VrEue3Mn7gvTp0wenT5/G/v378f79e+zbtw8nTpxAhw4d8vzMpUuXok6dOujRo0eBse3fvx8TJkyAtbU1AgMD8csvv2DJkiUKyxS2j+Z1zKWkpKB///5ITEzE/PnzERwcDAsLCyxfvjxXP6u5c+eiVatWCAgIgLa2NoYPHy4+ZHj69Cn69++P0qVLY9myZQgMDETdunXh4+MjJmz79u3DsmXLMGDAAKxevRrdu3eHv78/Nm/eDCA72ezfvz8uXLiAKVOmYMGCBXj9+jX69euXb1JHRD8eNt8i+kE9f/4cAFC9evUiLb9mzRp8+PABoaGh4s2Zubk5WrVqhbCwMDGpKIqUlBQEBwejcuXKALKbUM2ZMwdv375F1apVUbVqVSgrK+fbzCcrKwt+fn7o0KEDpk+fDgBo3rw5NDU14ePjgyFDhkBXV7fQ7QDZN43W1ta5pmdmZir8raysnGcTmmfPnsHNzU3sJGxubo6kpCT4+voiJSUFZcqUQWxsLFxdXeHg4AAAaNiwIcLCwpCWloby5csXOv9LdO3aFb///jvu3r2LevXqITMzE/v378fw4cPFZVJTU7Fs2TLY2toCAOrVq4fffvsNe/fuRdeuXREUFISKFSsiJCRErEWqU6cO+vbti/Pnz8PMzOybxK6iooLg4GCUKlUKAHDz5k3xhjQnVVVVNGrUCABQu3ZtNGjQQJxXt25dLFmyBEpKSrCyskJ0dLRYQ5eUlISQkBAMGTJETKqaN2+OlJQULF68GO3atcs3Ph0dnQJj79ChA8LCwqCrqwsLCwtcvXpVrKVKTU0FkN0UsEuXLgoJXffu3RVqeuQePXqEo0ePYubMmYU22woJCYGNjQ2mTZsGALCxscHTp08VEofC9tG8jrmrV6+iVq1a8PPzE493S0tLHD58GOfPnxd/bwAYNGgQ3NzcxGVat26N1atXY+7cubh9+zaMjY2xaNEilCxZEgBgZGQECwsLxMbGQldXF+fPn0eNGjXQp08fKCkpwdzcHCVKlBDPDzt37sS9e/cQFRUl9kWytLSEvb091q1bB3d39wLLiIh+DExKiH5Q8s7rnzeRKMj58+dhYWGh8LRYS0sLlpaWxW5qVL16dfGGAwCqVq0KAEhLS1NoLpaf+Ph4JCQkoG3btgrTO3bsCB8fH/FmpyiysrKgpKSkMO3ixYu5mrT17dtXvPH73JQpUwAACQkJuHv3Lu7evYujR48CyG6yU6ZMGZiYmGD58uWIi4tDixYt0KJFC0yYMEHcRmHzv0SLFi1Qvnx57N27FyNHjsTJkyeRlJSk8GReU1NTTEiA7JvvX375BbGxsejatStiYmLQsmVLKCsri0masbExNDQ0cPbsWZiZmX2T2GUymZiQANn7w+dNxIrKyMhI4besWbMmbt++DSD7N/348SPs7OwUEk5bW1ts27YNjx49wi+//FLszwSAyZMnIzU1Fb///rsYv7u7O6ZMmQJ1dXUA2c0WDx06BC8vLxgbG+Py5csIDAyEpqZmrvLasmULypYti86dOxf4uWlpabhx4wYmT56sML1NmzYKSUlR9tGc9PX1sWHDBmRkZODOnTu4f/8+rl+/jszMzFyj2LVp00b8v6qqKpo3b47Lly8DgLhPfPz4ETdv3sT9+/fFpnvy7ZiYmGDDhg3o1q0b2rZtCzs7OwwePFjcZkxMDGrXro3atWuLv52amhpMTU0RHR3NpIToJ8GkhOgHJR/N59mzZ/ku8/LlS2hra0NZWRlJSUniE+rPVaxYEXfu3CnWZ8tv1OTkT4OLmiDJm2xUrFhRYbqGhgZKlSqVqx9BQapXr56r6Y2Ojg62bt0q/u3q6prv+vHx8Zg6dSouXLgAdXV16Orqijd58rb3S5cuRWBgIPbv34+9e/eiZMmScHR0xJQpU6Cqqlro/C9RsmRJdOzYEXv27BGbMllaWiokg3mN1qWlpSWWb2JiIiIjI3ONCgVk9yUpyncripz7g5KS0he9M6Og7cibk/Xq1SvPdV+9evXFSUnZsmWxfPlyJCYm4s2bN6hduzb+/vtvANnN/549e4aoqCiMGzcOQ4YMAZBdW1GmTBnMmDEDvXv3Rq1atcTtHT58GK1atSq0/JKSkiAIQq5EPueADEXZR/OycuVKrFq1Cu/fv0eNGjVgYmKCEiVK5Fon53GopaWFly9fAsgeCGDevHmIjIxERkYGatWqJY6aJt/Ob7/9hk+fPiEiIgJLlizB4sWLoauriyVLlqB+/fpITEzE3bt3oaenlyvGOnXqFFhGRPTjYFJC9IPS0tJC48aNcfLkSYwdOzbPZQYOHAhtbW2Eh4ejXLly4khPn3v9+rXYTEdJSSlXYpGSkvLNY5d/Xs7RkJKSkvDx48diNXmys7NDaGgoXrx4gSpVqgAASpcuDQMDA3GZ/G4Os7Ky4OrqivLlyyMqKgoNGjSAsrIyIiIicOrUKYV4vb294e3tjRs3bmD37t0ICwtDzZo1MWzYsELnf6nOnTtj/fr1uHLlCo4dO4YZM2YozM+r38ebN2/EPicaGhpo2bIlevfunWs5+Y3wvxX7t6apqQkACAwMFH/nz+XXp6gojhw5gsqVK8PAwEDc9+Li4qCkpASZTCYOBpGzGaGpqSkEQUB8fLyYlDx9+hTx8fFFqm0qV64clJSUch0Hn/+uRd1Hc9q5cyf8/Pzg4+ODjh07iuVnaWmZa9mkpCSFZPfzc8KKFSuwefNmzJ8/Hy1atEDp0qWRlpamkPQD2c0Nu3btijdv3uDo0aMIDAyEu7s79u/fD01NTejq6mL27Nm5PvtLk3Yi+u9hR3eiH5iLiwtu3ryJLVu25Jq3a9cu3LlzR3zPgqmpKWJiYhReyJaQkICzZ8+iSZMmALJvYj98+ICkpCRxmQsXLhQ7rsLa0detWxcVKlTI9RI7+UhK8niKwtnZGRoaGpg8eTI+fPiQa/6LFy/yrXlJSEjAgwcP0KNHD+jo6Ihxnzx5UmEZOzs7HDp0CED20L0TJkxA9erV8ezZs0Lnfw1DQ0PUr18f8+bNA5DdryFn/PJmNkD2jfSjR4/QrFkzANm/+d27d6Gvrw8DAwMYGBigWrVqWLx4MW7fvv2vxl6QL3lvjpGREUqWLIk3b96I38XAwAC3b99GYGDgV8UTERGhsI309HRs2bIFJiYmqFChAmrVqgVlZWWx9kTu0qVLAKAwlLP89zA0NCz0c9XU1GBsbIzDhw8rTP/rr7/E/xdlHwVyH3P//PMPqlatit69e4sJybVr15CQkJCrpuTzbX348AEnTpyAubk5gOxmc/r6+mjXrh1Kly6tsLx8O97e3vDw8ACQXesi73gv34eaNGmCx48fo0aNGuLvpq+vjzVr1uD48eOFlhMR/RhYU0L0A+vcuTOOHz+OadOm4fLly2jZsiWUlJRw6tQpbNy4Ee3atUO3bt0AAAMGDMCOHTvETq2CIGDFihVQVVUVO9Da2NjA19cX3t7e6Nu3L27evIkNGzYUO66yZcsiLS0Nhw8fhqGhocJTWCD7ptTd3R2zZs1CuXLl0LJlS3GY27Zt2xbYMTmnqlWrYunSpRg9erT4AkkdHR18+PABZ86cwdatW6GqqgobG5tc62pra6N69eoIDw8Xm7nt3LlTvFFKS0tDjRo1ULt2bcyePRspKSmoVq0ajh8/jidPnqB169bQ0tIqcP7X6tKlCxYvXowuXbqIN4VySkpK8PT0FGvKli5dikaNGuHXX38FALi5uaFXr14YNWoUunXrhvT0dAQFBeHZs2do3Ljxvx57fuQ3yWfOnEGdOnWK1H9IS0sLzs7OmDdvHt69ewdDQ0PcvHkTS5cuRcuWLXMN11scvXv3hru7O1auXAkjIyOEh4fj7t27WLNmDYDsG+3evXuLI4EZGRnh+vXr8Pf3R5s2bRRGdrt9+zYqVKhQpL5VADBy5EgMGTIEkyZNQvv27REdHa2QpBRlHwVyH3MGBgbYtGkTAgICYG5ujvj4eAQGBkJJSSlX8i4fpKBGjRoIDQ1FWloahg4dCgAwMDDAH3/8gfXr10NHRwdXrlzJtR0zMzNMmDABS5YsgZWVFZ4/f46NGzeK+5CTkxPWrVuHQYMGiTWLkZGROHjwIH777bdi/lpE9J8lENEP7dOnT8LGjRuF7t27C+bm5oKxsbHQtWtXYePGjUJGRobCsrdu3RKGDh0qGBsbC6ampoKrq6tw9+5dhWU2b94s2NvbC/r6+kLfvn2FixcvCjo6OkJ0dLQgCIIwYcIEoUOHDgrrHDp0SNDR0REePXokCIIgvHnzRnB0dBT09PSEkJCQfGPfsmWL0K5dO0FPT0+wt7cX/Pz8hPT0dHH+8uXLBWNj4yKVw9OnT4X58+cLbdu2FYyMjAQTExOhe/fuQnBwsJCYmCgut23bNkFHR0d48+aNIAiCcOXKFaFnz56CkZGRYG1tLYwYMUI4c+aMoKOjI+zZs0cQBEF4/fq1MH78eMHa2lrQ09MTOnbsKERFRYnbLGx+QXKWXU43btwQdHR0hNOnTytMl5fNtm3bBGtra8HU1FQYO3askJCQoLBcbGys0K9fP8HQ0FAwMzMThg8fLty6deuLY4+OjhZ0dHSEy5cvC4IgCP369ROGDRumsExYWJigo6NT4PdeuHChYGRkJHTs2FEQBEGwt7cXZsyYobDM7NmzBXt7e/HvT58+CSEhIUKrVq3EfWbx4sXCx48fC/wsuUePHgk6OjrC/v37c83buHGj0Lp1a8HY2Fjo2bOncPbsWYX5mZmZQlBQkHhstGnTRggKCsr12T4+PkLr1q2LFI/coUOHhI4dOwr6+vpC7969hfXr1xd7H815zH369ElYsGCBYG1tLRgZGQnt27cXQkNDBW9vb6Ft27aCIPzfb3ngwAGhQ4cOgr6+vtCzZ0/hypUrYmypqamCt7e3YGFhIRgbGwtdunQRtm/fLgwaNEgYNGiQuNy6deuEtm3bCgYGBoKVlZUwY8YMITk5WZz/7NkzYfTo0YKZmZlgZGQkODk5CUePHi1WORHRf5uSIHxBb0MiIvou/PHHH4iIiMDRo0cVmuj4+/sjNDQU//zzj4TRERERFQ2bbxER/Qf9+eefuHz5MiIiIjBy5MhC++kQERF9z3gVIyL6D3rw4AHWr18PBwcH9O/fX+pwiIiIvgqbbxERERERkaRYU0JERERERJJiUkJERcaKVSIiIvo3MCkhojzJZDKsXr0aQPYbnb28vHDt2rU859OX8/X1RdOmTdGkSZMvehHlz+Tx48eQyWS5Xqr5rf2X9u3Tp0+jdevWMDAwwKxZs3LNl5fZ5/8MDAzQpk0bhISEfJMHDTExMZDJZLhy5cpXb4uIfl4cfYuI8hQZGYnq1asDAG7cuIE9e/ZgwIAB0gb1g4mLi8OaNWvg4uKC1q1bo1GjRlKHRP8xixcvhpqaGv744w9Uq1Yt3+XGjBkDCwsLANkvVLx48SKWLVsGABg2bNhXxaCnp4fIyEjUr1//q7ZDRD83JiVElCdjY2OpQ/jhvXv3DgDQsWNHGBoaShwN/RclJiaiRYsWaNasWYHL1a5dW+GYtrS0xMOHD7Fp06avTko0NDR4viCir8bmW0Q/gcOHD0Mmk+Hx48fitDlz5kAmk+HRo0fitJkzZ8LJyQnA/zVhiYmJEYecdXJywsSJE8XlExMTMWbMGJiYmMDCwgJz585FRkZGgbHs3r0b3bp1g5GREYyMjNCrVy+cP38+3+XlzU/Cw8Ph4OAAa2tr/P333xAEAeHh4ejUqRMMDAxgYmKCgQMHIi4uTlzX2dkZvr6+WLp0KaytrWFkZAQ3Nze8ePFCXCYrKwsBAQGwtbWFkZERRo4ciTVr1kAmkynEsWfPHvGzWrVqhXXr1hX4PQHg5s2bGDJkCMzNzWFubo5x48bh9evXALJfbujs7AwA6N69u/j/nCZOnAgPDw+sXr0atra2MDY2hoeHB5KTkxEQEAArKytYWFhg9uzZyMrKEtd78+YNxo8fD3Nzc5iYmGDEiBEKvzUAnDx5Ev369YOJiQkMDAzQuXNnHDx4UJz/6dMnLFiwAHZ2dtDX10f79u2xceNGcb6/vz9MTEwUtnnjxg3IZDLExMSI8bu5ucHLywtNmjTB6NGjAQCpqamYNWsWrKysYGhoCGdnZ1y/fl1hW5cuXULv3r1hZGSETp065ZqfV1m1adMm13RHR0eMHz8eAJCcnIzZs2fD3t4e+vr6aNasGSZMmICkpKQ8t7l9+3bIZDIkJCSI05KSkiCTybB9+3Zx2oMHD+Dm5gYTExM0bdoU48aNU1gnNTUV3t7eaN68OQwNDdG1a1eFss5LSkoK5s+fDwcHBxgaGsLJyQmnTp0C8H/HxZMnT7Bhw4Zcx3dRaGpq5ppW2PGZ1++Zs/mWs7Mzhg8frrDdnMeUg4MDQkJCMGXKFJiamsLCwgLLly/H+/fvMXbsWJiYmMDe3l6hjInox8akhOgn0KxZM5QsWRLR0dHitHPnzgGAQj+G06dPw9bWVmFdPT09TJs2DUB2/wc3Nzdx3qpVq1ChQgUEBQWhV69eCA8Px6ZNm/KN48CBAxg/fjzs7OwQEhICX19fJCUlYfTo0UhPTy/wOyxbtgxjx47FuHHjoK+vj9DQUCxatAhOTk5YvXo1pk6dijt37mDSpEkK623btg2XLl3C3LlzMX36dMTExMDX11ec7+fnh5UrV6JPnz5Yvnw5gOwmMZ/bsWMHvLy8YGZmhhUrVqBLly7w9fXFqlWr8o33xo0b6NmzJzIyMjBv3jxMnjwZsbGx6NevH1JTU9G9e3eFcvXx8cl3W6dOncKhQ4cwa9YsjBs3DocOHUK3bt1w6dIlzJs3D127dsW6deuwb98+AMCHDx/Qv39/XLhwAVOmTMGCBQvw+vVr9OvXT6yduXz5MoYNG4aGDRsiKCgIS5cuhbq6Ory8vMSb6dWrV2Pbtm3w9PTE6tWrYWNjg+nTp+PkyZMF/lY5/fXXX/j48SMCAwPRs2dPCIIAV1dX7N27F56enli2bBlUVVXh7OyMhw8fAsi+6R4wYABKlSqF5cuXo1u3brl+25w6duyI+/fv4+bNm+K0R48e4dq1a+jQoQMAwMvLC0ePHoWXlxdWr16NQYMGYc+ePQgKCirWd/rc69ev0adPHzx9+hQLFizAjBkzcPHiRQwePFjcr+fPn4/o6Gh4e3sjODgY9evXx6hRoxAfH5/nNrOysjBkyBBs374dw4YNg7+/P6pXr45hw4bh5MmTqFy5MiIjI1GpUiW0adMGkZGRqFy5cr4xZmVlITMzE5mZmUhNTcXp06exa9cu9O7dW1ymqMdnzt/zS61cuRKfPn1CQEAA2rVrh8DAQDg5OaFSpUpYtmwZ6tWrh2nTpuHp06df/BlE9N/B5ltEPwENDQ2YmJggJiYGTk5OePfuHW7duoXGjRsjNjYWXbp0wZMnT3D//n20aNEi17oNGjQAADRs2BC1atUS51lZWWHq1KkAspuDHD16FDExMfk+9X/48CH69u2LkSNHitNKliwJd3d33L9/Hzo6Ovl+hy5duqB9+/bi38+ePYObmxtcXFwAAObm5khKSoKvry9SUlJQpkwZAICKigqCg4NRqlQpANm1F5s3bwaQ/dQ8LCwMw4cPx4gRIwAAtra26Ny5s1jjkpWVhSVLlqBTp05iEtG8eXMoKSkhKCgIffr0QenSpXPFGxQUBC0tLfzxxx9QVVUFAOjr66NTp07Ytm0bnJ2dFcpV/v+8pKamYvny5eJN5+7du3Hnzh1s27YNGhoasLW1xf79+3Hp0iV07NgRO3fuxL179xAVFSW287e0tIS9vT3WrVsHd3d33L59G61bt1ZIhqpXr46uXbvi0qVLsLe3R2xsLPT19dGlSxcAgIWFBdTU1KCurp5vrHnJzMzEzJkzoaWlBSC7hiY6OhphYWGwsrICANjY2KBDhw5YsWIFfH19sW7dOqiqqmLFihVQV1dHixYtIAgC5s2bl+/nWFpaQltbGwcOHICuri4AYP/+/ahQoQKsra3x8eNHZGRkYPr06WLybWFhgX/++UdM0r9EeHg4Pn78iNDQUPE7Ghoaok2bNti3bx+6dOmC2NhYWFtbo127dgAAU1NTaGtrIzMzM89tHj9+HH///TdWrVoFGxsbAECLFi3Qs2dPLF26FNu3b4exsTFUVVWhra1daPMpeQ3V54yNjdG3b1/x76Ienzl/T3mtWHFVqVIFc+fOhZKSEkxMTBAZGYkqVapgwoQJAIA6deqgdevWuH79uti/jYh+XExKiH4SNjY2WL9+PQDg/PnzqFy5Mjp06IBt27YByH4aX6FCBRgYGBR5mzmb7tSoUSPfZjDA/3WoTUpKwt27d3Hv3j0cPXoUAAqtKcnZiXbKlCkAgISEBNy9exd3795V2JY8KZHJZGJCAgBVq1ZFWloagOzmQenp6WjVqpU4X0lJCb/++quYlNy7dw8vX76EnZ2dwg2kra0tli9fjsuXL+fZnv/8+fPo2LGjmJAAQIMGDSCTyXD+/Pl8E7e8VKtWTeEpeMWKFfHp0ydoaGiI08qXL4/3798DyL5JrF27NmrXri3GrKamBlNTU0RHR8Pd3R3dunVDt27dkJqaivj4eNy/f1+sSZP/FiYmJvDz84OzszNatWoFBweHPG9uC6OlpSXewMrjU1dXh5mZmUKZNm/eXPwN//77b5iZmSkkQL/++muBSYmKigratWuHAwcOwNPTE0B2UtK2bVuUKFECJUqUQGhoKIDsmpj79+/j9u3biI+PV9hHiismJgbGxsYoW7as+H2qVauG+vXr4+zZs+jSpQtMTEywefNmvHz5Evb29rCzs1NoCpnT+fPnUaZMGTEhkWvfvj3mzp2L5ORkhd+/MGPHjhX3048fP+LmzZvw9/eHi4sLNmzYgJIlSxb5+Mz5e34pQ0NDKCkpAcjeP8uUKQN9fX1xfvny5cV4iOjHx6SE6Cdha2uLxYsX4969e4iJiUHTpk1hamqKhQsXIiEhAadPn4aNjQ2UlYveqjPnE3NlZeUChxh99eoVvL29ceLECZQsWRINGzZEjRo1ABT+DpSKFSsq/B0fH4+pU6fiwoULUFdXh66urpiIfL6tnDEqKSmJ89++fQsAuW6wtLW1xf8nJiYCyG724+Xlled3yktSUlKumOXfIzk5Oc918iP/Xp8rqLYiMTERd+/ehZ6eXq55derUAZBd+zJt2jTs378fAFC3bl2xdkFePsOGDYO6ujq2bt2KuXPnYu7cuTA3N8eiRYtQpUqVIsefsxwSExORlpamcAMqV7JkSQDZ5SePR65SpUqFflbHjh2xbt06xMXFQV1dHdevX4e3t7c4/8iRI/D19cWjR49QoUIF6OvrQ01NTaE/TnElJibi0qVLeZa3POYpU6agcuXK2LVrF44dOwZlZWW0bt0ac+fOzTO5SEpKUtgP5eTTUlJSipWU/PLLLwoPHJo2bYoKFSpgzJgxOHLkCNq2bVvk4zOv/fpLFHe/JqIfG5MSop+Erq4uKleujJiYGMTGxqJHjx7Q19eHuro6zp07h+joaLEp1r/Fy8sLL168QGRkJPT09FCiRAn89ddfhXb4zSkrKwuurq4oX748oqKi0KBBAygrKyMiIkLsCFwU8tqHhIQEhZvszzsoyzsDT5s2Lc8RsmrWrJnntsuVK4c3b97kmv769et/fehUTU1N6OrqYvbs2bnmyWtuZs2ahdOnTyMkJARmZmZQVVXFnTt3EBUVJS6roqKCAQMGYMCAAXj69CkOHz4Mf39/eHt7Y9WqVVBSUsp1M5+SklKk+CpWrIjg4OB8lylfvnyu8pMnkQUxNjbGL7/8goMHD0JVVRXVqlWDqakpAOD+/fsYNWoUunbtivXr16Nq1aoAUGDfDvmT/M9vylNTUxWWkTeh8/DwyLW+/MZbTU0NHh4e8PDwwN27d/Hnn38iKCgICxcuxIwZM3KtV65cOXFQhM/Jk2B5LcLXkHc8l/fj+VbHJ4Bc+0XOMiMiyokd3Yl+IjY2Njh27Bji4uJgZmaGkiVLwtjYGGvWrMH79+/RvHnzPNdTUVH5Jp9/8eJFtG/fHkZGRihRIvuZiLzTdHFe4paQkIAHDx6gR48e0NHREWt3itsBu1GjRihTpgyOHDmiMF3eZAUA6tWrh/Lly+PFixcwMDAQ/yUmJmLZsmX51nqYmpriyJEjCs1e4uPjcevWLTRp0qRYcRZXkyZN8PjxY9SoUUOMV19fH2vWrMHx48cBZP8WNjY2sLa2FhOVnL/FoEGDxEEBqlevjv79+6NVq1Z49uwZgOyb8Q8fPig0rynKCyBNTU2RkJCA0qVLK5RpVFQUdu/eDSC7r0dMTIzCtk+cOFGk79+hQwccP34cBw8eRPv27cXE4vr168jIyMCwYcPEhCQ1NRUXLlzId/+T10a8fPlSnBYbG5vr+9y9e1d8MaGBgQF0dHQQEBCACxcu4NOnT+jYsSPWrFkDIHufcnV1hbGxsViWeZVRSkpKrn16//790NPT+6rmZnLy0bLk/cS+1fGpoaGhUF5A0fYLIvq5saaE6CdiY2MDT09PVKhQQXxa37RpU3Fo1woVKuS5nry24K+//kLp0qW/+Em/gYEBduzYAZlMhnLlyuHQoUPiELMfPnwo8na0tbVRvXp1hIeHQ1tbG8rKyti5c6d4wy3vM1IYTU1NuLi4IDg4GKqqqmjUqBF27dqFa9euiTeyJUqUwMiRI8W+DJaWlnj8+DEWL16MOnXq5FtTMmLECPTq1QtDhw7FgAED8P79e/j5+aFGjRpix/F/i5OTE9atW4dBgwZh2LBhKF++PCIjI3Hw4EH89ttvALJ/i6NHj2LHjh2oVq0aoqOjxbeYy38LU1NTrFixApUqVYKBgQHi4+Nx4MABcXABGxsb+Pr6wtvbG3379sXNmzexYcOGQuOzt7eHgYEBhg0bBnd3d1SrVg0HDx5ERESEWGvg4uKCyMhIDB06FCNGjMDz588REBBQpO/fqVMnsRbm87ecN2rUCCoqKli4cCF69+6Nt2/fIjQ0FK9fv1bo+/M5CwsLlCpVCnPmzIGrqyuePn2KFStWKCw/cOBA7Nq1C0OGDEH//v1RsmRJhIaG4uLFi/D09ISKigoMDQ0RGBiIUqVKoV69erh06RIuXLiQZy0JANjZ2cHIyAjjxo3D6NGjUa1aNWzfvh2XLl3CypUri1QOn3vw4AEuXrwIIDvBuHXrFpYsWYI6derAwcEBwLc7Pm1tbTF9+nT4+/vDzMwMf/75J65evVrsmIno58KaEqKfiLW1NVRUVNC0aVPxptvc3BwAcg0F/LmGDRuic+fOCA4OxsKFC7/48319fVG/fn1MmjQJo0ePRnx8PNatW4fSpUuLN0xF5e/vjzJlysDT0xOTJ09GWloawsLCAKBY23J3d8fAgQMRHh4Od3d3ZGRk5BpRq1+/fpg+fTqOHj2KoUOHYtmyZWjbti2Cg4PFcsxJX18f4eHhyMzMxKhRozBnzhw0bdoUGzduLFZfgC+hoaGBiIgI1KtXD9OnT4ebmxuePn2KoKAgcXS1iRMnwsrKCnPnzsXIkSMRHR2NgIAA1KlTB//88w+A7MRq+PDh2LhxIwYPHoyQkBC4uLjA3d0dQPbgA7Nnz8a1a9cwdOhQHD58WBxWuSAqKipYvXo1rK2tsXDhQgwbNgznz5+Hr68vevXqBSC738L69euhrq4OT09PrFmzJt8b+JwaNGgAHR0d1KlTB40bNxan161bF/Pnz0dcXByGDRuGRYsWQV9fHz4+Pnj27JnC+2vkypYtCz8/PyQkJGD48OHYsGEDFixYoLB/VK9eHRs2bIC6urqYRGRlZSEsLAyNGjUCkN2npHPnzli5ciUGDx6Mbdu2YcKECejevXu+ZbRq1Sr8+uuvWLp0KUaOHInnz58jJCQEdnZ2RSqHzy1ZsgQ9e/ZEz5490bdvX/j7+6Nt27ZYu3atmGB9q+Oze/fucHFxwfr16+Hq6ork5GRMnjy52DET0c9FSShOnSwR0Q8kPT0d+/btQ/PmzRU6FXt5eeHu3bvYsWOHhNERERH9PNh8i4h+WqqqqggKCsKWLVswZMgQqKur4+zZs9i3b1+encSJiIjo38GaEiL6qd27dw+LFi3ChQsXkJqairp162LAgAHo2rWr1KERERH9NJiUEBERERGRpNjRnYh+anwuQ98D7odE9LNjUkJE/znyIYy/1u3bt8Xhbf+Lbt26BRcXF5iYmMDOzg4hISGF3tweOHAAMpks17/169eLy1y9ehX9+/eHoaEhmjdvjnnz5hVrSNhvzcHBATNnzixwmdjYWHTv3h1GRkb49ddfsXXr1kK3++zZM4waNQrNmjWDlZUVxo0bl+uFjadPn4aTkxMMDAzg4OCAFStWfNXb33N6/vw5Bg8erPBiyAMHDsDBwQFmZmaYPn06Pn36pLCOk5MTdu3a9c1iICL6HrCjOxH9tA4cOCC+QO6/5s2bNxg4cCAaNmwIPz8/XLt2DX5+flBRUcHgwYPzXS8uLg61a9fGggULFKbL37dy7949ODs7o3bt2li8eDEyMjKwbNky3L59W3yPyfcmPj4eQ4YMgb29PUaOHInTp0/D29sbGhoaaNu2bZ7ryF+i+OHDB0yfPh0AsHDhQri6umLjxo1QUVFBbGwshg4dCjMzM/j7+yMhIQFLlizBs2fPCk2SiurMmTM4deqU+Pe7d+8wceJEDB48GI0bN8aUKVOgp6cnDh188OBBfPz4EZ06dfomn09E9L1gUkJE9B8UERGBzMxMrFixAurq6mjRogXS09MREhIivsAvL3FxcdDT04OxsXGe89evXw9lZWWEhoZCS0sLAKCrq4t27drhr7/+Et9z8j0JCQlBjRo1sGTJEigpKcHW1hYJCQkIDAzMNym5fv06bt26hTVr1sDS0hJA9vtdBg8ejOvXr8PAwACrVq1C1apVERISIr5BvVKlShg6dCicnZ3RsGHDb/5dHjx4gLS0NAwbNgylSpXCnj17cP36dQBAVlYWli9fDk9PTygrs6EDEf1YeFYjov+snTt3wt7eHkZGRhg+fDgePHigMP/q1atwcXGBkZERmjVrhlmzZolve/f390dAQABSU1Mhk8kQGhqKRo0aYfv27eL6hw8fhkwmw7Zt28RpBw4cgJ6eHt6/fw8gu3lP9+7dYWhoCFtbWyxbtixXc5s9e/agU6dOMDAwQKtWrbBu3TqF+TKZDNu3b8fo0aNhYmICCwsLzJkzB5mZmfl+9zNnzsDS0hLq6uritFatWiExMbHA2p+4uDjIZLJ859+/fx+NGzcWExIAqFevHipUqICTJ0/mu15ycjJmz54Ne3t76Ovro1mzZpgwYQKSkpKK9T1fvXoFDw8PmJqawsbGBjt37sz3M+XOnDkDOzs7hRdZtmrVCrdu3crzhYhA9jtqACi8yLJ8+fIAsmsrgOxaI3NzczEhAbLfci8IgkLthlxWVhYsLCzg7+8vTrtx4wZkMpnCSyWvXLkCmUyGiIgITJo0CQBgaWkJf39/VKtWDcrKyjh58iRevXqF69evo0aNGgCA3bt3o3Tp0mjVqlWhZUJE9F/DpISI/pPS0tKwaNEieHh4YMGCBbh//z4GDRqEjIwMAMCdO3fQr18/KCkpwc/PD2PHjsW+ffvg6ekJIPut005OTlBTU0NkZCS6dOkCAwMDREdHi58RExMDILu/gtzp06dhYmICTU1NnD17FkOHDkXNmjUREBCAwYMHIywsTOEdJzt27ICXlxfMzMywYsUKdOnSBb6+vli1apXC95k7dy60tLQQFBSEvn37Yu3atdi8eXO+3//+/fuoXbu2wrRffvlFnJeXlJQUPHnyBNevX0ebNm2gp6eHTp064a+//hKXqVq1Kp4/f67QNyUpKQlJSUl4/PhxvvF4eXnh6NGj8PLywurVqzFo0CDs2bMHQUFBRf6enz59wuDBg3H16lXMmjULEydOxPLly/NNLAAgNTUVL1++LHZZmJiYoFGjRli6dCmePn2Kp0+fYtGiRahWrRpMTU0BANWqVcPTp08V1pOXwZMnT3JtU1lZGVZWVkXah2rUqIF27drB1dUVALBq1Sp0794dlSpVgpubG9zd3dG8eXOUL18evXv3RkZGBgICAjB69Oh8y4KI6L+MzbeI6D9JEAQsXLhQbHpTr149dOrUCXv37kWXLl0QFBSEihUrIiQkBKqqqgCAOnXqoG/fvjh//jzMzMxQtWpVKCsri02ZbG1tsWXLFvEzzp07h8aNG+PChQvitNOnT6NXr14AAD8/PxgZGWHp0qXi+uXKlcOkSZMwePBgVK9eHUuWLEGnTp0wbdo0AEDz5s2hpKSEoKAg9OnTB6VLlwaQfZM8depUANlPzY8dO4YTJ06gT58+eX7/5ORklClTRmGa/O/k5OQ814mLi4MgCHj8+DEmTpwIFRUVbNiwASNGjEBYWBiaNWuGTp06YevWrZg1axZcXV2Rnp6OWbNmQUVFRaxlyunjx4/IyMjA9OnTYWtrCwCwsLDAP//8g3PnziksW9D3PH78OOLi4hAZGSn+JnXq1IGjo2Oen/v5dy1uWZQoUQKzZs3C0KFDYW9vDyC7pmTt2rVi7dNvv/2GSZMmYeXKlejZsycSEhIwbdo0lCxZEqmpqXlu19bWFlOnTkVaWhrU1dXFfejy5cvIyMhAyZIlcfr0adjY2EBLSwu1atUCAOjp6Ym1UyNHjkSfPn2QkpKCX375BUpKSoiIiEDNmjVhbm6OefPm4fjx42jUqBGmTp2qUKtFRPRfxZoSIvpP0tTUFBMSAGjYsCF++eUXselSTEwMrK2toaysjMzMTGRmZsLY2BgaGho4e/Zsntu0sbHB8+fPcf/+fbx79w63bt3C4MGD8eDBA7x69Qr37t3DkydP0KJFC6SlpeHy5cuwt7cXt5+ZmQlbW1tkZWUhJiYG9+7dw8uXL2FnZ5drmZSUFFy+fFn8bCMjI4VYqlSpku+Nb2Hy62/QoEEDhISEYO3atbC3t4etrS0CAwNRv359rFixAgDQrFkzTJ8+HTt27EDz5s3Rpk0bNGjQAHp6egpNxT5XqlQphIaGwtbWFo8fP8apU6cQFhaG+Ph4seaqKN/z77//Rrly5RT6u+jp6YnNl/Iir9H5vOnW59PzK4sbN26gf//+qF+/PoKDgxESEgKZTCb+3gDg6OiI33//HQEBAWjWrBmcnJzQoUMHaGlp5VsWNjY2yMzMxN9//w1BEHDhwgUMHjwYaWlpuH79OlJTU/HPP/8U2jenYsWKqFWrFpSUlPDhwwesXLkSo0ePRkREBE6fPg1/f38oKyuLnfSJiP7rWFNCRP9JFStWzDVNS0sLL1++BAAkJiYiMjISkZGRuZZ79epVnts0MDBAhQoVEBMTg4oVK0JbWxtt27bF1KlTceHCBbx+/RpVq1aFTCbDixcvkJWVhcWLF2Px4sV5fkZiYiKA7KZNXl5eBcaR8yZXWVm5wOF9NTQ0kJKSojBN/vfn/SQ+V7Zs2Vw3wyoqKrCyslIYYrZ3795wcnLCw4cPUalSJZQtWxYODg65mkh97siRI/D19cWjR49QoUIF6OvrQ01NLdfwuQV9z6SkJFSoUCHXtitVqpTv58q/a86ykCc6mpqaea63fv16qKurIyQkRKxVadasGdq1a4fAwEBxdDIPDw8MGzYMjx8/RvXq1aGmpgZfX1+UK1cuz+1qa2ujUaNGiImJgZaWFpKTk+Hg4IA6deogNjYWCQkJUFJSQrNmzfL9TjmtW7cOBgYGMDIywoIFC/Dbb7+hYcOGcHFxQa9evfDp0yeoqKgUeXtERN8jJiVE9J/0eQdqudevX0NHRwdA9s1qy5Yt0bt371zL5XXjC2TfIFtbWyMmJgaVKlVC06ZNUaJECZiYmCA2NhZPnjwRmyfJb2RdXV3RsmXLXNuqXLmy2GF62rRpMDQ0zLWMfBjeL1GnTp1cfTwePXoEILspW16uX7+Oa9euicPLyn348EEskzt37uDWrVto37496tevDyC7rJ8/f45GjRrlud379+9j1KhR6Nq1K9avX4+qVasCAEaNGoX4+Pgif6fy5cvnek8IADG5y0uZMmVQqVIl8bvLyf+uU6dOnus9f/4cDRs2VGj2VapUKejr6+POnTsAgMuXL+PNmzewt7dHgwYNAGQ3gfv06VO+ZQFk15bExMRAW1sbenp6KF26NMzNzREbG4tnz57BzMxMbLZXmOTkZISGhiI8PBxA9lDQ8g75ZcuWxadPn/D27Vtoa2sXaXtERN8rNt8iov+khIQEXLt2Tfz72rVrePz4MczNzQFkj5J09+5d6Ovrw8DAAAYGBqhWrRoWL16M27dvA8i7aY+NjQ3OnTuHCxcuoGnTpgCApk2bIjo6GufOnRNrGjQ0NKCrq4tHjx6J2zcwMEDJkiWxZMkSPH/+HPXq1UP58uXx4sULhWUSExOxbNmyfPs7FEWzZs1w5swZhSZehw8fRvny5aGrq5vnOjdu3MCUKVPEIWaB7ITkxIkTYrnduHEDY8eOVUgONm3aBEEQxIQsp+vXr4vv/ZAnJKmpqbhw4UKx3lRuYWGB9+/fKzSvu3fvHh4+fFjgevK+KZ+Penb48GHo6Ojke7Nep04dxMXFKdSwpKen4/r162KyeObMGUyaNEkcqQvIHopZnmTkx9bWFlevXsXJkycV9qELFy7g9OnTCuVY2NC+q1evRvPmzcVku2LFimIN26tXr6CioiImKURE/2VMSojoP0lVVRVjxozBoUOHsG/fPowcORK6urpo06YNAMDNzQ1Xr17FqFGj8Ndff+HQoUMYOnQobt68icaNGwPIftKclpaGw4cPi82+bGxs8Pr1a1y5cgVmZmYAADMzM9y+fRsfP35U6Mfi4eGBvXv3wsfHB6dOnUJUVBR+//13PH78GDo6OihRogRGjhyJ0NBQLFmyBGfPnsWWLVswbtw4JCcnf1VNSZ8+fcRE4NixY1ixYgVCQkIwbNgwsWN/cnIyLl68iISEBABA27ZtUadOHYwaNQr79u3DkSNHMGjQIKSmpoqjQNnb20NbWxteXl44ffo0wsLC4Ofnh549e+ZbA9OoUSOoqKhg4cKFOHv2LPbt24f+/fvj9evX+XaOz4u1tTXMzMwwbtw47Ny5EwcPHsTvv/+e7ztX5AYPHox79+6Jv/W8efOwe/duuLm5icskJCTg4sWLYiLo4uIilt+RI0dw7NgxuLq64sWLFxg6dCiA7I7uHz9+xKRJk3D27FksWbIEkZGRcHNzy7f5FgAYGxtDXV0dJ06cEPchc3NzvHv3Dnfv3lVoQle2bFkAwKFDh3LVfCUkJGDDhg3w8PAQp7Vo0QJbtmzB8ePHsXLlStja2qJEiRJ5fkciov8UgYjoP2b58uVCmzZthI0bNwrW1taCsbGxMGrUKOH169cKy8XGxgr9+vUTDA0NBTMzM2H48OHCrVu3xPlv3rwRHB0dBT09PSEkJESc3rVrV8Hc3FzIysoSBEEQPn78KBgYGAguLi65Yjly5Ijg6Ogo6OvrC82aNRPGjBkjPH36VGGZLVu2CB06dBD09PQEa2trwcfHR0hMTBTn6+joCKtWrVJYx9XVVejXr1+B5XD58mWhZ8+egr6+vmBnZycEBwcrzI+OjhZ0dHSEbdu2idOePn0qjB49WrC0tBSMjIyEQYMGCXFxcQrrxcXFCf369ROMjY0Fe3t7ISAgQMjMzCwwlt27dwu//vqroK+vL9jb2ws+Pj7Chg0bBF1dXeH58+dF/p5JSUnCxIkThaZNmwrNmjUTQkJChO7duwszZswo8PNPnDgh/Pbbb4K+vr7w66+/KnxnQRCEbdu2CTo6OkJ0dLQ47caNG8LgwYMFY2NjwdzcXBg0aJBw/fp1hfXOnz8vdOvWTTAyMhLatGkjREREFBiH3MiRIwVdXV3h3bt34jR7e3uhVatWCsulpqYKgwYNEvT09HJ9R19fX2HatGkK0z58+CCMGzdOaNKkieDs7CyWbX7fkYjov0JJEIpRt05ERERERPSNsfkWERERERFJikkJERERERFJikkJERERERFJikkJERERERFJ6qdPStjP/+fzvf/m33t8RAX5t/bfH+W4+C99j/9SrPRj4j743/Slv1uxkpLt27dDJpMV+M/Z2RkA4ODggJkzZxa4PZlMhtWrV39R4N9i/cOHD8PHx0f829/fHyYmJl+8vZ9RRkYGxo4dC2NjY5iZmeHJkydSh1SgzZs3w8/P76u2cePGDXTq1An6+voYMWLEtwns/4uNjVV4J4H8mJO/Z+JbcnBwgKWlJd6+fZtr3uHDhyGTyRTem5DzeMvv+Nu7dy8aNWqEkSNHIjMzs9Dv0LlzZ0ycOFFhuzKZDOvXr89z+XPnzkEmk+V7rLq5uUEmk2H//v15zs/rvKWvrw87Ozv4+Pjk+Y6H+Ph4jB8/HjY2NtDX14eDgwOmTp2Ke/fu/c+3X5iCzmNz5syBTCbDqlWr8pzv4OCQZ+zNmzfH2LFj8eLFiwI/O+fxNXHiRHTs2LHY3yGn27dvw8XFpcBlvuRcFBMTA5lMhitXrnzTeOU+P0aSkpLg5eWl8MLPwq5hjx8/hkwmw4EDB75ZTEVVlDLP6d88X/3X7Nq1C7a2tjA0NMQff/zxzbd/69YtuLi4wMTEBHZ2dggJCSnWjeDcuXMxfPjwXNNXr16d5zns2LFjXxVvYdfNnPv68+fPMXjw4DyvT/T9Sk9Px+zZs3HkyJEvWr9EcRa2s7NDZGSk+Hd4eDjOnz+PgIAAcZqGhkaRtxcZGYnq1asXJ4Rvun54eDhKly79xesTcPLkSURFRcHLywsmJiaoVq2a1CEVaOXKlbCzs/uqbQQFBeHt27dYuXIlqlSp8m0C+/+2bt36RTeiXyohIQHz58/HvHnzvsn2Dh8+jPHjx8PBwQFLliwRX+pWXEpKSvjzzz/Rr1+/XPP+/PPPfNd7+/YtTpw4gYYNG2Lr1q1o165dnss5Ozsr3HympKTg7NmzWLVqFd6+fYvly5crfKcxY8agYcOGGDNmDGrUqIGHDx9i3bp1cHR0xNKlS3PtU//29r9EZmYm9u7dK5bNkCFD8lyuTZs2GDRokPh3WloaLl26hKCgINy/fx9bt27N9zO+xfGVlwMHDoiJQ36+x3PR59eoGzduYM+ePRgwYECR169cuTIiIyNRp06dfyfAAhSlzCl/c+fOhUwmw8iRI/HLL798022/efMGAwcORMOGDeHn54dr167Bz88PKioqGDx4cKHrr1+/HuHh4Xkeq3FxcWjatCnGjRunMD2/F6cWVWHXzZz7+pkzZ3Dq1Kmv+kz633v58iXWrVuHpk2bftH6xbpj0NLSgpaWlvj33r17oaqqCmNj4y/68C9d71utT1/v3bt3AAAnJyeFfeNHlpiYiMaNG6N58+ZSh/LVNDU1sWPHDvz222+wsrL6qm2dOnUKo0ePho2NDfz8/Ap9C3dBTExMcOHCBSQkJCjsV1lZWfjzzz8hk8nw6NGjXOvt3bsXZcqUwciRI+Hp6YknT56gRo0auZarVq1arvOHtbU1nj17hr179yIlJQVlypTB48ePMW7cONjZ2SkkWebm5vjtt98wYsQIjB07Fnv27EHVqlX/Z9v/EidPnsTbt2+xePFiDBgwALGxsXleOLS1tXPFbmlpibS0NKxcuRJ37txBgwYNviqWf8P3eC762mvU11xfSVqJiYmwtbWFmZlZoctu374dkyZNQlxcXJG2HRERgczMTKxYsQLq6upo0aIF0tPTERISgv79++d77n3z5g0WLVqEnTt3QlNTM89l4uLiYGNj8833u8Kum9zXCfiX+5R8+PAB06dPh7m5OUxNTTFhwgSFpgufV11/+vQJCxYsgJ2dHfT19dG+fXts3LixwO1/zfrOzs44d+4cjh8/nquZyr59+9CmTRsYGBjA0dERf//9t8K6V69ehYuLC4yMjNCsWTPMmjULaWlp+X6WvInApk2b0Lx5c7Ro0QKPHz9GRkYGli9fjjZt2kBfXx9mZmZwd3fHs2fPxHUdHBzwxx9/wMfHB+bm5mjSpEmucvz48SNmz54NS0tLNGnSBN7e3liyZAkcHBwU4li7di1+/fVX6Ovro0OHDti3b1+B5QsA58+fR9++fdGkSRNYWVlh5syZSElJAZDd1EHe7MbS0lKhCU5OBw8ehKOjI4yMjODg4ICVK1cqVDUfOnQI3bp1g7GxMVq0aAE/Pz9kZGQolEPO5oBz5sxR+I4ymQzbt2/H6NGjYWJiAgsLC8yZMweZmZniNp48eYKIiAjIZDIAQGpqKry9vdG8eXMYGhqia9euOHjwYL7fQyaT4dy5c/jrr78gk8kQExNTaDkB2fvb1KlTMXjwYDRp0gTz58/Pte2JEydix44duH37tsK2ASA6OhqdO3eGgYEBOnTokKtq9MGDB3Bzc4OJiYn4lKsoTSjat2+PRo0aYdq0aQXuw4U5f/483N3dYWVlheXLl39VQgIANjY2KFWqVK7v+ffffyMlJQW2trZ5rrdz505YW1vD3t4eZcqUwbZt24r1uTlreteuXYuMjAxMmzYtV62PqqoqZsyYgeTk5Hybmv2vt1+QnTt3wtDQEJaWlqhXrx62bNlSrPXLlClT4Py8ji+5tWvXwt7eHoaGhnB2dkZ8fLw4TxAEhIeHo1OnTjAwMICJiQkGDhwo3qD5+/sjICAAqamp4jGeU37nort378LDwwPNmjUTm8UFBgZ+UXvnvJo0ypvDfZ4gz5w5E05OTgD+7xoVExOD/v37A8hOmj4/VyYmJmLMmDHiOWvu3LniuS9nk5aJEyfCw8MD4eHh+ZZnXopbDvmVeUJCAqZMmQJbW1sYGRmhf//+hdamnD59Gt27d4ehoSFsbW2xbNkyfPr0SZzv4OCARYsWoUePHmjatCnWrFkDIDuJ7tevH0xMTGBgYIDOnTsrnJv9/f3h6OiIPXv2iNfrbt265bpex8TEoG/fvjAxMYGtrS3mzZuHjx8/ivOLez0vrBzk13sAWLhwYa5j4Vs4c+YMLC0toa6uLk5r1aoVEhMTC/w9Vq5ciQsXLmD16tVo1KhRrvmZmZm4e/dusWPOyMhASEiI+Dt06tQJUVFR4vz8rpuf+3xflydpQPbx7O/vDwD466+/xPsIS0tLTJo0CYmJifnGJT9eVq9eDVtbWxgbG8PDwwPJyckICAiAlZUVLCwsMHv2bGRlZYnrvXnzBuPHj4e5uTlMTEwwYsSIXA/BCts/C7snzauZ7Y0bNxTKZ+LEiXBzc4OXlxeaNGmC0aNHA8i+b5k1axasrKzEc8D169fF7Wzfvh0WFhY4ceKEeF51cnJCfHw8Dh06hDZt2sDExATDhw/HmzdvxPUyMzOxbNky2NnZife/Z8+eFefL9+3Y2Fj06tULBgYGaNmypXgtefz4MVq2bAkAGDVqlNidozj+1aRkx44dePfuHfz8/DBy5EhERUUpNPX63OrVq7Ft2zZ4enpi9erVsLGxwfTp03Hy5MkifVZx1/fx8UHjxo3RpEkTREZGonLlygCymyosXboUHh4eWLZsGdLS0sS28QBw584d9OvXD0pKSvDz88PYsWOxb98+eHp6FhpjUFAQZs6cidGjR6NmzZrw9fXF+vXrMXToUISGhsLT0xNnz57F3LlzFdYLDg5GUlISlixZAk9PT+zduxcrVqwQ50+ePBnbt2+Hu7s7Fi9ejIcPHyIsLExhGwEBAZg/fz7at2+PlStXwsrKCmPGjMm33T2QfQLo378/KlWqhKVLl2LkyJHYu3cvhg8fjqysLLi5ucHV1RUAsGrVKri5ueW5nT///BMjR46ETCZDQEAA+vfvj4CAALGdbWRkJNzd3WFgYICAgAD069cPoaGh4kmpOObOnQstLS0EBQWhb9++WLt2LTZv3iyWQaVKldCmTRuxGeL8+fMRHR0Nb29vBAcHo379+hg1alS+F/nIyEiF/UZPT6/QcpLbvn07atasieXLl+fZrMjNzQ0tWrTAL7/8Im5bbs6cOXB2dkZQUBA0NTUxevRo8WTy+vVr9OnTB0+fPsWCBQswY8YMXLx4EYMHD0Z6enqB5VWiRAnMmjULT58+VWhSVByXL1/G8OHDYWJiAn9/f6iqqn7Rdj5XqlQp2Nra5koQDxw4AAcHB5QqVSrXOnfv3sWVK1fQqVMnqKqqol27dti+fbvCbyCXlZWFzMxM8d/bt2+xa9cu7Ny5E61atRJvwE+fPo1GjRpBW1s7zzh/+eUXNG7cGH/99df/dPvF9f79exw7dgydOnUCkN2P58CBA3j//n2uZQVBUIhdvm5oaCj09fXzbcaR1/EFZPeX2blzJ7y9vTFv3jzcu3dPoWlIaGgoFi1aBCcnJ6xevRpTp07FnTt3xOO/e/fucHJygpqaGiIjI/NscpLXuSglJQX9+/dHYmIi5s+fj+DgYFhYWGD58uVf1D6+WbNmKFmyJKKjo8Vp586dAwBcuHBBnHb69OlcSbOenh6mTZsGAPD19VU4V65atQoVKlRAUFAQevXqhfDwcGzatCnfOM6cOSOW58KFC/HgwYMCHwh9STnkVeYpKSno3bs3zpw5Ay8vLyxduhSCIKBfv375PuE/e/Yshg4dipo1ayIgIACDBw9GWFgYZs+erbBcWFgYbG1tsXDhQtja2uLy5csYNmwYGjZsiKCgICxduhTq6urw8vJSeNhy//59LF++HO7u7vD398fHjx8xatQo8Xp9+fJlDBo0CJqamuK5ecuWLZgzZw6AL7ueF1YOenp64v7v7OyscCx87vNzhPwc9flxV5D79++jdu3aCtPkTcTu37+f73q9e/fGvn378q0Vv3v3LtLT03Hy5EnY29tDT08PPXv2xKVLlwqMZ8KECQgKCkKPHj2wYsUKmJiYYOzYseLNal7XzYLY2dkpHM/du3fHkydP4O7ujiZNmiAkJAQTJkzAsWPHCu27fOrUKRw6dAizZs3CuHHjxAegly5dwrx589C1a1esW7dOfEj74cMH9O/fHxcuXMCUKVOwYMECvH79Gv369RNrY4uyf37tPa3cX3/9hY8fPyIwMBA9e/aEIAhwdXXF3r174enpiWXLlkFVVRXOzs54+PChuF5KSgqmT5+OYcOGYenSpXj+/Ln4/zFjxmDixIk4c+YMli1bJq4zdepUhIWFoX///ggMDES9evUwdOjQXIn+mDFj0KZNG4SEhKBx48aYMmUK7ty5g8qVK4v3+GPGjFHos11UX9bgu4jq1q2LJUuWQElJCVZWVoiOjs4zQwayO/jq6+ujS5cuAAALCwuoqakpPAkoSHHXb9CgATQ0NFC6dGmFKkNBELBw4UJxWmZmJkaOHIk7d+5AV1cXQUFBqFixIkJCQsSbrzp16qBv3744f/58gVW1Li4uCk/2ExISMH78ePGpmrm5Oe7du6fwhAEAqlatKpZj8+bNce7cOZw4cQLjxo3DvXv3sGfPHvj6+sLR0RFA9sVTnq0C2R0sQ0JCMGTIEPFk27x5c6SkpGDx4sX5trtftmwZDA0NFTqu1qxZE0OGDMHx48fh4OCAWrVqAci+6ObXZGLFihVo1qwZfH19AWQ/BX/16hX+/vtvZGVlwc/PDx06dMD06dPF2DQ1NeHj44MhQ4ZAV1c33zLNycTEBFOnTgWQ/YTl2LFjOHHiBPr06YPGjRtDVVVVoXlKbGwsrK2txTIwNTWFtrZ2vhcFY2PjXPtNUcoJyH7SPGXKlHxrEWrVqgUtLS08ffo0VzX25MmT0aFDBwDZzSgdHR1x8eJFtGzZEuHh4fj48SNCQ0PF38DQ0BBt2rTBvn37xGMiPwYGBnB2dkZ4eDg6duxY6AXjc3FxcQgODkZqauo379zapk0bjB8/Hu/fv4empiYEQcCff/6JadOm4ebNm7mW37lzJypWrAgbGxsA2TfemzdvxsmTJ9GiRQuFZRctWoRFixYpTCtfvjycnJzEp1EA8OTJk0L7SNSsWTNX2+d/e/vFtW/fPmRlZaF9+/YAssvGz88PUVFR6NOnj8KyGzZswIYNGxSmlSlTBg4ODpg4cSKUlfN+lpXX8SW3YsUKsR35ixcvMG/ePCQnJ0NDQwPPnj2Dm5ub2Kna3NwcSUlJ8PX1RUpKCqpWrYqqVatCWVk53+YdtWrVynUuunr1KmrVqgU/Pz/xuLC0tMThw4dx/vz5XDXJhdHQ0ICJiQliYmLg5OSEd+/e4datW2jcuDFiY2PRpUsXPHnyBPfv38+1v2loaIhN3ho2bCjGCgBWVlYK56yjR48iJiYm36eMKSkpCA4OFh+kvXjxAnPmzMHbt29RoUKFXMvfu3ev2OWQV5mvW7cODx8+RFRUlPhdmjdvjrZt2yIgIEB8mv05Pz8/GBkZYenSpQAAW1tblCtXDpMmTcLgwYNRs2ZNANn3Cu7u7uJ627ZtQ+vWrRVuaqpXr46uXbvi0qVLsLe3F8tizZo1MDQ0BJD9dNrNzQ03b96Evr4+goODUbNmTQQGBkJFRQVAdsuCHTt24NOnT190Pd++fXuh5SAvs7yaccoFBgbmekj7+bn3yJEjYvnklJycnKvmUv53XgNpyBXWL0SeXL5+/RqzZ8/Ghw8f8Mcff8DFxQXbtm1D/fr181xn7969mDFjBnr16gUguzySk5OxZMkSODo65nndLIiWllau43n//v1IT0/HsGHDxH2/TJkyhQ5okZqaiuXLl4vr7N69G3fu3MG2bdugoaEBW1tb7N+/H5cuXULHjh2xc+dO8T5M/n0tLS1hb2+PdevWwd3dHbdv3y50//zae1q5zMxMzJw5Uzx2T548iejoaISFhYnJpY2NDTp06IAVK1aI91nygT/k5/xLly4hJCQE69evF/fr2NhYMeGMj4/H9u3bMXv2bHTv3h1A9vH66tUr+Pn5Ye3atWJMzs7OGDhwoPj7HDp0CCdOnMCgQYPEGrjatWt/UTPff7WmxMjICEpKSuLfNWvWRFJSUp7LmpiY4NSpU+LN0aNHjzB69Ogid5b52vXlVFRUxBMcALE9uvyJYkxMDKytraGsrCw+0ZAfcJ9Xc+Ul5w/k5+cHJycnvHjxAmfPnkVERAT+/vvvXE+3DQwMFMqxatWqSE1NBZDdbAbIrrqVk7cxlbt48SI+fvwIOzs7hScxtra2ePToUZ5t81NSUnD9+nW0bdtWYbqNjQ3KlSsnfm5hPnz4gBs3bogXEblx48Zh5cqViI+PR0JCQq7PkXcSjo2NLdLnyBkZGSn8XaVKFbGs8mJiYoLNmzdjxIgRiIyMxNu3bzFx4sQiV18Xp5xq1ar1xc2aPq/mzWufNDY2RtmyZcXftlq1aqhfv36h+6TcqFGjULVqVUyZMkWhaUVhdu3aBT09Pfj7++PmzZt5jmz2+b6bn7yWsbOzg4qKivg098KFC0hOTs6z6ZYgCIiKikLLli2RmpqKpKQk6OjooHr16nl2zO7fvz+2bt2KLVu2wN3dHSVLlsSwYcMwdepUhSZWgiCINzP5yWv+v7394tq5cycsLCxQokQJJCUloUyZMjAxMcmzbNq1ayfG7u3tjVKlSqFbt25YuHBhvjU6BalevbpCx1b5/iu/FkyZMgWurq5ISEhAbGwsNm/ejKNHjwJAoTV9BdHX18eGDRugqamJO3fu4PDhwwgICEBmZuYXb9fGxkahyWblypXRoUMHsabk1KlTqFChAgwMDIq8zZxNOGrUqJHvdRLILk/5DRYAsa9Rfk2OvlU5nD9/Hg0aNFC4jqmqqqJVq1ZijdHn0tLScPnyZdjb2+e67mRlZSk8oMx5s9utWzcsX74cqampuHLlCqKiohAREQFAcZ8oUaIE9PX18y2Lf/75B7a2tgrHUL9+/bBt2zaoqKh80fW8uOWQnx49emDr1q3YunWrmJDJ/966dSsqV66cq8a1KOfm/B4aFEWzZs2wcuVKBAcHw9raGi1btsSqVatQunTpfEeJk1+jc14D27dvj4SEhEKbFhaVvr4+VFVV0b17d8yfPx8xMTFwcHAodOCIatWqKRwvFStWRN26dRXOw+XLl1e4ntauXRu1a9cWy11NTQ2mpqZiLWlR9s9vdU+asy93TEwM1NXVYWZmplCr1rx5c4VaXAAK56GKFSuK5ZjX95bvu7a2tgr7XIsWLXLdl36eWJYtWxalS5cu8D6rOP7VmpKcGaGSklK+bViHDRsGdXV1bN26FXPnzsXcuXNhbm6ORYsWFWmEo69dX65UqVIKB7X8//Lq1cTERERGRuZZJfvq1asCt52zJuHvv//G9OnTERcXB01NTTRq1CjPZikFlePbt29RsmRJlC1bVmGZz28e5G0u5U8x8oo75+gg79+/hyAI4o6c83sU9DTmc/Lqzry2U9B8DQ0NlCpVqsifI5ezrJSVlQtsPz5lyhRUrlwZu3btwrFjx6CsrIzWrVtj7ty5RRpJrjjllF8ZFIWampr4/7z2yUuXLuVZw1GpUqUibb906dKYPn06hg4dirCwsCKP9mNqaooVK1ZATU0N3bp1Q1hYGFq0aAELC4tcsX/eR+hzGRkZCt/v85iaN2+OQ4cO4bfffsOff/4Je3v7PI+R6OhoPH36FJs3bxab68m9evUKb968USj/qlWriids+UOIBQsWoFy5cmLNJZB9g/h5H6+8PHnyJNdIT//29ovj0aNHYvV7Xk9+r1+/jsaNG4t/a2lpKcSuoaGBSZMmQUNDA6NGjSr25+d1TAL/t//Gx8dj6tSpuHDhAtTV1aGrqys+9f3adxSsXLkSq1atwvv371GjRg2YmJigRIkSX7xdW1tbLF68GPfu3UNMTAyaNm0KU1NTLFy4EAkJCTh9+jRsbGyKdWNY3HNWYeWZl29RDklJSXkmpdra2gr95z5fPisrC4sXL8bixYtzzf/8epnz3Jiamopp06aJzYvr1q0r1ph/HrOqqmqB1+t3794VeN79kut5ccshP1WqVBHvTW7fvg0AuZJZed8euRo1auDo0aPQ0NDI9Vnyv4szAmpOlSpVyvUAUV5DmFftNJBdxiVKlED58uUVpsvLqLjX8Pz88ssvWLNmjfi0PzQ0FJUqVYKPjw9at26d73p59YUrqLYiMTERd+/ezfN6Kr8uFmX//Fb3pDn338TERKSlpSkkF3I5H3oW57vL7xPz66/5+dDMOa/XhZ2ziuNfTUqKQ0VFBQMGDMCAAQPw9OlTHD58GP7+/vD29s53TP1vuX5RaWhooGXLlujdu3eueXlVnefn/fv3GDFiBJo0aQJ/f3+xfeiCBQvyPfjzUrlyZWRkZCApKUkhMfm8KY18lI3AwMA8D4a6devmmqapqQklJSWFTlByr1+/znUCyo/8oMjZtOf58+d48OCBmKjl/JykpCR8/PhR4XNyXnS/RWaupqYGDw8PeHh44O7du/jzzz8RFBSEhQsXYsaMGYWu/63K6WvIq6A/f7+JXGGdkz9na2uLjh07wt/fv0h9pACgZcuW4glq8uTJOHPmDCZOnIhdu3aJ+6P84vTq1atc+58gCHj58mW+T+B//fVX+Pj4IDU1FQcPHsSUKVPyXG7Xrl2oXr16rqGNExMT4eHhgR07duQ7BC4AjBgxAgcOHMDcuXNhY2MjxmlnZ4e1a9fi5cuXCk/b5J4/f45r164V+j6Hf2v7r169woYNG9C1a1exucOnT58U+vXs3LkTampqWLlypcIN3KdPnzBixAhs2bKlwLa/8s7EwcHBaN26tUIC87WysrLg6uqK8uXLi81hlJWVERER8dVN1nbu3Ak/Pz/4+PigY8eO4nnQ0tLyi7epq6uLypUrIyYmBrGxsejRowf09fWhrq6Oc+fOITo6WmyK9b34VuVQrlw53L17N9f0V69e5Xmek597XF1dFZoTy+W1v8vNmjULp0+fRkhICMzMzKCqqoo7d+7katpcGA0NjVzXnsTERFy7dg1NmjT5out5ccvha/To0UOheefnTcw+H3ABgNji4WuG7j1//jxevnwpNhWW+/DhQ4HlkZmZicTERIXv//r1awD4pmViamqK4OBgpKWliUOsjxo1CseOHftmw/NrampCV1c3V78n4P/Kvyj7Z2H3pEpKSrnuaYqS1GpqaqJixYoIDg7+ym+ae7tKSkrYuHFjnkP5V6hQocD+St/Kd/NG90GDBolt4apXr47+/fujVatWhT5F/Jr1v6Sa09TUFHfv3oW+vj4MDAxgYGCAatWqYfHixeLTjqK4e/cu3r17BxcXFzEhycrKwpkzZ4qVcTZp0gTKyspicwcAYkc1OSMjI5QsWRJv3rwRYzYwMMDt27cRGBiY53bLlCmDRo0a5Xpp18mTJ/H+/Xs0adKkSPFpaGhAR0cHx48fV5i+bt06jB07FvXq1UOFChVyfY6805n8czQ0NPDy5UtxflZWFv75558ixfC5nDdlHTt2FEd7qVevHlxdXWFsbFzk/e5blVNe8RWVfJ+UyWTib6ujo4OAgACFDrhFMXnyZKipqeU7IEVBNDQ0MHv2bDx9+lQhoTMwMICamhoOHTqUa53o6Gi8f/8+3yptBwcHZGZmYuXKlUhKShL7i3wuLS0NBw8eRNu2bWFhYaHwTz6yXUHv1gCynzBNnjxZ7Gcl179/f5QqVQrTpk3LVdOTkZEBHx8flCpVKle/jP/V9j98+ICgoCBcvXpVnPbixQuFC/Tu3bvRokULWFpaKpSNlZUV7OzssGfPHnz48KHA+L29vaGkpJRrEI6cirv/JiQk4MGDB+jRowd0dHTE9XN2Bv2S4+Kff/5B1apV0bt3b/FG/Nq1a0hISPiqp3o2NjY4duwY4uLiYGZmhpIlS8LY2Bhr1qzB+/fv8x3y9Fs0w/sSX1oOOcvc1NQUd+7cUWiOk56ejsOHD+d5ntPQ0ICuri4ePXqkcN0pWbIklixZgufPn+f72RcvXoSNjQ2sra3FG0H5PlGc387ExAQnTpxQuPnbt28fhg8fjk+fPn3R9by45fA1qlSpolB28mbFzZo1w5kzZxQezB0+fBjly5cvVh/MnM6ePYsJEyYo1BLJ+3+am5vnuY6pqSkA5HkNr1ix4he/Yyfn/rdlyxa0bNkSGRkZUFdXh4ODAzw9PfHp06dCX+paHE2aNMHjx49Ro0YNsdz19fWxZs0a8T6mKPtnYfekGhoa+PDhg0JTzaJcr01NTZGQkIDSpUsr7BtRUVHYvXv3F39vU1NTCIKAlJQUhe2ePXsWa9asKfI7x772PPfd1JTIm4FUqlQJBgYGiI+Px4EDB4r8RtkvWb9s2bK4ceMGYmJicvVFyI+bmxt69eqFUaNGoVu3bkhPT0dQUBCePXtWrCeI9erVQ5kyZRAUFISsrCx8+PABGzZswM2bN8XmWUVpi1+7dm106tQJs2fPRmpqKmrUqIG1a9fi1atX4ku7tLS04OzsjHnz5uHdu3cwNDTEzZs3sXTpUrRs2TLf6t6RI0fCzc0Nnp6ecHR0xLNnz7BkyRJxaMWi+v333zFq1ChMnToVbdu2xa1bt7B27VqMHz8eKioqcHd3x6xZs1CuXDm0bNkScXFx8Pf3R9u2baGjowMg+yl+WFgY1q1bhwYNGmDTpk148+ZNsWoCgOzf/Nq1azh//jyaNm0KQ0NDBAYGolSpUqhXrx4uXbqECxcuFKmW5FuXkzy+58+f4/Tp03lWz+Zl4MCB2LVrF4YMGSKOUR8aGoqLFy8WucZDrmLFipgwYcIXjXwGZLdr7dGjBzZv3gw7Ozt06tQJampqcHV1xbJly5CSkgI7OztkZWXh+vXrWLVqFezt7RWae31OU1MTVlZWCA0Nxa+//ppnM6+DBw8iJSUlV5tmuU6dOsHX1zff93LIWVtbw9bWFrt370a/fv1gaGiIqlWrYvHixRg9ejR69eoFZ2dn1KxZE48fP8b69esRHx+PhQsX5tsh9d/efo0aNSCTybBixQqUKVMGL168wO7du8XRnWJjY/Hw4UOFzvWfkzeLO3DgQIEDItSvXx/du3fHxo0bceDAgXzLOufxVRhtbW1Ur14d4eHh0NbWhrKyMnbu3Cle/OV9A8qWLYu0tDQcPnwYhoaGBT5llzMwMMCmTZsQEBAAc3NzxMfHIzAwEEpKSoUmYQWxsbGBp6cnKlSoIPaFaNq0qTjEZ35PlOUJwV9//YXSpUvn2Wn43/Cl5ZCzzB0dHREeHo6hQ4fC09MTmpqaWLNmDV6/fp3nG7oBwMPDA7///js0NDTQunVrvH37Fn5+flBWVhbP7fnFfPToUezYsQPVqlVDdHS02KehOL/diBEj0LdvX3h4eKBHjx54/vw5/Pz80K9fP2hoaHzR9fxLyqEwjo6O4kA1RdGnTx+sX78ew4YNw+DBg3Hz5k2EhITAy8tLvElOTk7GnTt3xAFUiqJnz56IiIjA8OHD8fvvvyM9PR2BgYGoUKFCni+yBbJrD9u0aYN58+YhJSUFMpkMR44cwd69ezFt2rQv7uMir2k/dOgQrK2t0bRpU7x69QqjRo1Cnz59kJGRgRUrVqBmzZp5Dm/8pZycnLBu3ToMGjQIw4YNQ/ny5REZGYmDBw/it99+A1C0/bOwe1IbGxv4+vrC29sbffv2xc2bN3MNMJIXe3t7GBgYYNiwYXB3d0e1atVw8OBBREREFOu+JadGjRqhTZs2GDduHNzd3VG/fn2cO3cOK1aswJAhQ4r8O8rPc2fOnEGdOnWgq6uLhIQEPHz4UBxgqiDfTU3JiBEjMHz4cGzcuBGDBw9GSEgIXFxcFEbk+NbrDxgwAOnp6RgyZIjCGM8F0dfXR3h4ON6+fQsPDw94e3ujSpUqWLduXbGqDzU1NeHv74+kpCS4urpi5syZKF++PJYtW4asrKxCh+D73PTp09G+fXv4+flh9OjRqFGjBtq2bavwtvpx48bBzc0NW7ZswZAhQ7B27Vq4uLgU+CZv+Vj2Dx8+hJubG/z9/dGxY0esWrWqWNlw27Zt4efnh4sXL4q/0YQJE8TRZfr164c5c+YgJiYGI0aMQEREBAYOHKgwetGIESPQsWNHLF26FKNGjUKlSpUwfPjwIscgN3z4cDx48ABDhgzBixcvMGXKFHTu3BkrV67E4MGDsW3bNkyYMEEcfaIovlU5AdkXhYoVK2L48OE4ffp0kdapXr06NmzYAHV1dYwbNw6jR49GVlYWwsLCvuhk7ejo+FVNXCZMmIAaNWpg5syZePr0KYDs32/u3Lm4cuUKPD09MXLkSOzevRuDBg0qdCjiX3/9FRkZGfneCO/evRvVq1dXGKDicx06dICKikqR3ssxfvx4KCsrK9QI2NnZYdu2bWjUqBGWL1+OgQMHYvny5WjUqBG2bdumMMjE/3r7ysrK8PPzg7a2NkaPHg0/Pz/0798fw4YNA5BdNmpqarlGg5KTj4ZUWE0SkH2DqaGhgYULF+bbQTrn8VUU/v7+KFOmDDw9PTF58mSkpaWJQ5pfvHgRQPZvqKenB09PT+zatatI23V0dMSQIUOwadMmDBs2DOvXr8fgwYPh5OQkbvdLWFtbQ0VFBU2bNhUfHMmfIhf0EKJhw4bo3LkzgoODsXDhwi/+/OL60nLIWeYaGhqIiIiAkZERZs6ciTFjxohN7fK7gW/ZsqVYk+fq6oq5c+fC2NgYa9euLbBd/8SJE2FlZYW5c+di5MiRiI6ORkBAAOrUqVOsGnJjY2OsXr0ar169wu+//44VK1bA2dkZXl5eAL7sev4l5fCtVa5cGWFhYcjMzISHhwc2b94MT09Phbe5X7t2DT179szVSqEgVapUQUREBCpVqoRJkyZhypQpqFevHtavX1/gA8BFixahb9++WLNmDVxdXfH3339j4cKF6Nu37xd/R0tLSzRv3hyzZs1CaGgo6tati5UrVyIhIQEeHh7w8vJCxYoVERYW9tXvxfqc/PetV68epk+fDjc3Nzx9+hRBQUHiebQo+2dh96T169fH7Nmzce3aNQwdOhSHDx8u0rD8KioqWL16NaytrbFw4UIMGzYM58+fh6+vb779hotq0aJFcHR0FEdr3bt3L7y8vDBmzJgib0NDQwNDhw7F7t27xaHfjx8/jp49e+LatWuFrq8kfKveKSQJeedKe3t7hQy0V69e0NbW/qJmOERERERE/0vfTfMt+jJqamqYMWMGDhw4gF69eqFEiRLYv38/Ll68mOsFikRERERE3yPWlPwALl++jKVLl+Lq1avIyMiATCaDq6troS9lIyIiIiL6HjApISIiIiIiSX03Hd2JiIiIiOjnxKSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkxaSEiIiIiIgkVULqAIjov83CwgI1atSQOgwi+g948uQJYmJipA6DiL5DTEqI6KvUqFED27dvlzoMIvoPcHR0lDoEIvpOsfkWERERERFJikkJERERERFJikkJERERERFJikkJERERERFJikkJERERERFJikkJERERERFJikkJERERERFJikkJERERERFJikkJEYnevXuHsWPHwtzcHDY2Nli0aBE+ffokdVhERET0g+Mb3YlINGPGDLx69Qrr169HQkICxo4di/Lly2PIkCFSh0ZEREQ/MNaUEJHor7/+gouLC3R0dNCsWTN07NgR0dHRUodFREREPzgmJUQkKl++PHbv3o20tDS8ePECJ0+ehJ6entRhERER0Q+OSQkRiXx8fHDu3Dk0adIEtra20NbWxsiRI6UOi4iIiH5wTEqISPTw4UM0btwY69evR0hICJ48eYL58+dLHVahMj+kSx3Cd4NlQURE/0Xs6E5EALITkrlz5+Lo0aOoWrUqAKBUqVIYNGgQhg8fDm1tbYkjzF8JNVWEKbeUOozvwsCsI1KHQEREVGysKSEiAMDVq1dRpkwZMSEBAH19fXz69AlPnz6VMDIiIiL60TEpISIAQOXKlZGUlIRnz56J0+Lj4wEANWvWlCosIiIi+gkwKSEiAICxsTEaNWqESZMm4ebNm7h48SKmTp2Kzp07Q0tLS+rwiIiI6AfGpISIAAAlSpRAcHAwypUrBxcXF7i7u8Pc3BwzZ86UOjQiIiL6wbGjOxGJqlSpgmXLlkkdBhEREf1kWFNCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSYlJCRERERESSKiF1AET0fdi+fTsmTZqU57xjx46hevXq/+OIiIiI6GfBpISIAADt27eHjY2N+HdWVhZcXV1Rs2ZNJiRERET0r2JSQkQAADU1NaipqYl/r1+/Hk+fPkVYWJiEUREREdHPgH1KiCiX5ORkBAQEwMPDA+XKlZM6HCIiIvrBMSkholwiIyOhqqqK7t27Sx0KERER/QSYlBCRAkEQEBkZiX79+qFkyZJSh0NEREQ/ASYlRKTg2rVrePjwITp37ix1KERERPSTYFJCRApOnDgBIyMjVKlSRepQiIiI6CfBpISIFFy6dAlmZmZSh0FEREQ/ESYlRKTg9u3baNCggdRhEBER0U+ESQkRKXj9+jXKly8vdRhERET0E+HLE4lIweXLl6UOgYiIiH4yrCkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIiIiIiJJMSkhIlFGRgZ8fX1hYWEBCwsL+Pj4ID09XeqwiIiI6AfHpISIRAsWLMChQ4cQFBSEFStW4OTJkwgMDJQ6LCIiIvrBlZA6ACL6PiQlJWHjxo0IDg6GqakpAMDd3R379u2TODIiIiL60TEpISIAwIULF6CmpgYrKytxmqOjIxwdHSWMioiIiH4GbL5FRACAhw8fokaNGtizZw86dOgAe3t7zJ8/n31KiIiI6F/HmhIiAgCkpKTg8ePHWL9+PWbMmIGUlBTMmDEDmZmZ8Pb2ljo8IiIi+oGxpoSIAAAlSpRAcnIyFi5ciKZNm6JFixYYP348IiMjkZWVJXV4RERE9ANjUkJEAIDKlSujRIkSqFWrljitbt26+PjxIxISEiSMjIiIiH50TEqICABgbGyMzMxMxMXFidPi4+NRpkwZlC9fXrrAiIiI6IfHpISIAAB16tRBy5YtMWnSJFy9ehWxsbFYtGgRevTogRIl2P2MiIiI/j280yAi0YIFCzBnzhy4uLigRIkS6NKlC7y8vKQOi4iIiH5wTEqISKShoQFfX1/4+vpKHQoRERH9RNh8i4iIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIhEUVFRkMlkCv/c3NykDouIiIh+cCWkDoCIvh937txB69at4ePjI04rVaqUhBERERHRz4BJCRGJ4uPjIZPJUKlSJalDISIiop8Im28RkejOnTuoW7eu1GEQERHRT4ZJCREBANLT0/Ho0SMcO3YMv/76K1q1aoVFixYhPT1d6tCIiIjoB8fmW0QEAHjw4AEyMzNRunRpLF++HA8fPsScOXOQkpKi0MeEiIiI6FtjUkJEAICGDRsiOjoaFSpUAADo6upCEAR4eXnB29sbJUrwdEFERET/DjbfIiKRPCGRq1+/PjIyMpCQkCBRRERERPQzYFJCRACAgwcPwsrKSqEPyfXr11G2bFmOxkVERET/KiYlRAQAMDMzgyAImDZtGu7du4fjx49jwYIFGDx4MJSUlKQOj4iIiH5gbCRORACym26tXr0avr6+cHR0hIaGBnr16oXhw4dLHRoRERH94JiUEJGocePGWLdundRhEBER0U+GzbeIiIiIiEhSTEqIiIiIiEhSTEqIiIiIiEhSTEqIiIiIiEhSTEqIiIiIiEhSTEqIiIiIiEhSTEqIiIiIiEhSTEqIiIiIiEhSTErohyAIgtQhEBEREdEX4hvdqdgOHz6MjRs34vr16/jw4QNq166N7t27o0ePHihZsuT/PJ7bt29j1qxZWLt27VdtJykpCV5eXoiJiYGmpiaOHTsGVVVVcb6zszPOnTtX4DZ8fX3h6Oj4VXEQERER/WyYlFCxzJgxA5s2bUKXLl3Qu3dvlC5dGufOncP8+fMRHR0NPz8/qKio/E9jOnDgAK5cufLV29m1axdOnDiBefPmoXbt2goJCQD4+PggOTlZ/HvgwIFo3749unfvLk6rVavWV8dBRERE9LNhUkJFtnPnTmzYsAEzZ85Ez549xelWVlbQ0dHB6NGjERUVhS5dukgX5Fd49+4d1NXV0bVr1zznN2jQQOFvFRUVVK1aFcbGxv+D6IiIiIh+XOxTQkW2evVqyGQyhYRErn379hg0aBAqVKggTnv8+DFGjRoFKysrmJiYwNXVFffv3xfn+/v7w8TERGE7N27cgEwmQ0xMDABg4sSJ8PDwQHh4OOzt7WFoaAhnZ2fEx8eL2wgICEBqaipkMhm2b9+eb/yHDh1Ct27dYGxsjBYtWsDPzw8ZGRkAsptm+fv7Iy0tDTKZDP7+/l9URiNHjkTHjh1zTW/Tpg3mz5+Px48fQyaTYf/+/ejXrx8MDQ3Rvn177Nu3T2H51NRUzJo1C1ZWVuJ3vn79+hfFRERERPS9Y1JCRfLy5UvcunULLVq0yHeZCRMmiPOfP3+O7t2748GDB/Dx8YGvry8eP36MPn364MWLF8X67DNnzmDnzp3w9vbGwoUL8eDBA0ycOBEA0L17dzg5OUFNTQ2RkZGws7PLcxuRkZFwd3eHgYEBAgIC0K9fP4SGhmLSpEkAsptmfb6dz5tkFUfnzp1x+/ZtxMXFidMuX76M+/fvo3PnzuK0qVOnQldXFwEBAdDT08OYMWNw6tQpANmd9l1dXbF37154enpi2bJlUFVVhbOzMx4+fPhFcRERERF9z9h8i4rk+fPnAIDq1asXafk1a9bgw4cPCA0NhZaWFgDA3NwcrVq1QlhYmJhUFEVKSgqCg4NRuXJlAMCLFy8wZ84cvH37FlWrVkXVqlWhrKycbzOqrKws+Pn5oUOHDpg+fToAoHnz5tDU1ISPjw+GDBkCXV3dQrdTFC1atICWlhb27NkDmUwGAIiKioKOjg50dXXx+PFjAICNjQ2mTJkCALC1tcW9e/cQHByM5s2b49SpU4iOjkZYWBisrKzE5Tt06IAVK1bA19f3i+MjIiIi+h6xpoSKRN55PSsrq0jLnz9/HhYWFmJCAgBaWlqwtLQsdASrnKpXry4mJABQtWpVAEBaWlqR1o+Pj0dCQgLatm2rMF3ezCo2NrZY8RSkZMmS6NChA/bu3QsA+PTpE/bt26dQSwIAHTp0UPjbwcEB//zzD7KyshATEwN1dXWYmZkhMzMTmZmZALITqejo6G8WKxEREdH3gjUlVCTVqlUDADx79izfZV6+fAltbW0oKysjKSkJjRo1yrVMxYoVcefOnWJ9trq6usLfysrZuXRRE6R3796Jn/05DQ0NlCpVSmFErW+hS5cuWLduHf755x8kJycjISEBnTp1UlimUqVKCn9raWkhIyMDqampSExMRFpaGvT19XNtW4ohl4mIiIj+bUxKqEi0tLTQuHFjnDx5EmPHjs1zmYEDB0JbWxvh4eEoV64cXr9+nWuZ169fo3z58gAAJSWlXIlFSkrKN49d/nlv3rxRmJ6UlISPHz+K878VfX19NGzYEH/++SdSUlLQrFkzVKlSRWGZxMREhb/fvHmDUqVKoUyZMtDU1ETFihURHBz8TeMiIiIi+l6x+RYVmYuLC27evIktW7bkmrdr1y7cuXNHrBEwNTVFTEwMEhISxGUSEhJw9uxZNGnSBEB2TcWHDx+QlJQkLnPhwoVixyWvOclP3bp1UaFCBRw4cEBhunzEK3k839Jvv/2GI0eO4Pjx47mabgHAsWPHFP4+cuQIzM3NoaSkBFNTUyQkJKB06dIwMDAQ/0VFRWH37t3fPFYiIiIiqbGmhIqsc+fOOH78OKZNm4bLly+jZcuWUFJSwqlTp7Bx40a0a9cO3bp1AwAMGDAAO3bswKBBg+Dm5gZBELBixQqoqqrCxcUFQHbnbV9fX3h7e6Nv3764efMmNmzYUOy4ypYti7S0NBw+fBiGhoYK/U+A7P4w7u7umDVrFsqVK4eWLVsiLi4O/v7+aNu2LXR0dL6+cHLo3Lkzli5dilKlSqF169a55m/ZsgVaWlowMTHBzp07ERcXh/Xr1wMA7O3tYWBggGHDhsHd3R3VqlXDwYMHERERgRkzZnzzWImIiIikxqSEikxJSQlLlizB5s2bsX37dhw8eBDp6emoW7cupkyZAicnJygpKQHI7oMSERGBhQsXYsKECVBRUYG5uTmWLl0qdlSvX78+Zs+ejRUrVmDo0KEwMjLC8uXL0aNHj2LF1aFDB+zcuROenp4YNWoUhg4dmmuZfv36QU1NDaGhodiyZQsqV66MgQMHws3N7esLJg9VqlSBTCZDgwYNUKZMmVzzPT09cejQIaxatQo6OjpYtWqV+M4WFRUVrF69GosWLcLChQuRnJyM2rVrw9fXF46Ojv9KvERERERSUhIEQZA6CKIfzcuXL2FnZ4dVq1aJw/oC2S+UbNmyJZYtW5ZrNLD/KkdHxwJfWvm/EqbcUuoQvgsDs45IHQJRvr6X8wURfX9YU0L0DT18+BC7d+/G4cOHUb9+fVhaWkodEhEREdF3jx3dib4hQRAQHh6ODx8+YOHChWJzNiIiIiLKH2tKiL6h2rVr4/z58/nOr1mzJuLi4v6HERERERF9/1hTQkREREREkmJSQkS5eHt7w9nZWeowiIiI6CfBpISIFJw9exZbt26VOgwiIiL6iTApISJRamoqpk6d+q+85Z6IiIgoP0xKiEi0dOlSmJubw9zcXOpQiIiI6CfCpISIAAD//PMPDhw4gAkTJkgdChEREf1kmJQQEdLT0+Ht7Y3JkyejXLlyUodDREREPxkmJUSEwMBA1K5dG+3atZM6FCIiIvoJ8eWJRISoqCi8evUKJiYmAICMjAx8+vQJJiYm+OeffySOjoiIiH50TEqICOvWrUNmZqb495o1a3D16lUsWrRIwqiIiIjoZ8GkhIhQo0YNhb/Lli0LNTU11K5dW6KIiIiI6GfCPiVERERERCQp1pQQUS6jR4+WOgQiIiL6ibCmhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIiIiIiIJMWkhIhE8fHxGDBgAExMTGBvb49Vq1ZJHRIRERH9BJiUEBEAICMjA0OHDkW1atWwc+dOTJs2DUFBQdi9e7fUoREREdEPjkkJEQEAXrx4AUNDQ/j4+KB27dqwt7eHlZUVzp8/L3VoRERE9INjUkJEAICaNWvCz88PampqEAQBFy5cwPnz52FpaSl1aERERPSDKyF1AET0/bG1tcXLly9hb2+PNm3aSB0OERER/eBYU0JEuQQFBSEoKAjXrl2Dr6+v1OEQERHRD441JUSUi4GBAQDgw4cPmDBhAsaPHw9VVVWJoyIiIqIfFWtKiAhAdkf3I0eOKEyrX78+MjIykJycLFFURERE9DNgUkJEALLfUTJy5Ei8efNGnHbt2jVoaWlBS0tLwsiIiIjoR8ekhIgAAGZmZqhfvz4mTpyI+Ph4HDt2DIsXL8aIESOkDo2IiIh+cExKiAgAULJkSQQHB0NFRQXdu3fHtGnT4OLigv79+0sdGhEREf3g2NGdiETVq1fHypUrpQ6DiIiIfjKsKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEiIiIiIkkxKSEi0cOHDzFixAiYmZnB1tYW8+bNw8ePH6UOi4iIiH5wJaQOgIi+D+np6RgxYgQaNGiATZs24c2bN5g8eTIAYOLEiRJHR0RERD8y1pQQEQDg8uXLePjwIXx9fVG/fn2Ym5tj1KhRiIqKkjo0IiIi+sExKSEiAEC9evUQEhKCMmXKiNOUlJSQnp4uYVRERET0M2BSQkQAAC0tLVhZWYl/Z2VlYf369TA1NZUwKiIiIvoZsE8JEeXJ19cXN27cwNatW6UOhYiIiH5wTEqISIEgCJgzZw42btyIZcuWoWHDhlKHRERERD84JiVEJMrKyoK3tzeioqKwdOlStGrVSuqQiIiI6CfApISIRPPmzUNUVBT8/f1hb28vdThERET0k2BSQkQAgIsXLyI8PBxeXl7Q19fHq1evxHmVKlWSMDIiIiL60TEpISIAwJ9//gkAWLx4MRYvXqww79q1ayhRgqcLIiIi+nfwLoOIAAATJkzAhAkTpA6DiIiIfkJ8TwkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkREREREUmKSQkR5ZKeno6OHTvizJkzUodCREREPwEmJUSk4OPHjxgzZgxu374tdShERET0k2BSQkSiO3fuoEePHnj48KHUoRAREdFPhEkJEYliY2NhbW2NyMhIqUMhIiKin0gJqQMgou9Hr169pA6BiIiIfkKsKSEioh9S5od0qUP4brAsiOh7x5oSIiL6IZVQU0WYckupw/guDMw6InUIREQFYk0JERERERFJikkJERERERFJikkJERERERFJikkJERERERFJih3diShPcXFxUodAREREPwnWlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBARERERkaSYlBCRKD09HVOnToWZmRmsra3xxx9//L/27jyq6jr/4/gTElDBA6LCMck9CVEhNgtZ3MY0MpW0YUqdGdFRR1Irw7Ss5LiERwUla3IhHTcQdXLUxiWX6oBgIiikqIALWTFMEG6UwOX3h4fvr5tohdl1eT3Ouefw/Sz3+77few/n+/5+Pp/v19IhiYiIyH2ggaUDEJE7x7x588jKyuKDDz7gm2++ITo6mgcffJCwsDBLhyYiIiL3MI2UiAgAV65cYcOGDUyfPp0uXbrQt29fRo8ezZo1aywdmoiIiNzjlJSICAB5eXlcvXoVX19fo8zX15ecnByqqqosGJmIiIjc65SUiAgAJSUlODo6YmdnZ5Q1b96cyspKSktLLRiZiIiI3Ou0pkREAKioqMDW1tasrHb76tWrN+x3/vx5wsPDb2tsv8hgR0tHcEfYeid8F3cS/S6AO+d3cf78eUuHICJ3KCUlIgKAnZ3ddclH7XajRo1u2C8jI+O2xiUiIiL3Pk3fEhEAXF1duXDhglliUlJSgq2tLY6OutosIiIit4+SEhEBwMPDAxsbG7KysoyyzMxMPD09adBAg6oiIiJy+ygpERHg2hStwYMHM3PmTI4ePcqePXtITExk5MiRlg5NRERE7nFWNTU1NZYOQkTuDBUVFbz11lvs2rULe3t7Ro0axahRoywdloiIiNzjlJSIiIiIiIhFafqWiNxTevfuTUpKynXlaWlpuLu735Z9hoSEsHnzZgBGjBhBXFzcbdmP/L8xY8YQHR1tVvbJJ5/g7u7O7Nmzzco3bNhA9+7dqampueHv45dISUmhd+/e9Y75l8jIyMDd3b3O14oVKwBwd3cnLS3tN9nfjh07KCkp+U3eS0TkVmj1qoiI3HX8/f3ZtGmTWVl6ejouLi6kp6eblWdnZ+Pn54eVlRUbN26kcePGv2eo9fLpp59ibW1+3dDBweE33cf58+eZNGkSu3bt+k3fV0SkPjRSIiIidx0/Pz/Onj3LhQsXjLKMjAwiIyM5deoUpaWlRnl2djYBAQEAODs707Bhw9893l+rWbNmtGjRwux1s+cF1Ydmb4vInURJiYjcdy5evMjUqVPx9fWlR48ezJgxg0uXLhn1+/btY8iQIXTt2hVfX18mT55sVp+UlERoaCi+vr68//77N9xPZWUlsbGxhISE4OnpSa9evVi3bp1R37t3b5KTk3nmmWfo1q0bkZGRnD9/nqioKLy8vBg8eDAFBQUAbN68mWeffZa4uDh8fHwIDQ0lKSnpNhydu0PXrl2xs7MjJycHgAsXLpCXl8fAgQNp3bq1MVpy8eJFCgsL6d69O2A+vW/EiBEsWbKEyMhIvLy8GDhwIJ988omxj+LiYkaPHo23tzfh4eF8+eWXZjEUFBQQGRmJj48PQUFBJCQkYDKZOHbsGB4eHnz33XcAlJeX88gjj7By5Uqj77hx43jvvfdu+Tj88MMPzJ8/n9DQULy9vRk3bpzZU9OzsrJ47rnn8PLywtvbm8jISIqLiwHo06cPAP369TOmH4qIWIqSEhG570yfPp2ysjLWrl3L+++/z+nTp5k2bRoARUVFvPDCC0RERPCf//yHRYsWkZ6ezvr16wH47LPPmD17Ni+++CJJSUlkZ2cbJ3k/tWzZMvbu3cvixYvZsWMHQ4YMYdasWWbtFy9ezIsvvsjatWvJzc1lyJAhBAcHk5KSgrW1NfHx8UbbY8eOkZubS1JSEhMnTmTWrFlmJ9H3ExsbG7y8vDhy5AgABw8epF27djRr1oyAgAAjKcnOzsbR0fGG64mWLl1KWFgYmzdvpl27drz22mtUV1cDMGnSJEwmEykpKYwePZp//vOfRr/S0lKee+45XFxcSElJ4a233mLt2rUkJibi4eGBs7Mzhw4dAuDzzz8H4PDhwwBUVVVx8OBBQkJCbvk4vPnmm+zatYvY2FiSk5Opqqpi/PjxVFdXc+nSJcaOHUtgYCDbtm1jxYoVfPnll0YyVJucJScn8+STT95yLCIit0JrSkTknhMTE8OcOXPMympPNM+dO8fu3btJT0/HyckJgNjYWHr37s3XX39NdXU1r732Gn/84x8BcHNzIzAwkPz8fODaiVxYWBiDBw8GYPbs2YSGhtYZR6dOnZg9ezbe3t7AtavjS5Ys4fTp07i6ugIwaNAggoKCAAgICKCsrMzY99NPP33douzY2FiaN29Op06dOHjwIMnJyTfc/73O39+fo0ePAtfWk9SOhnTv3p133nkHgCNHjhjrSeoSEhJCeHg4AH//+98ZNGgQxcXFXL58maysLPbs2YObmxsPP/wwOTk57Ny5E4Bt27ZhZ2dHTEwMNjY2dOjQgZKSEhYtWsTo0aMJDAwkIyODvn378vnnnxMSEmIkJdnZ2TRs2JDOnTvf9LP9WGBgIEuWLDErKy8vZ8uWLfzjH//gscceA2D+/Pn07NmTzz77DE9PT8aOHcuoUaOwsrLioYceol+/fsYDUp2dnQFo2rTpXTGlTUTubUpKROSeExUVRf/+/c3KsrKymDp1KgUFBdTU1NCrV6/r+p05c4bHH38cW1tb3nvvPU6dOsWpU6fIz88nLCwMuDZlZ9iwYUYfZ2dnWrVqVWccffv2JTU1lbfffpvCwkKOHTsGgMlkMto89NBDxt92dnY8+OCDZttXr141a9u8eXNju0uXLqxZs+YXHZN7kZ+fnzGFLSMjg6ioKOBacnfmzBm+/fZbsrKybjoi8ePjX7uQvKqqivz8fBwcHHBzczPqu3TpYiQlBQUFdO7cGRsbG6P+0UcfpaysjNLSUoKCgkhMTASujZS89NJLjBs3jnPnzpGWlkZwcPANEyWATZs28cADDxjbda0nOXPmDCaTCS8vL6PMycmJdu3aUVBQQM+ePRkyZAgrV67k+PHj5Ofnc+LECbp163bD/YqIWIqSEhG55zg7O9OmTRuzstp59tXV1TRu3JgPP/zwun4tWrQgLy+PP/3pT/Tq1QtfX1/+8pe/sGrVKrN2P10g/OMT0x+Li4sz1owMGjSIN99887pbyjZoYP5v+Kd3XLpZ2+rq6pu2v9d5e3tTXl7OF198QX5+vjG64OrqStu2bcnMzCQnJ4dXXnnlhu9R13dX+/3+9Hv+8fG3s7O7rl9tsmkymQgKCmLatGkUFRVRUFCAv78/Xbt25fDhw6SlpTF8+PCbfrbWrVtf933/VF0xwLXfRXV1NcXFxTzzzDN4eHgQFBTEs88+y/79+8nMzLzp+4qIWIKSEhG5r7Rr144rV65QXV1N+/btATh79ixz584lJiaGLVu24OPjw8KFC40+Z8+eNZKc2mk8tS5dukRRUVGd+0pKSmLGjBk89dRTAMYUsPre9aioqIhLly4ZV/Rzc3Nv27NX7gaNGjXC09OT9evX8/DDDxvTkeDaFK6dO3diZWVVr2PUqVMnLl++TGFhofE7qR3pAujQoQM7duygsrLSSGyysrJwcnLC2dkZa2tr3N3dWbp0KZ07d8bOzg4/Pz/27dtHbm4uPXr0uMVP//+Jy5EjR4wpfGVlZZw9e5b27duze/du7O3tWbZsmdFn9erVxu/vZiM1IiK/t/v3EpuI3Jc6dOhAcHAw0dHRHDlyhLy8PKZOncq3336Li4sLTk5OnDx5kiNHjnDmzBnefvttcnJyqKysBOD5559n165dJCUlUVBQwOuvv84PP/xQ576cnJzYt28fRUVFZGZmGg/7+/GUrF+joqKCN954g4KCAjZs2MCOHTt4/vnn63cg7hF+fn5s377dWE9SKyAggD179tx0PcnNdOjQgccee4zp06eTl5fHxx9/bNzsAOCpp57CZDIZ38eePXtISEggIiLCGL0KCgriX//6F35+fkasO3fupHPnzjRt2vQWPvU1jRs3JiIigtmzZ5Oens6JEyeIjo7G1dWV4OBgnJyc+O9//0tqaipFRUUsXbqUXbt2Gb+/2ue15OXlcfny5VuOR0TkVigpEZH7zrx582jTpg2jRo1i+PDhuLi48O677wLXbhPr4+PDX//6VyIiIoxb9B4/fhy4tgB57ty5LFu2jKFDh+Lq6kqnTp3q3M+cOXM4efIkYWFhTJ06lf79++Pt7W12xf3XcHFxoVWrVgwdOpTly5czb9686xZE32/8/f25cuVKnUlJRUWF8XyS+oiPj6d58+ZEREQQFxfHiBEjjDp7e3uWL19OUVERgwcPJiYmhpEjRzJp0iSjTVBQEJWVlUZS4uvri7W1NcHBwfWO6adeeeUVgoKCmDRpEhEREdja2rJq1Srs7OwYMGAATz/9NJMnTyY8PJz09HSmTZvG6dOn+f7772natCnh4eG8/PLLbNy48TeLSUSkPqxq9PQkEZE73ubNm4mPj+fTTz+1dCgiIiK/OY2UiIiIiIiIRSkpERERERERi9L0LRERERERsSiNlIiIiIiIiEUpKREREREREYtSUiIiIiIiIhalpERERERERCxKSYmIiIiIiFiUkhIREREREbEoJSUiIiIiImJRSkpERERERMSilJSIiIiIiIhFKSkRERERERGLUlIiIiIiIiIWpaREREREREQsSkmJiIiIiIhYVANLByAiIrcmJyeH1atXc+jQIUpKSnBwcODRRx8lMjISX19fo11CQgKJiYlkZWVZJM6MjAxGjhx50zatWrVi7969v1NEIiJyp1BSIiJyF9uwYQMzZ87Ex8eHiRMn0qpVK/73v/+xceNGRowYQVxcHE888YSlwwTA09OT5ORkY/ujjz5i1apVZmW2traWCE1ERCxMSYmIyF0qLy+PmJgYwsLCiI2NxcrKyqgbMGAAkyZNYubMmfTq1euOONl3cHDA29vb2M7OzgYwKxMRkfuT1pSIiNylli9fjq2tLdOnTzdLSGpNnDgRPz8/ysrK6uxfWVnJ4sWLeeKJJ+jSpQv+/v5ERUXx9ddfG20KCwsZPXo0fn5++Pj4EBkZSV5e3i+u/7VOnDiBu7s7O3bsMCvfunUrXbp0oaysjFdffZWxY8eSmJhIYGAgfn5+vPzyy3z33XdmfVJTUxk2bBjdunUjJCSERYsWUV1dXe/YRETk9lFSIiJyl9q/fz+PP/44Tk5OddZ36NCBxYsX4+rqWmf93LlzWbNmDWPGjCExMZHJkydz4MAB5syZY7SZMGEC1dXVxMXFERcXR1lZGWPHjjVO7n+u/tdyd3fHw8OD7du3m5Vv3bqV0NBQmjZtCkBmZibr1q1jxowZvP7666SlpTF+/Hij/YEDBxgzZgxubm688847REZG8sEHHzBr1qx6xSUiIreXpm+JiNyFysvLuXjxIq1btzYrr6mpuS4heOCBB+ocSSktLSU6OpqhQ4cCEBAQwOnTp9m6datRX1hYyIQJEwgODgagZcuWbNu2jStXrlBZWXnT+iZNmtTrsw0ePJgFCxZw8eJFmjRpQmlpKampqcTFxRltLl26RFJSEh07dgTAycmJsWPHcvDgQQICAoiPj8fLy8voExISgqOjI9OmTSMyMhI3N7d6xSYiIreHRkpERO5CtYnHT5ONjz76CE9PT7NXYmJine8RHx/P0KFDKS4u5sCBA6xdu5bDhw9z9epV4NqJftu2bZkxYwbTp09n586dtGrVipdeeokmTZr8bH19DRw4EJPJxO7du43PZG9vT8+ePY027u7uRkICEBoaio2NDYcOHaKiooKjR4/Sq1cvqqqqjFdISAgmk4mMjIx6xyYiIreHRkpERO5Czs7ONG7cmK+++sqsPCgoiI0bNxrbtaMgdTl8+DBvvfUWJ06coEmTJnh4eGBnZ2fUW1tbs3LlShISEtizZw+bNm2iYcOGREZG8sILL/xsfV2jM79Es2bNCA4OZvv27YSHh7N161b69+9vtli/RYsWZn2srKxwcnKivLycCxcuYDKZWLBgAQsWLLju/UtKSuoVl4iI3D5KSkRE7lKhoaGkpqZSUVFBo0aNAHB0dKRr164/2/fixYuMGzcOHx8fEhISaNOmDQDz5s0zW6jesmVL5syZg8lkIjs7m5SUFJYsWULHjh158sknf7a+vgYNGsSUKVM4efIk2dnZREdHm9X/dFG7yWSirKyMZs2aYW9vD8D48ePp06fPde/t4uJS77hEROT20PQtEZG71N/+9jcqKiqIiYmpc2F5fn7+DfsWFhZSXl7On//8ZyMhMZlMpKWlUVNTA1y75XBQUBBffPEF1tbW+Pj4MGvWLBo0aMBXX331s/W3ok+fPjRu3JiZM2fi5uZm9hDI2ti++eYbY3v//v1UVVXRvXt3HBwceOSRRygqKqJr167Gy8bGhoULF5r1ExGRO4NGSkRE7lKdO3dm1qxZvPHGG5w6dYphw4bRtm1bLly4wL59+/j3v/9Ny5Yt8ff3v65v+/btsbe3591338VkMvH999+zbt068vLysLKyoqamho4dO2Jvb8/UqVOJiorC0dGRDz/8ECsrK3r27Enbtm1vWn8rbG1tGTBgAMnJyUyYMOG6+qqqKsaNG0dUVBTl5eXMnz+fnj174uXlBVy7HfKECRNwcHDgD3/4A2VlZcTHx2NtbU2nTp1uKTYREfntWdXUXhITEZG7UkFBAatXryY1NZXi4mIaNmyIu7s7/fv3Jzw83JjalZCQQGJiIllZWcC153jMmzeP06dP07RpU/z8/OjXrx8TJ04kOTkZb29vzp07R2xsLJmZmVy5cgV3d3cmT55Mjx49AH62/mZWrlzJ3LlzOXHiRJ31H3/8MRMmTGDnzp20bdvWKH/11VfJzc1l4MCBrFixAisrKwYOHMiUKVNo2LCh0W7v3r0sWbKEkydP4uDgQGBgIFOmTKFly5b1PdQiInKbKCkREZE7Uu0i/PXr15uV1yYl27Zts1BkIiLyW9P0LRERuaNs3LiR48ePs2HDBhYuXGjpcERE5HegpERERO4oubm5bNmyheHDh9O/f39LhyMiIr8DTd8SERERERGL0i2BRURERETEopSUiIiIiIiIRSkpERERERERi1JSIiIiIiIiFqWkRERERERELEpJiYiIiIiIWNT/ARRdTTOc7ED0AAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "dataset.replace({\"GlassType\": dict},inplace=True)\n", "element='Ba'\n", "Badict={'clear facsimile':[0.22, 0.52],'Nikumaroro jar':[0.59,0.89]}\n", "low=0\n", "high=dataset[element].max()\n", "\n", "#Run report for all glass types\n", "numgroups(element,low,high,None)\n", "\n", "#Run reports for the clear facsimile and Nikumaroro jar\n", "for jar_type in Badict:\n", " low=Badict[jar_type][0]\n", " high=Badict[jar_type][1]\n", " numgroups(element,low,high,jar_type)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Barium Analysis" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Barium:
\n", "The ratio of 2:1 for the element Barium in the two glass samples (.74 for the Nikumaroro jar and .37 for the clear facsimile jar) suggests that the glass maker, Hazel-Atlas, used a recipe. This is not uncommon in the glass industry. \n", "For the clear facsimile, one window, one container and one headlamp have measurements of barium in the database that are within +- 0.15 of its measured value of barium.\n", "For the Nikumaroro jar, eight headlamps and one window have measurements of barium in the database that are within +-0.15 of its measured value of barium.\n", "\n", "Conclusion: We have not yet applied machine learning to the identification of the Nikumaroro jar or the clear facsimile, but it would appear from the weight percent of some of the key ingredients from these jars that windows, of the float or non-float variety, and, to a lesser extent, headlamps, are strong candidates for how the 1987 database might predict their identity, if machine learning were employed as a predictive tool to do this. Nevertheless, some containers did appear in the bar graphs for the clear facsimile, indicating that there is at least a slight resemblance between that jar and containers from the U.K. database. The results are not exactly encouraging to the success of a machine learning model, but they are not discouraging enough to dampen curiosity at discovering what machine learning might be able to do with the data from the Nikumaroro jar and the clear facsimile." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Turning the question around: If one were to subset the 1987 database only for containers, how would that database compare with the two jars? Both of the jars are, of course, containers." ] }, { "cell_type": "code", "execution_count": 86, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Container summary report:\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Ref_ixNaMgAlSiKCaBaFe
count13.00000013.00000013.00000013.00000013.00000013.00000013.00000013.00000013.000000
mean1.51892812.8276920.7738462.03384672.3661541.47000010.1238460.1876920.060769
std0.0033450.7770370.9991460.6939201.2823192.1386952.1837910.6082510.155588
min1.51316011.0300000.0000001.40000069.8900000.1300005.8700000.0000000.000000
25%1.51666012.7300000.0000001.56000072.1800000.3800009.7000000.0000000.000000
50%1.51994012.9700000.0000001.76000072.6900000.58000011.2700000.0000000.000000
75%1.52119013.2700001.7100002.17000073.3900000.97000011.5300000.0000000.000000
max1.52369014.0100002.6800003.50000073.8800006.21000012.5000002.2000000.510000
\n", "
" ], "text/plain": [ " Ref_ix Na Mg Al Si K \\\n", "count 13.000000 13.000000 13.000000 13.000000 13.000000 13.000000 \n", "mean 1.518928 12.827692 0.773846 2.033846 72.366154 1.470000 \n", "std 0.003345 0.777037 0.999146 0.693920 1.282319 2.138695 \n", "min 1.513160 11.030000 0.000000 1.400000 69.890000 0.130000 \n", "25% 1.516660 12.730000 0.000000 1.560000 72.180000 0.380000 \n", "50% 1.519940 12.970000 0.000000 1.760000 72.690000 0.580000 \n", "75% 1.521190 13.270000 1.710000 2.170000 73.390000 0.970000 \n", "max 1.523690 14.010000 2.680000 3.500000 73.880000 6.210000 \n", "\n", " Ca Ba Fe \n", "count 13.000000 13.000000 13.000000 \n", "mean 10.123846 0.187692 0.060769 \n", "std 2.183791 0.608251 0.155588 \n", "min 5.870000 0.000000 0.000000 \n", "25% 9.700000 0.000000 0.000000 \n", "50% 11.270000 0.000000 0.000000 \n", "75% 11.530000 0.000000 0.000000 \n", "max 12.500000 2.200000 0.510000 " ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "Mg : min to max: 0.0 to 2.68\n", "reference: Clear facsimile jar = 2.4 wt%\n", " Nikumaroro jar= 4.3 wt%\n", "\n", "Ca : min to max: 5.87 to 12.5\n", "reference: Clear facsimile jar = 3.6 wt%\n", " Nikumaroro jar= 8.5 wt%\n", "\n", "Ba : min to max: 0.0 to 2.2\n", "reference: Clear facsimile jar = 0.37 wt%\n", " Nikumaroro jar= 0.74 wt%\n" ] } ], "source": [ "sampcont = dataset[dataset['GlassType'].isin(['Container'])]\n", "sampcont=sampcont.drop(['ID', 'GlassType'], axis=1)\n", "print('Container summary report:')\n", "display(sampcont.describe())\n", "rangedict = {'Ba':['Barium',[.37,.74]],\n", " 'Mg':['Magnesium',[2.4,4.3]],\n", " 'Ca':['Calcium',[3.6,8.5]]}\n", "ELEMENTS=['Mg','Ca','Ba']\n", "for element in ELEMENTS:\n", " print()\n", " print(\n", " element,': min to max:',sampcont[element].min(),'to'\n", " ,sampcont[element].max())\n", "\n", " facsimile_ref=rangedict[element][1][0]\n", " artifact_ref=rangedict[element][1][1]\n", " print('reference: Clear facsimile jar =',facsimile_ref,'wt%'\n", " '\\n Nikumaroro jar=',artifact_ref,'wt%')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "When we restrict the U.K. database to only containers, the comparisons with the Nikumaroro jar and the clear facsimile enter into sharper relief. The clear facsimile has a high magnesium content for a container but this is still within range of the lowest and highest values within the U.K. database. The jar from Nikumaroro has a magnesium content much higher than that of the highest nearest neighbor in the U.K. database.\n", "\n", "The clear facsimile has a calcium measurement lower than that of all the containers in the U.K. database. This indicates, perhaps, that calcium was not as heavily used in glassmaking in the early part of the twentieth century. The Nikumaroro jar has a calcium content, 8.5, that is lower than the 25th percentile. This is a low value for a container, especially when one considers that the mean for containers in the U.K. database, 10.12, is skewed toward the higher end of the range.\n", "\n", "The clear facsimile jar and the Nikumaroro jar are within the range of minimum to maximum in barium weight percentage for containers; however, most of the measured values of barium in the database are zero. The measured values of barium for the clear facsimile and the Nikumaroro jar are still far higher than those usually found in the U.K. database." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### What do the correlations between elements for the different types of glass in the 1987 database reveal about late 20th century glassmaking, as compared with early 20th century glassmaking?\n", "To answer this question, we can generate custom diverging colormaps from the U.K. database (excluding our samples of clear facsimile and Nikumaroro jar). The correlation colormaps, also known as heatmaps, show, for example, which elements of the glass most strongly correlate with refractive index, which may be considered synonymous with brilliance. \n", "\n", "To keep the number of graphs to a manageable amount, we will restrict them to the glass types of Window Float, Window Non-Float and Containers. These three heatmaps provide a good summary of the correlations of ingredients in glassmaking of the late twentieth century." ] }, { "cell_type": "code", "execution_count": 87, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj8AAAHoCAYAAACmf449AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAB7tklEQVR4nO3dd3QUVR/G8e+md3pN6JhQQiAU6b0KqKgUQVFBlI6g0kEQ84r03hRBRISANCkqItKkCdJ77zWUkF523j+QlRWEQGQ3cZ8PZ89h5t6dPLOb3f3lzp1Zk2EYBiIiIiIOwsneAURERERsScWPiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY9IKuhkSfkn+t0QSbtU/Mh/xurVq3n77bepWLEioaGhNGnShDlz5pCYmPiv/6xLly7x9ttvc+PGjRTf59y5cwQFBfHjjz/+63n+LUFBQXz55Zf2jpGmLFq0iKCgIK5fv57i+xw9epQ333zTsrx161aCgoLYu3fv04goIo/Jxd4BRP4NH3/8MfPmzaNJkya0bNkSLy8vtm3bxrBhw9iyZQtjx47F2dn5X/t5mzZtYuPGjY91n+zZsxMeHk7+/Pn/tRySNv34449WhU7x4sUJDw+nUKFCdkwlInep+JF0b8mSJXz77bcMGTKEFi1aWNZXqlSJwMBAevTowbJly2jSpIn9QgJubm6UKlXKrhnEPnx8fPTci6QhOuwl6d6XX35JUFCQVeFzV8OGDWnbti2ZMmWyrDt37hzvvfcelSpVIjQ0lI4dO3Lq1ClL+4QJE3j55ZdZvnw59evXp0SJErzyyiv88ccfwJ3DIH379gWgYsWKTJgwAYArV67Qt29fqlSpQvHixalSpQr/+9//SEhIsPzcew979enTh27dujFr1ixq1qxJSEgIrVu35vjx41b78Ntvv9GsWTNCQkKoVq0a48aNIzk52dJeq1YtRo4cSfPmzSlbtixfffXVAx+n+Ph4wsLCqFixIqVLl6Z///6MHj2aWrVq/eNju2fPHt555x3Kli1LcHAw9evXZ968eVZ9pk+fTt26dSlRogR16tRh0qRJmM3mFLc/SHh4OI0aNSIkJIQGDRowf/58S5thGMyfP5/nn3+ekJAQ6tWrd98+BwUFMXXqVBo1akT58uX58ccf6dOnD506deKDDz6gdOnS9OjRA4CYmBg++eQTKlWqZHkODhw48I/ZDMNg1qxZPP/885QoUYLQ0FDatGnD4cOHgTu/PxMnTiQmJoagoCAWLVr0wMNeP//8M6+88gqlSpWievXqjB071uoQba1atfjiiy8YNGgQzz77LKVLl6Z3795ERUU99LETkRQwRNKxy5cvG4GBgcbIkSNT1P/ixYtGhQoVjBdffNH48ccfjR9++MFo3LixUbFiRePSpUuGYRjG+PHjjdDQUKNu3brG999/b/z6669Go0aNjCpVqhiJiYlGRESEMWbMGCMwMNBYv369cfHiRSM5Odl4/vnnjRdeeMFYtWqVsWnTJmP48OFGYGCg8fXXXxuGYRhnz541AgMDjR9++MEwDMPo3bu3UaZMGaNJkybGzz//bPz4449G1apVjaZNm1rybtq0yShatKjRvXt3Y926dcZXX31llCxZ0hg8eLClT82aNY1ixYoZEyZMMNasWWMcP378gfv+/vvvG6GhocY333xjrFmzxnj99deN4OBgo2bNmpY+gYGBxvTp0w3DMIzz588bJUuWNLp162Zs3LjRWLt2rdG+fXsjMDDQOHjwoGEYhrFixQojODjY+Oabb4ytW7caU6dONYKCgoy5c+emqP1BZsyYYQQFBRlDhw41fvvtN8tjvWzZMsMwDGPkyJFG0aJFjdGjRxsbNmwwRo0aZVm+dz+KFy9uzJkzx1i5cqVx9epVo3fv3kaxYsWMzp07G5s2bTI2b95smM1m44033jDKly9vhIeHG2vWrDHatm1rlC5d2jh9+rRhGIaxcOFCIzAw0IiIiDAMwzCmT59uBAcHG1999ZWxdetWY+HChUaVKlWMl156yfI71q9fPyMkJMTYuXOnERERYWzZssUIDAw09uzZYxiGYcybN88IDAw0Bg0aZGzYsMH4/PPPjRIlShgffPCB1fNapkwZo3v37saGDRuMWbNmGcWLFzeGDx/+j4+diKSMih9J13bv3m0EBgYa3377bYr6Dx061ChVqpTlg8wwDCMiIsIIDQ01hg4dahjGneInMDDQ2L17t6XP6tWrjcDAQGPv3r2GYdz/gXjhwgXj9ddftxQFdz3//PNG165dDcN4cPFTpEgR4/Lly5b+s2bNMgIDA43r168bhmEYzZs3N1599VWrbS5evNgoUqSIcfbsWcMw7nxINmrU6KH7feLECSMwMNBYuHChZV1MTIxRsWLFfyx+1q5da7z55ptGQkKCpf3GjRtGYGCgMXv2bMMwDGPw4MFG/fr1DbPZbOkzffp045dffklR+98lJycbzz77rNG3b1+r9e+//74xePBg4/r160bx4sXvK3ZHjhxpFC9e3PJ8BAYGGu+++65Vn969e1s9Z4ZhGOvXrzcCAwON3377zbIuMTHRqFevntGnTx/DMO5/rj/55BNj8uTJVtueOXOmERgYaERFRRmGced3qFSpUpb2e4uf5ORko0KFCkaPHj2stjF37lyrwvLu83rvY9e5c2ejcePGD3zsRCTlNOdH0rW7k5gfdRjlrt9//53y5cuTOXNmy7rMmTNTsWJFtm3bZlnn4uJCcHCwZTlnzpwAxMbGPnC7uXLlYvbs2ZjNZk6dOsWpU6c4dOgQERER5M6d+x/z5M6dm+zZsz/w53h4eLBnzx569OhBUlKSpU+1atUwm81s3bqVgIAAgEdOpP39998BqFOnjmWdp6cn1atXZ+vWrQ+8T/Xq1alevTrx8fEcOnSIU6dOWQ7b3D2UFxoayrfffssrr7xCgwYNqFGjBm+//bZlG49q/7uTJ09y8+bN+w7FjRo1CoC1a9eSmJhIgwYNrNobNWrE559/zu7du6lZs+Y/PiaZM2e2eu63bt2Kp6cn5cqVs3qMq1Spwpo1ax6YccCAAQBcv36dEydOcOLECUvfhIQEvL29/3H/AI4fP87169fv24fGjRszaNAgtm/fTpEiRQAoUaIEJpPJ0idnzpwcPHjwodsXkUdT8SPpWq5cuQC4ePHiP/a5cuUKWbNmxcnJicjISIoWLXpfnyxZsnDs2DHLspubG05Of02Ju/v/hxVZCxYsYOzYsVy7do1s2bJRsmRJ3N3dH3q9F09PT6vle39OZGQkZrOZUaNGWT7873X16lWr/A9z48YNXF1d8fPzs1qfNWvWf7xPcnIyn332GeHh4SQmJpI3b17Kli0L/HUNmxdeeIHk5GTmzJnD6NGjGTVqFEWKFGH06NEUKlToke1/d/PmTQCrAuVet27demDuu/t/73yYBz0mf1938+ZNYmNjrQrdu1xdXR+Y4fjx4wwcOJAdO3bg6elJkSJFLAXPw57rv+/D37P4+Pjg7u5utQ9///0wmUy6fpDIv0DFj6RrmTNnplixYmzYsIEPP/zwgX3atGlD1qxZmTVrFhkyZODatWv39bl27RoZM2Z84hzbtm1j4MCBdOrUiddff93y4d20adMn3ubdD9SOHTtSu3bt+9rvHTF6lOzZs5OYmEhkZKRVAfSwa9dMmTKF+fPnM2zYMKpXr46XlxexsbF89913Vv1eeuklXnrpJSIiIlizZg2TJk2iS5cu/PDDDylqv5evr+8Dc508eZIbN25YnqNr166RI0cOS/vd5/Rxn0NfX1+yZMnCtGnTUtTfbDbTsWNHMmbMyLJlyyhcuDBOTk7MmTMnxZc+uJsxIiLCan1kZCTx8fGp+j0UkZTR2V6S7r355pscOnSIBQsW3Ne2dOlSjh07xvPPPw9AmTJl2Lp1q9WH6/Xr19m8eTOlS5dO8c+8d1QIYNeuXZhMJjp27GgpfC5fvsyRI0ee+C91Hx8fihQpwtmzZylRooTl5urqyujRo7l06VKKt1W6dGmcnJysDuUkJCSwYcOGf7zPrl27CA4O5rnnnsPLywvA0v/uPvXv359u3boBd0YymjVrRtOmTS0jcY9q/7uCBQuSMWNG1q5da7V+3LhxDB8+3LL/f79Q5MqVK3FxcSEkJCSlDwlw5/fh+vXreHl5WT3Gy5Yt4/vvv7+v//Xr1zl9+jTNmzcnMDDQ8nvw98fx778f9ypQoACZMmV64D4Aj/V7KCJPRiM/ku69+OKLrF27lo8++og9e/ZQu3ZtTCYTGzduZO7cuTz33HO88sorALz11lssXryYtm3b0qlTJwzDYMqUKbi5uVldkfdR7o6e/Pzzz1SuXJkSJUpgNpv59NNPadCgARcvXmTKlCkkJCT84zyhlOjWrRudO3fGx8eHunXrcuPGDcaOHYuTkxOBgYEp3k6+fPl4/vnnCQsLIyYmBn9/f77++muuXr36j3OSSpQowRdffME333xDYGAge/fuZdKkSZhMJuLi4gAoV64cvXv3ZvTo0VSqVIlLly4xd+5c6tatm6L2v3NxcaF9+/aMGDGCTJkyUbFiRX7//Xd+/PFHJk6cSObMmWndujVffvklzs7OlCtXjt9//50vv/ySNm3akCFDhsd6fGvWrEmJEiV499136dKlC7ly5WLVqlXMmTOHjz/++L7+WbNmJXfu3MyaNctyKHXJkiWWYu3uc+3n50dsbCyrV6++ryBzdnamS5cufPLJJ2TIkIHatWtz+PBhJkyYQIMGDR7reRWRJ6PiR9I9k8nE6NGjmT9/PosWLWLVqlUkJCRQoEABBgwYQNOmTS2TRnPlysWcOXMYMWIEvXv3xtnZmWeffZYxY8ZYJhunRMWKFalSpQqffPIJzZs356OPPqJv377MmjWLhQsXkjNnTp577jlcXFyYNWuWZYLw46pduzaTJ09m0qRJLFq0CB8fHypVqsSHH35433yQRxk8eDAeHh6MHTuWpKQkGjduTIMGDazmOt3r3Xff5erVq0ycOJH4+Hjy58/PwIEDWb58OTt37gSgSZMmREVFMWfOHL766it8fX2pX78+H3zwQYraH6Rt27a4u7sza9YsvvrqK/Lnz8/o0aMtk7V79uxJpkyZCA8PZ/r06fj7+9OrV6/HKl7vcnZ25ssvv2TkyJGMGDGCqKgo8uXLx9ChQ3n55ZcfeJ8JEyYQFhZG9+7d8fHxoUSJEsycOZO33nqLXbt24e/vT6NGjViyZAndu3fnvffeu68Aev311/Hw8GDGjBksWLCA7Nmz06ZNGzp16vTY+yAij89kaPacyH/e9evX+e2336hZsyY+Pj6W9a+++ipZs2Zl4sSJdkwnImJbGvkRcQAeHh58/PHH/Pjjj7z66qu4uLjwww8/sGvXLmbOnGnveCIiNqWRHxEHsWfPHsaMGcO+fftITEwkKCiIjh07UqNGDXtHExGxKRU/IiIi4lB0qruIiIg4FBU/IiIi4lBU/IiIiIhD0dleadiK3w/YO0KqNCpXjJsXzto7xhPLmDsPr46eZ+8YT2ze+69yfv0qe8dIFf9q9fQc2JF/tXr8sD19f5Hqc2WLcmn7b/aO8cRylq1s7wj/SRr5EREREYei4kdEREQcioofERERcSgqfkRERMShqPgRERERh6LiR0RERByKih8RERFxKCp+RERExKGo+BERERGHouJHREREHIqKHxEREXEoKn5ERETEoaj4EREREYei4kdEREQcioofERERcSgqfkRERMShPLL4qVWrFkFBQZZbkSJFePbZZ+nYsSMXL15M0Q/ZsGEDVapUITQ0lCNHjjy074QJE2jZsmXK0ouIiIg8phSN/PTp04eNGzeyceNG1q1bx5gxYzh69Ci9e/dO0Q8ZOXIkVatWZdmyZRQsWPChfdu2bcuUKVNStF0RERGRx+WSkk4+Pj5ky5bNspwjRw66detGz549uX37Nr6+vg+9/+3btwkNDSUgIOCRP8vb2zslkeQR9v/xO6sWz8fJ2Ylnq9emYs16Vu2RN28wZ/IYkpKS8MuYiZbtu+Hm7s7ubZtZs2wRmKBizXpUqFnX5tnNZjPDx47n6PHjuLm60q/nB+Tx97e0b9i0mS+/no2zszPPP9eAJo0bkZSUxMdDh3Hx8mWcnJzo9+H75M+b1+bZ71W6YG5eqVCcZLPB2v0nWLP3hFV7Fl8vOj9XARMQFZfAhJWbSUhKpmHpIGoGF+R2bDwAX6z+nYs3bts0u9lsZtyc+Rw/dx5XFxc+fLMV/tn/eg84dPI0U+YvwgAy+/nRr90buLm68u6QYXh7egCQM2sWerd53aa5/+6//Bys3vI7839eg7PJiQZVKvBijaokJCYy/Ks5XLwagZenB++1akZAjuw2zX2vfX9s46fF83FycqZC9dpUrHX/+9DsyaNJ/vN9qFX79/58H9rE6mWLMGGiYq16VLTT+9CYmd9w7MxZ3Fxd6NnuLQJy5rC0r960he9+/BknJycK5clDjzav4+R0Zzzhxq1I3hnwMaP6fki+3Llsnl0e7Ynn/Li5ud3ZgJMTt2/fpnfv3pQpU4bKlSszcOBAoqKiAAgKCuL8+fMMHDiQ1q1bP3K79x726tmzJ/Xq1SMhIQGAlStXEhISwokTJx62CYeXnJTEkm9m0L7PIDoPCGPLmp+JvHnDqs8vyxZRtmpNun70KTn887B5zU+YzcmsCJ9Nh76DeW/wZ/y6YglRtyNtnn/dxt9ISEjgy0kT6PRuO8ZNnmppS0pKYuykKYwfMYypY0ezZPkKIq5f57ctW0lOTmb6xPG8/UZrpkyfYfPc93J2MvFGjVA+XbiWj+evoXaJQmTw8rDq06h0EJsPn+Hj+Ws4F3GLmsF3RkUL5MjE5B+3MGTBGoYsWGPzD12Ajbv2kJCYyMS+H/DOyy8wZf5iS5thGIz6ei692rzO+N49KBdclEsR10lITARgTM/3GNPzPbsXPv/l5wBg6ndLGPl+F8b36cGCVWu4HR3Dig2b8HR3Z1K/D+jasinjv11g89x33X0f6thnMF0HhrHp11X3vQ+tXraQclVr0e2joeTwz8OmP9+Hls2bTae+H9P9489Ys2KxXd6HNu7YSUJiIlM+7s+7LZoyeU64pS0+IYEvFyxmbP9eTB7cn6jYGDbv3A3ceY8aOWMW7n9+Rkra9ETFz6lTpxg/fjxVq1bF29ubfv36cePGDebMmcO0adM4efIkffv2BWDjxo3kzJmTPn36MGHChMf6OX379iUyMpIZM2Zw8+ZNwsLC6Nat2yMPnTm6yxfOkTVHLry8fXBxcaVAUFFOHD5g1afJ620pU7k6ZrOZm9ev4ZMhI05OzvQePgFPL2+ib98GDNzdPR78Q56i3Xv3UeHZcgCUKFaMQ/fMEzt5+gwB/rnx8/XF1dWVkiWC2bVnL3nzBJBsNmM2m4mOjsbFJUWDmk+Nf2Y/Lt2MIjo+kWSzmcPnr1HEP5tVn1NXb+DtfucN0tPNlWSzGYCC2TPR5NliDG5RmxfLFbV5doB9R09QLrgYAMUKFeDw6TOWtrOXr+Dn4813q3+l+4hx3I6OIW/OHBw/e564hAR6jpnE+yPHc+D4Sbtkv+u//BwAFPTPTXRMLAmJiRiAyQSnL17i2T/vkzdnDs5cumzr2BZ/fx8qGHj/+9BLr79N2bvvQxHX8PXLgJOTM31HTPzrfcjALu9Dew4f5dmSwQAUf6YQh0+esrS5urgwaXA/PNzdAUhONuPm6grA5G/n82LtmmTNlNHWkeUxpOgTYsiQIXz66afAnarW1dWV2rVr069fP86cOcPPP//Mli1byJgxIwDDhg2jVq1aXLx4kVy5cuHs7IyPj4+lPaUyZ85M3759GTx4MLt37yYgIIA2bdo81jYcUVxsDJ5eXpZldw8P4mJirPqYTCbMycmM7NeDxMRE6r3UHABnZ2f2/L6ZhV99QbFSZXB2cbZpdoDomBh87jn86eTkRFJyMi7OzkTHRFu1eXl6ERUdjZenJxcvXaL5m224dSuSUZ+G2Tz3vTzdXImNT7QsxyYm4uXuatXn+u1YWlYpSeUi+XB1ceK7zfsA2HT4DKt2HSUmIYkPXqhC6Wu3+OPkBZvmj4mLsxy+AnB2ciI5ORlnZ2cio6LZf/wkXVs2JSB7dvpNmEpgvjxk9PWleb1aNKpaiXOXr9Jn/BS+/mQAzs62/x2C//ZzAFDAPxcdwkbg4e5G1dIl8fHyonCeALbs2UeV0BAOnjjFtRs3STabcXay/Ym9cbExeHje8z7k6UnsP7wPDe/bnaTEROq/1AK48z60+/fNLPzqc7u9D8XExuLt6WlZvvd9yMnJicwZMgCw8KfVxMbFUbZEcX5Yt5GMvr48GxLMnO9X2DyzpFyKip8uXbrQoEEDYmJimDhxIhcuXKBHjx5kypSJXbt2YRgGNWvWvO9+p06dIleu1B3vfPHFF1m8eDFr1qxh+fLldnsjTQ9WLpjDycMHuXD2NPkKPWNZHx8Xh+cD5lI5u7jQe/gEjuzbzbdTx9FlwP8ACClXkeAy5Zn3+QS2b1jLs9Vr22oXAPD28iLmnjdJs9nA5c/n3dvLm+jYWEtbTGwMPj4+zF2wkPLlytL5nXZcvnKFTu/35NsZX9h86Ll5pRIU8c9G3mwZOHYxwrLe09WVmPgEq76vVSvJlJ+2suf0JUIL5KJTgwoMX7KelX8cITbhzof2zpMXyJ89k80/eL08PIiNi7csm82G5bXn5+2Ff/as5P9zLkO54KIcOX2Wl2tXxz97VkwmE3lyZsfP24uIW5Fkz5zJptkd4Tk4fu48W/buZ87QwXh6uPPp9Fms3b6T5ypX4PTFS7w/cjzBhQvyTL48Ni98Vsyfw4kjB7h45jR5CwVa1sfHxuLp9eD3ob4jJnJ4327mTB1H14F33odKlqtIiTLl+XbaeH7fsJbyNn4f8vL0JCYuzrJs3PM+BHfmBE2du4Czly7zSffOmEwmVq7biMkEO/Yf4NjpM3w6ZTqfftCNLBkz2DS7PFqKXhWZM2cmX758FC1alDFjxpCcnEznzp1JTEwkOTkZLy8vlixZYnVbtWoVJUuWTHXAmJgYzpw5g7OzM9u2bUv19v7LGjZ7jc4DwhgyaSbXLl8iOuo2SUmJnDi0n3yFg6z6fjdzGkcP7AXA3cMTk8mJuJgYJob1JykxEScnJ9zc3TGZTDbfj5Dg4mzaeue53nvgAIULFrC0FciXl7PnznMrMpLExER27t5LiWLF8PX1tYwI+fn6kpyUhDk52ebZ52/ay5AFa2g/dQk5Mvri7eGGs5MTRQKyceSeD2KA6PhEYv78gL0RFYuPhxuebq6MfPM53F3v/F0SnCcHJ65ct/l+BBcuyNa9+wE4cPwkBQP++iMmV7asxMbFc/7KVQD2Hj1O/tw5+eG3LUxZcGdeyrWbt4iJiyNLBj+bZ3eE58Db0xN3Vzfc3VxxdnIik58vUTExHDp1hhKFCzGm53tUCS1J7mxZbZ67UfPX6Drgf3wy+SuuXb5oeR86fmg/+Z+xfh9aMHMqR/ff+z5kIi4mhgmf3Ps+5GGX96ESgYXZuutOtv1Hj1Mgj79V+8gvvyYhMZH/9ehiOfw14aM+jB/Yh3EDelM4X176dWynwieNeuyJEW5uboSFhdGiRQtmzpxJ7dq1iYmJITk52TIX5/Tp0wwdOpQhQ4bgdc/hlycxbtw4PD09LYfeateuTc6cOVO1zf86ZxcXXnytDZ8PG4JhmHm2em0yZs5CdNRt5k+fRJvufahavxHfzZh654wwk4mmb72Lh5cXZSpVZ2JYf5ydncmVJz9lqlS3ef4aVauwbccftOvSDcMwGNi7Jz+t/oWY2Fheer4x3Tt14L1efTCbDZ5/rgHZs2WlZbNXCBs2gne7dScpMYmO7driec+Qta0lmw1mr9tJv5erYzKZWLvvBDeiYvH2cKN93XKMXvYbM9fsoE2tMjg5mTABM37ZTmxCIvM27uGjZjVJTDaz78xldp1M2fW0/k1VQkPYceAQXT4bDYZBr7de45et24mNj6dxtcp8+GYrwr74CgwoXrgAFUKCSUxKYtjMb+g2bAwmoOebr9l1pPa//hw0rlaZbsPG4OriQu5sWalfqTwxsXHMXLqc+at+wcfLkw/ffM3mue9ydnGhyettmDrsYwyzmfLV61jeh8K/mETbHn2oVr8x82dM4afF4ZicTDRt0/7O+1Dlaoz/pB/Ozi7kzpuPsnZ4H6patjTb9x6g0+D/YRjQp31bfv5tC7HxcRQpUICV6zYQEvQMPT4dAcAr9etQrVwZm+eUJ2MyDMN4WIdatWrRsWNHmjVrZrV+4MCBLF++nB9//JH+/ftz8+ZNBg4ciLu7O4MHDyY5OZkFCxY8dBsPMmHCBDZt2sTcuXPZt28fLVq0YMaMGTz77LO89tpr+Pn5MXXq1Edu579gxe8HHt0pDWtUrhg3L5y1d4wnljF3Hl4dPc/eMZ7YvPdf5fz6VfaOkSr+1erpObAj/2r1+GH7QXvHSJXnyhbl0vbf7B3jieUsW9neEf6TnvhgcI8ePXB1dWXYsGEMHz6cfPny0bZtW15//XWyZ8/O5MmTUxUsOTmZgQMH0qBBA8qXL4/JZGLQoEFs2LCBlStXpmrbIiIi4rgeedhrzZo1D1yfOXNmqzk4o0aNeuxtPEjXrl3p2rUrAIsXW1/XIigoiP3796d4WyIiIiJ/py82FREREYdi0yvB/fTTT/Tp0+cf24OCgpg3L/0e3xcREZG0z6bFT5UqVViyZMk/trvpcuAiIiLylNm0+PH29tYXl4qIiIhdac6PiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY+IiIg4FBU/IiIi4lBU/IiIiIhDUfEjIiIiDkXFj4iIiDgUFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQTIZhGPYOISIiImIrLvYOIP9tG/Ydt3eEJ1Y1uBDn16+yd4wn5l+tHpNWbrJ3jFTp3LAS1w7vt3eMJ5Y1qDiLN++1d4wn9lLFEun6NQB3XgeREdfsHeOJ+WXJau8I/0k67CUiIiIORcWPiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY+IiIg4FBU/IiIi4lBU/IiIiIhDUfEjIiIiDkXFj4iIiDgUFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQVPyIiIiIQ7Fb8VOrVi1atGiBYRhW67du3UpQUBBJSUl2SiYiIiL/ZXYd+dm1axfz58+3ZwQRERFxMC72/OH+/v6MHj2aunXrkjlzZntGkTRo1+9bWb7gW5ycnalSqx7V6jawao+4eoWvJo0l2ZwMhsEbHbqR0z+Ak8eOMP+rLzAMgwwZM9HuvZ64urnZNLvZbGbcnPkcP3ceVxcXPnyzFf7Zs93Xb9TXc/H19uLdV14k2Wxm1NdzOXvpMk5OTvR667UH3sdWTuzbxbZVS3FycqZY+aoEV6xu1R4XHcXXQ/uSJac/AIVKlKZU9XoAJCbEs2TKSGq/2pbMOXLZPLvZbGbk1M85dvIUbq6u9OnSiYDc1jni4uPpPnAwfbt1Jl9AAABfL1jIxm2/k5SUxEvPNeD5enVsnv1eB3ZuZ833C3BycqZs1Zo8W6OuVfvtmzeYN20cyclJ+GbIRLN2XYiPjeHbKWMsfS6eOUWDZq9RoVZ9m2Z/ktcAwLcrV7Fp914Sk5J5sUZVGlataNPcd5nNZoaNHMnRo8dwdXNjQN8+5Pnz9wRg/caNTJ8xExdnZ55v3JiXXnzB0nb9+g1at23LpLFjyZ8/nz3iyyPYdeTnrbfewtvbmxEjRjyw/fjx47Rr147Q0FBKlChBy5YtOXr0qI1Tij0kJSUR/tXn9PgojF5DhrH+5x+4deO6VZ8l82ZT67nG9BoyjIYvt2DhnK8wDIOvp4znrc496PO/kRQPLUPE1Ss2z79x1x4SEhOZ2PcD3nn5BabMX3xfn2XrNnLy/AXL8ubdewGY0Od92rzY6IH3sZXk5CQ2LJ1Lkw4f8kqXPuzbvJboyFtWfa6cO01gaHle6dKHV7r0sRQ+l8+cZOGEz7gVYfvH/a71W7aRkJDI5yM+o8MbrzNhxldW7QePHqNT3wGcv3TZsu6PvfvYd+gwU4d9ysRPP+HKtWs2Tm0tOSmJFXO/ou2HA3m378dsW7ea2zdvWPVZu2IJpavUoEO/MLLnDmDr2lX4ZsxE+75DaN93CA2avUbu/AV5tobti7gneQ3sOnyUfcdPMr53D8b2fI8rN27cdx9bWbt+PfEJCcz44nO6dOzA2PETLG1JSUmMGTeeiWPHMG3yJBYvXcq1iAhL29Dhw/Fwd7dXdEkBuxY/np6e9OvXj8WLF7Njxw6rNsMw6NSpE7lz52bp0qXMmzcPs9nM8OHD7ZRWbOniubNkz5kbbx9fXFxdKVy0OEcP7rfq0/zNdpQo8ywAZnMyrq6uXL5wHm9fX1YvX8Lwgb2IjrpNTv+AB/2Ip2rf0ROUCy4GQLFCBTh8+oxV+/7jJzlw4hSNq1W2rKsSWpIPWr8KwOWI62Ty87Vd4L+5cfkiGbJmx8PLG2cXF3IXCOTCiSNWfa6cO8XVc6f5buJnrPxqEtG3bgJ3PrQbte1Cpuy2H/G5a8/Bg1QoHQpAcJEgDh07btWemJjI0L69yRfgb1m39Y9dFMyXl76fDqNX2FAqlStr08x/d+XiObJkz4mXtw8uLq7ke6YIJ48ctOrTuNVbhFashtls5tb1CHz9MlraDMPg+2++5KU33sHJydnG6Z/sNfD7/oMU9M/FR5On03/iNCqGBNs08712795DpfIVACgRHMzBQ4csbSdPnSIgIAA/Pz9cXV0pVTKEXbt3AzB2wkRebtKErFmz2iW3pIzdz/aqU6cONWrU4OOPP7aa5BwXF0fTpk3p3bs3efPmpXjx4rz00kscO3bMjmnFVuJiY/D08rIse3h4EhMTbdXH1y8DLi4uXDp/jvmzvuSF5q9x+/Ytjh8+SM0GjXl/0Kcc2rObg3t22Tg9xMTF4e3pYVl2dnIiOTkZgIibt5j1/Urea9X8vvs5Ozvz2YzZTJj7HdXKlLJV3PskxMXi7uFpWXbz8CA+LsaqT6bsuSjfoAlNu/ShYInSrF00B4DcBZ/BN1MWm+b9u+iYGLy9//r9cXZyIunPxx8gpFhRcmSz/nC6dTuSQ8eOE9b7Q3p2bM/Ho8fed0KGLcXHxuJxz2vA3cOTuFjr58BkMmGYzYwd8D4nDu0j3zNFLG0Hd20nh38esuXyxx6e5DVwKyqKw6fPMqhDW7q/3oL/TZ9lt+cgOiYabx9vy7KTs7PlMyo6Ohqfe9q8vLyIiopi2YoVZMqYkYoVyts8rzweu875uWvAgAE0atSI2bNnU6zYnb8UPD09adWqFUuXLmXfvn2cOHGCAwcOkDFjRvuGladq8bezOHroAOdOn6TgM0GW9XFxsXh5+9zX/9De3XzzxWTadfuAnP4BGOcMsufMRe48eQEoHlqGUyeOUjSklK12AQAvDw9i4+Ity2azgbPznb++1+3Yya2oaPqOn8L1yEjiExLJmzMHDSrf+SuzT9vWXL8VSadPRzJzSH88bTh8vnnlQi6cOMq1i+fImbegZX1CXBzuHl5WffM8UxQXtzvZCpUozZYf7HeY7u+8vbyIiY21LJsNMy7ODx/9yODrSz5/f1xdXckX4I+7qys3b90ik43fc35aOJdTRw5y6dwZ8hQsbFkfHxeLp5f3ff2dXVx4/9OxHN2/h/lfTKB93yEA7Ny0nsr1Gtks9989yWvAz9ubvDlz4OriQt6cOXBzdeHm7Si7jIJ6e3kTE/NXsWmYzbi43PnI9Pa2bouJicHXx5fwBQvAZGLb9u0cOXqUQZ98wqjhw8iaxb5/DMj90kTxExAQQIcOHZgwYQKDBw8GID4+nqZNm5IhQwbq1KlD48aNOXHiBJ9//rl9w8pT9VKrN4E7x80/eq8DUbdv4+HhwZED+6j/wstWfQ/t3c3cGdPoMWAIWbLnACBbjpzEx8Vx+eIFcuTKzdGD+6lau57N9yO4cEE2795LjXKlOXD8JAUD/joE9HLtGrxcuwYAP/62hTOXLtOgcgVWbd7GtRs3adWwHu5urjiZTDg72XZwtmLDV4A7c36++aw/cdFRuLp7cP7EYUrXtJ5w/kv4TAqFlCUw9FnOHjlI9oD8Ns36MCWKFuG3bdupXaUy+w4dplC+R086DSlWlPnfL+fVJi9w7foNYuPi8fO1/Ydu/VdaAncOH47u352YqNu4eXhw6vBBqj33glXfJV9/QYlyFSlUNBh3Dw9MJpOl7fypE+QrHIS9PMlrYPPufSz6ZS3N6tYi4lYkcfEJ+PncX/DZQsmQEqz/7Tfq1q7N3n37KFSokKWtQP78nD17jluRkXh5erJz125eb9WK2rVqWvq079yFvj17qvBJo9JE8QPw9ttvs3TpUsaOHQvAtm3buHTpEt9//z2urq4AbNy40a7D0GI7Li4uNH/rHcZ+MgCzYVClVl0yZclK1O3bzJoyjs69BjBv5uckJyXx5cTRAOTMHcAbHbryZqf3+GLscDAMCgUVJeTPeUG2VCU0hB0HDtHls9FgGPR66zV+2bqd2Ph4qzkO96pauiTDZ87hveFjSU5OpvOrr+D25+++rTk7u1D1xZYsmTYKwzAoVr4qPhkzERcdxS/hM2nUtiuVGjdj9bwZ7P1tDa5u7tRu0cYuWR+keoXy/L5rN+179cUwDPq/14VV69YTGxvHiw0eXAxXLleWXfsO0O6DXhiGwQcd3rGMVNiDs4sLjV59ixmjwjDMBmWr1iRDpizERN1m4cwptO7ai0p1GrJ41jR+WboAk8lEkzfeASAq8hbuHp5WxZCtPclroGLJYPYcPUan/43EbBi816q5zf8AuKtG9eps/f132r7bHgyDj/r358dVq4iJieXlJi/SvVtXunbvgWEYPN+4Edmz2e/MTHl8JsNO1UStWrXo2LEjzZo1s6zbvHkzb731FgBz5szh9ddfZ/To0YSEhLB582bCwsLw8PBg69at9ogsT2DDvuOP7pRGVQ0uxPn1q+wd44n5V6vHpJWb7B0jVTo3rMS1w/sf3TGNyhpUnMWb99o7xhN7qWKJdP0agDuvg8gI+565lxp+WTRx+mmw+4Tne1WsWJHGjRsDUKpUKbp06cInn3zCCy+8wMKFCxk0aBA3b97kwoULj9iSiIiIyIPZ7bDXmjVrHrh+1KhRjBo1CoAuXbrQpUsXq/aXX375QXcTERERSZE0NfIjIiIi8rSp+BERERGHouJHREREHIqKHxEREXEoKn5ERETEoaj4EREREYei4kdEREQcioofERERcSgqfkRERMShqPgRERERh6LiR0RERByKih8RERFxKCp+RERExKGo+BERERGHouJHREREHIqKHxEREXEoKn5ERETEoaj4EREREYdiMgzDsHcIEREREVtxsXcAkbTs0o7N9o7wxHKWqcj2I6ftHSNVygbmY/PBk/aO8cQqFi3Ahn3H7R3jiVUNLsT59avsHSNV/KvV49qhvfaO8cSyFilh7wj/STrsJSIiIg5FxY+IiIg4FBU/IiIi4lBU/IiIiIhDUfEjIiIiDkXFj4iIiDgUFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBR9vYWIiIikyuN8DYp/tXpPMUnKqPgRERGR1DGZ7J3gsaj4ERERkVQxOaWv4kdzfkRERMShaORHREREUseUvsZSVPyIiIhI6qSzOT82L9WCgoIICgri7Nmz97XNnTuXoKAgxowZY+tYIiIi8oRMJqcU39ICu6RwdXVlzZo1961fvXo1pnRWPYqIiDg8J1PKb2mAXYqfsmXL3lf8REVFsXPnTooVK2aPSCIiIuIg7DLnp3bt2gwbNozbt2/j6+sLwLp16yhbtiyxsbFWfb///nvGjx/P1atXqVOnDoZhUKBAAbp27WqP6CIpYjabGTPza46dPoubqws932lLQM4clvbVm7bw3Q+rcHJyolDeAHq0eQMDGPHFTM5evIiTkxN92rfDP0d2u+3DH9s2s3juHJycnaletz616je0ar925Qqfjx+FOTkZwzB4u0t3cgfkASA+Lo6hH/Xh3a7vkztPXnvEB2Dnti18P//OPlStXZ8a9Z6zao+4eoUvJ4wmOTkZgLc6dSOXfx5L+8xJ4/D29aX5G21tmvuuXb9vZfmCb3FydqZKrXpUq9vAqj3i6hW+mjSWZHMyGAZvdOhGTv8ATh47wvyvvsAwDDJkzES793ri6uZm0+xms5lxc+Zz/Nx5XF1c+PDNVvhnz2ZpP3TyNFPmL8IAMvv50a/dG7g4Oz/0PrbOP3LqFxw7dRo3Vxf6dOlIQK5cVn3i4uPp/tEQ+nbtRL4AfwC+/m4RG7dtJykpiZeeq8/zdWvbI77tpZHDWSlll7SFChXC39+f9evXW9b98ssv1KlTx6rf9u3b6devH23btmXRokV4enqycuVKW8cVeWwbt/9BQmIiU4YM5N1XmzF5zjxLW3xCAl/OX8jYAb2Z/PEAomJi2bxzN5t27ARg0uABtG36EpO+mWuv+CQlJfHN9Gn0+WQoA4eO5NcfV3LzxnWrPt/N+Yp6jV5gwNCRvNi8JeFfzwDgxNEjfNL3A65cvGiP6BZJSUnMnTGNDwd/St+wEaxbdf8+LPr2a2o3fIG+/xtB46Yt+G72TEvbrz+t4NyZk7aObZGUlET4V5/T46Mweg0Zxvqff+DW3/IvmTebWs81pteQYTR8uQUL53yFYRh8PWU8b3XuQZ//jaR4aBkirl6xef6Nu/aQkJjIxL4f8M7LLzBl/mJLm2EYjPp6Lr3avM743j0oF1yUSxHXH3ofW1u/dRsJiYl8PvxTOrzxOhNmzLJqP3j0GJ36DuT8pcuWdX/s3ce+Q4eZ+lkYE//3MVeuXbN1bLsxmUwpvqUFdivVatWqZTn0lZiYyMaNG6lVq5ZVn7lz51K/fn1atWpFoUKFGDx4MDlz5rRHXJHHsufwUZ4NKQFA8WcKc/jEXx+iri4uTPp4AB7u7gAkJyfj5upK1XJl+LDdWwBcvhZBJj8/m+e+68LZM+TIlRtvH19cXF0JLFacw/v3WfV5rW17SpUrD9zZB1fXOyMLiYmJ9Og3yDIKZC8Xz50h+z378EzRYI4csN6HV9u8Q8myzwLW+3Ds0AGOHz5EjXoN79uurVw8d5bsOf/KX7hocY4e3G/Vp/mb7ShR5k5+szkZV1dXLl84j7evL6uXL2H4wF5ER90mp3+AzfPvO3qCcsF3pjEUK1SAw6fPWNrOXr6Cn483363+le4jxnE7Ooa8OXM89D62tufAISqElgIgOCiQQ8dOWLUnJiUxtG8v8gXktqzbunM3BfPlpe/Q4fQK+4xKZcvYMrI8BrsVP7Vr12bDhg0kJSWxZcsWChcuTNasWa36HD58mBIlSliWXVxcCA4OtnVUkccWExuLt5eXZdnJyYmkPw+tODk5kTlDBgAW/vQzsfHxlC1RHAAXZ2c+nfIF42Z9Q43y5Wwf/E+xMTF4enlblj09vYiJjrbq45shAy4uLlw4d5ZvZ3zOyy1fByCoWHGyZLPf4bq7YmNi8LpnHzw8PYmN+ds++N3Zh4vnzxL+1XRefPU1bl6PYMm8b2jdvrOtI1uJi43B857fIQ8PT2L+If+l8+eYP+tLXmj+Grdv3+L44YPUbNCY9wd9yqE9uzm4Z5eN00NMXBzenh6WZWcnJ8vhxcioaPYfP8mLNaoyskcX/jh4mD8OHn7ofWwtOiYWb++/Hn/ne17DACFFi5Ajm/Vn1q3ISA4dO05Yrw/o2fFdPh49HsMwbJbZnkxOzim+pQV2u85P6dKlcXZ2ZseOHfzyyy/UrVv3vj7Ozs73/eI4yi+SpG9enp7ExMVZlg3DwMX5rxe92Wxm6tz5nL14iU+6d7EaCu7X8R0ibjaj40efMGv4p3h6uNss9/zZMzlyYD9nTp2kUGCQZX1sbAxePt739d+/ZxdfTZlAx/d7232k566Fc77iyIH9nDt9koKBRSzr42Jj8fL2ua//wb27+XraRN7t3pNc/nn4efkSbkdGMnrIQG7dvEFCfDy5/AOoWts2X8a4+NtZHD104E7+Z/56DuLiHpz/0N7dfPPFZNp1+4Cc/gEY5wyy58xlmWtVPLQMp04cpWhIKZvkv8vLw4PYuHjLstls4Pzna8DP2wv/7FnJn/vOHJpywUU5cvrsQ+9ja95ensTE/vUaNhtmq9fwg2Tw9SVfgD+urq7kC/DH3c2Vm7ciyZQxw9OOK4/JbsWPk5MTNWrUYM2aNfz6669888039/UpXLgw+/b9NUydnJzMwYMHKVKkyH19RdKSEkGF2fTHLmpVeJb9R49RII/1YYeRX36Fm4sr/3u/G05OdwZgf9rwG1ev3+D1Fxvj4eaOyWTCycanhTZv3Qa4M9+kV6d2RN2OxMPDk0P799Lo5WZWfffv2cXsL6bQ6+NPyZY9x4M2ZxevvPYWcGcf+nd9l6jbt/Hw8ODw/r081+QVq74H9+5mzvQpfPBRGFn/3Ie6jZtQt3ETADb8soqL58/ZrPABeKnVm5b8H73XwZL/yIF91H/hZau+h/buZu6MafQYMIQsf+bPliMn8XFxXL54gRy5cnP04H6b5r8ruHBBNu/eS41ypTlw/CQFA/6aLJwrW1Zi4+I5f+Uq/tmzsffocRpWqUjubFn/8T62VqJoEX77fTu1q1Ri3+EjFMr36In7IcWKMH/ZSl598XmuXb9BbFw8fr73F6z/Sf/Se5XZbGbw4MEcPnwYNzc3wsLCyJcvn6V9+fLlzJo1C2dnZwIDAxk8eLDlPfRx2PUKz7Vr16ZXr17kyZOHPHnu/6vx9ddfp3Xr1pQvX55y5coxZ84czp8/n2YmTIn8k6ply7B97346DQrDMAz6tH+bn3/bTGxcPEUK5mfl2g2EBAXS43/DAHilQT2qlSvLZ9Om03XIpyQlJdO1dSvcbXyGzl0uLi683q49wz7qh9kwU71uAzJnyUrU7Ui+mDCGHv0G8c0XU0lKTGTamBEA5PIP4O0u3e2S90FcXFx4tc27jPq4H2azQdU69ciUJStRt28zc9IYuvb5iG+/nEpSUhJfjBsJ3NmHtzq9Z+fkd7i4uND8rXcY+8kAzIZBlVp1LflnTRlH514DmDfzc5KTkvhy4mgAcuYO4I0OXXmz03t8MXY4GAaFgooS8ue8IFuqEhrCjgOH6PLZaDAMer31Gr9s3U5sfDyNq1XmwzdbEfbFV2BA8cIFqBASjNlsvu8+9lK9wrP8vms37Xv1wwD6d+vMqnUbiI2L48X69x+pAKhcriy79h+k3Yd9MAyDD9q3s9vIlc39S2d7rV69moSEBMLDw9m1axefffYZU6ZMASAuLo6xY8eybNkyPD09ef/99/n111+pXfvxz6gzGTY+jhQUFMTMmTOpVKkScXFxVKhQgbffftty6nrr1q0pXbo0PXr0AGDBggVMmjSJGzduUL9+fc6ePUuNGjVo3769LWOLg7q0Y7O9IzyxnGUqsv3IaXvHSJWygfnYfNB+Z1ylVsWiBdiw77i9YzyxqsGFOL9+lb1jpIp/tXpcO7TX3jGeWNYiJR7dKQ24vPv3FPfNUfKf5zMOHTqUkJAQGjVqBEDVqlXZsGEDcGdU6Pr165b5wd26daN58+ZUqVLlsfPafOTn8OHDlv97eHiwa9cuq/bZs2db/r9nzx7KlCnD2rVrLesaNWpEtmz2ue6DiIiI3O9xJjKHh4cTHh5uWW7RogUtWrQA7lzw2Mfnr0OFzs7OJCUl4eLigpOTk6XwmT17NjExMVSuXPmJ8qbpLzbduXMns2fPZvjw4WTLlo0VK1Zw8eJFqlatau9oIiIi8gTuLXb+zsfHh+h7ziw1m824uLhYLY8YMYKTJ08yYcKEJ54Gk6aLn9dee41z587RtWtXbt++TdGiRfniiy808iMiIpKW/EsTnkuXLs2vv/5Kw4YN2bVrF4GBgVbtH330EW5ubkyePPmJJjrflaaLHxcXF/r370///v3tHUVERET+wb/1be1169blt99+49VXX8UwDD799FOWLVtGTEwMwcHBfPfdd5QtW5Y337xzVuQbb7zxwEvlPEqaLn5ERETEcTg5OTFkyBCrdYUKFbL8/9ChQ//Kz1HxIyIiIqmTzi5Bo+JHREREUiW9XX9PxY+IiIikzr8058dW0ldaERERkVTSyI+IiIikjo2/hzC1VPyIiIhIqvxbp7rbSvpKKyIiIpJKGvkRERGR1HmM7/ZKCzTyIyIiIg5FIz8iIiKSKrrOj4iIiDiWVHzJqD2kr7QiIiIiqaSRHxEREUkVk0Z+RERERNIujfyIiIhI6qSzixyaDMMw7B1CRERE0q+bF86muG/G3HmeYpKU0ciPyH/YKyO+tXeEVFnYsxXn16+yd4wn5l+tHjUHfWnvGE/s14/fTtePP9x5Di5eu2HvGE8sV9ZM9o7wn6TiR0RERFInnR32UvEjIiIiqaOLHIqIiIgjSW/f6q7iR0RERFInnY38pK9STURERCSVNPIjIiIiqWJySl8jPyp+REREJHU050dEREQciub8iIiIiKRdKn5EREQkVUwmpxTfHsZsNvPRRx/RokULWrduzenTp63a16xZwyuvvEKLFi2YP3/+E+fVYS8RERFJnX9pwvPq1atJSEggPDycXbt28dlnnzFlyhQAEhMTGTp0KN999x2enp60bNmSmjVrki1btseP+6+kFREREUmlHTt2ULVqVQBKlSrFvn37LG3Hjx8nb968ZMiQATc3N8qUKcP27duf6Odo5EdERERSxXByTnHf8PBwwsPDLcstWrSgRYsWAERFReHj42Npc3Z2JikpCRcXF6KiovD19bW0eXt7ExUV9UR5VfyIiIiIzdxb7Pydj48P0dHRlmWz2YyLi8sD26Kjo62Kocehw14iIiKSKsnmlN8epnTp0qxfvx6AXbt2ERgYaGkrVKgQp0+f5ubNmyQkJLB9+3ZCQ0OfKG+aLX6WL19OUFAQM2bMsFrfunVrxowZY6dUIiIi8nfGY/x7mLp16+Lm5sarr77K0KFD6du3L8uWLSM8PBxXV1f69OnD22+/zauvvsorr7xCjhw5nihvmj3stXz5cvLly8fixYtp27atveOIiIjIU+bk5MSQIUOs1hUqVMjy/1q1alGrVq3U/5xUb+EpuHXrFhs3bqRLly4cOXKEAwcO2DuSiIiI/APDSPktLUiTIz8//fQTbm5uNGzYkEmTJrFo0SKKFStm71giDqVsIX+aVQwm2TBYs/c4q/cct2rP6utFt0YVMWEiKi6eMcs3kZCUTNWi+XmhXBHMZoM1+47z065jNs9uNpsZN2c+x8+dx9XFhQ/fbIV/9jvXArl+K5JPPp9p6Xvs7HneefkFXqhRhXeHDMPb0wOAnFmz0LvN6zbPfq+KgXl4o0YoyWaDH3YeYcWOw1btOTP60Oel6phMcPlmFKOWbSQ+MZk6JQrRrFIwZuPO/b7//ZDNsz/sOQA4dPI0U+YvwgAy+/nRr90buLm6AnAj8jYdwoYzokdn8ubKafPsd23auIFZM2fg7OxMw8aNafxCE6v2mzdvEjb4I+Lj48maNSu9+w/Ew8ODQwcPMGn8OAwMMmfOQv+PBuPu7m6fnbARc1qpalIoTY78LFu2jGrVquHi4kLt2rVZvnw5iYmJ9o4l4jCcnUy8VbM0Qxas4aO5q6kbUpiM3h5WfRqXLcJvh84wcN5qzl67Re0Sd4am36wRysfz19D/2595vmxRvN1dbZ5/4649JCQmMrHvB7zz8gtMmb/Y0pY5gx9jer7HmJ7v0e7lF3gmbwCNqlUi4c/3mLtt9i58nJ1MdG5QgZ5f/0j3mStoXCaITD6eVn061HuWZdsP8t6MFew6dZFmFUvcWV//WT78+ge6frmc5hVL4OPhZvP8D3sODMNg1Ndz6dXmdcb37kG54KJcirgOQFJSMqNnz8Pd1fa/N/dKSkpi4vhxjBwzjnGTprBs6VIiIiKs+nw980tq163HhCnTKBwYxLIlizEMgxGfDaV3/wFMnPI5z5avwOVLl+y0F7ZjGEaKb2lBmit+Ll++zPbt26lTpw4A9erV48aNG6xbt87OyUQcR0CWDFy6eZvo+ESSzGYOnr9KUX/rq6ieunLD8qHq6e5KsvnOaRynr97Ey90VVxcnTCYeMb3x6dh39ATlgu+MFhcrVIDDp8/c18cwDCbMXUD311vg7OTE8bPniUtIoOeYSbw/cjwHjp+0dWwr+bJl5Pz1SKLiEkhKNrP3zGVC8ua4r8/Wo+cA2HfmMiX+bD9x+Tre7m64uThDGnwOzl6+gp+PN9+t/pXuI8ZxOzqGvDnvZJ/63WJeqF6FLBkz2CH1X06fOol/QAC+fn64urpSIqQke3fvsuqzd89unq1QEYDyFSqyY/vvnD1zhgwZMvBdeDjvde7I7duR5M2Xzw57YFtmw0jxLS1Ic4e9Vq5cibOzM9WrVwegZMmSZM+enSVLllgKIhF5ujzdXImJ/2u0NS4hES9369GDiNsxvF6tFFWL5sPF2Znw3/YCcObaTYa3bkB8YhJbj56z2o6txMTFWQ5fATg7OZGcnIyz818XYtu0ex/5c+eyfOi6u7nRvF4tGlWtxLnLV+kzfgpffzLA6j625OXuRnR8gmU5Nj4R77+N4By/dJ3KQXn5afcxKhXJh4fbnbf0k1duMK19E+ISE9lw8DTRcQnY2sOeg8ioaPYfP0nXlk0JyJ6dfhOmEpgvD1eu3yCDjw/lgovy7Q+rbJ75XtHR0fh4e1uWvby87rugXnR0DD4+3n+1R0dx69ZN9u3dS7ceHxCQJw99e35AYFARypQtZ9P88nBprvi5e4irfPnylnVms5m1a9dy/fp1OyYT+e9rWSWEIv7ZyJctI0cv/jXE7+HmavVBDPBGjVAm/rCFXacuUrpgbro1rMicDbspUzA3nT7/nrjEJN5rVJGKgXnYfOSsTffDy8OD2Lh4y7LZbNxXxKze8juv1KlhWQ7IkQ3/7FkxmUzkyZkdP28vIm5Fkj1zJlvFBqBtrTKUyJuDgjkyc/D8Fct6T3dXov5WxEz+aSvvNapIrRKF+OPkBW7FxFEwRyYqPJOHVmPDiU1Iot8r1aleLD/rDpyy6X487Dnw8/bCP3tW8ufOBUC54KIcOX2WzXv2YTKZ+OPgYY6dPc/QGbP5X5f2ZM7gZ7Pc0z+fyt49uzlx7DhFi/811zQmJgafv11Qz9vbi5iYGNzdPe60+/jilyED/gEB5C9QAIBny1fgyOFD//niJ40M6KRYmip+Tp06xb59++jbty+VKlWyrL948SLvvvsuK1assGM6kf++uRv3AHfmm4xr2xgfDzfiEpIoFpCd738/aNU3Ki7BUhDdiIrF28ONmPgE4pOSSUhKxmwY3IqJt8t8k+DCBdm8ey81ypXmwPGTFAzIdV+fI2fOUrxQAcvyD79t4eT5C3R/rQXXbt4iJi6OLDb80L1rxpodwJ3n4Ksur+Dr6UZsQhIl8+Vk/p+ja3eVLeTPrLU7OXH5Bs0qBbPj+AWi4xKJT0om/s/n4GZ0HL6etp9s+7DnIFe2rMTGxXP+ylX8s2dj79HjNKxSkVcb/DW632PEOHq83sKmhQ9Au3c7AHfm/Lz52qtERt7C09OLPbt30qJVK6u+wSVC2LJpE881aszWLZsJKVmS3Ln9iY2N4dy5swQE5GHP7l00fP4Fm+6DPaSVuTwplaaKn+XLl+Pn50fLli2tZsYHBgYSGhrK4sWL8b5nGFJEno5ks8FXv/7BwKY1MZlMrNl3nOtRsfh4uNGxfnlGLN3Al79sp13tsjg5mTBhYvrq7VyNjOHn3ccIa1WXpORkLt2M4td9tp87UyU0hB0HDtHls9FgGPR66zV+2bqd2Ph4GlerzM3bt/Fyd8dk+uubqBtWqciwmd/QbdgYTEDPN1+z2yEvuPMcTP5xK8NbN8DJZOKHnUe4djsGX083PnyhKoPCf+HstVv0erEqiclmTl25wdgVm0g2GyzbfojxbRuTlJzMhRu3+XHXUZvnf9Rz8OGbrQj74iswoHjhAlQICbZ5xodxcXGhc9f36NmjO4Zh5rlGz5MtW3YiI28xYuinfDJ0GK3fasPQsE9YsWwpGTJkZMDgIbi6utKrb3/CBg/CMAyCS5SgYqXK9t6dpy6tzOVJKZORhsq15557jgoVKjBo0KD72pYuXUqvXr3ImDEjr776Kj169LBDQpH05ZUR39o7Qqos7NmK8+vtO/cjNfyr1aPmoC/tHeOJ/frx2+n68Yc7z8HFazfsHeOJ5cpq28OuT+rs5YhHd/pTnhxZnmKSlElTIz8//PDDP7a9+OKLvPjiizZMIyIiIv9Faar4ERERkfQnDR1EShEVPyIiIpIq6W3Oj4ofERERSZX0Vfqo+BEREZFUSm8jP2nu6y1EREREniaN/IiIiEiqpLcJzxr5EREREYeikR8RERFJlXQ28KPiR0RERFInvU14VvEjIiIiqZLe5vyo+BEREZFUSWe1jyY8i4iIiGPRyI+IiIikiub8iIiIiEPRnB8RERFxKOb0Vfuo+BEREZHUMdLZV5uq+BEREZE0Ky4ujp49exIREYG3tzfDhg0jc+bMVn2++uorVqxYAUD16tXp0qXLQ7eps71EREQkVQzDSPHtcc2dO5fAwEC+/fZbmjRpwuTJk63az549y/fff8+8efMIDw9n48aNHDp06KHbVPEjIiIiqWI2Un57XDt27KBq1aoAVKtWjc2bN1u158yZk+nTp+Ps7IyTkxNJSUm4u7s/dJsmI71N0RYREZE0ZcfRMynue+yPzYSHh1uWW7RoQYsWLQBYsGABs2bNsuqfJUsWPvroIwoVKoTZbKZGjRqsX7/+vu0ahsHw4cOJjo5myJAhD82gOT8ikqZ9OGulvSM8sZFvNuT8+lX2jvHE/KvVY+jCX+0dI1X6vlKTmxfO2jvGE8uYO4+9I6TI44yj3Fvs/F2zZs1o1qyZ1bouXboQHR0NQHR0NH5+fvfdLz4+nn79+uHt7c2gQYMemUGHvURERCTNKl26NOvWrQNg/fr1lClTxqrdMAw6depEUFAQQ4YMwdnZ+ZHb1MiPiIiIpMrTvMJzy5Yt6d27Ny1btsTV1ZVRo0YBMHPmTPLmzYvZbGbbtm0kJCSwYcMGAN5//31CQ0P/cZsqfkRERCRVnubsYU9PT8aPH3/f+jZt2lj+v3fv3sfapg57iYiIiEPRyI+IiIikSrLZbO8Ij0UjPyIiIuJQNPIjIiIiqZLerhio4kdERERS5Wme7fU06LCXiIiIOBSN/IiIiEiqGKSvkR8VPyIiIpIq6eyol4ofERERSR3N+RERERFJwzTyIyIiIqnyON/qnhao+BEREZFUMaev2kfFj4iIiKROehv5SXNzfpKSkpg8eTJ169YlODiYqlWrMnDgQCIiIgBo3bo1Y8aMsXNKERERucswjBTf0oI0N/IzatQo1q9fz+DBg8mfPz8XL15kxIgRvPPOOyxcuJAJEybg6upq75giIiKSTqW54mfRokV8/PHHVK5cGQB/f39GjhxJnTp12L17N6VKlbJvQBFJN4oFZKduycIkmw1+P3aOrUfPWrW/UK4o/pn9APD1dCc2IZEJKzc/8n62YDabGTdnPsfPncfVxYUP32yFf/ZslvbVW35n/s9rcDY50aBKBV6sUZWkpGSGz5rD5WsRJCQl8XqjBlQuVcLm2e8qnCsLVYrkx2wY7D51kd2nLlq1e7i60L5+ea5GRgNw5MI1th87R3DeHJQPzEt8YhJ7Tl9iz9/uZwtms5nhY8dz9Phx3Fxd6dfzA/L4+1vaN2zazJdfz8bZ2Znnn2tAk8aNSEpK4uOhw7h4+TJOTk70+/B98ufNa/Ps9qA5P/+CLVu2ULduXZydnQHIkycPK1asICAggNatW1O6dGl69Ohh55QikpY5mUy8UK4o41b8RkJSMl2eq8iBs5e5HZdg6fP97wctfbs8V5EFm/al6H62sHHXHhISE5nY9wMOHD/JlPmLCevyrqV96ndLmPFxPzzd3Wnz0f+oVa4MG3ftwc/bm35vv8GtqGjafzLMbsWPk8lEnZDCfLVmBwlJybxRozTHLkYQHf/X45gzky8Hzl7h591HLes83VypVqwgM375nbjEJFpWLcXpKze4FRNn0/zrNv5GQkICX06awN4DBxg3eSoj//cJcGd6xthJU5g5dRKeHh680/U9qlaqyL4DB0lOTmb6xPFs3b6DKdNnMGzIYJvmtpe0cjgrpdLcnJ833niDuXPnUrNmTQYMGMCKFSuIjIykcOHCeHh42DueiKQTOTL6cO12DLEJSSSbDU5euUGBHJkf2LdK0fwcvnCVSzdvP9b9nqZ9R09QLrgYAMUKFeDw6TNW7QX9cxMdE0tCYiIGYDJBjTKhtG3SyNLH2cl+b/FZfL24ERVLXGISZsPgbMQt8mTNYNUnZ0Zfcmb05bVqoTQpXxxvDzcyentw+dZt4hKTALh4I5Lcf47O2dLuvfuo8Gw5AEoUK8ahI0csbSdPnyHAPzd+vr64urpSskQwu/bsJW+eAJLNZsxmM9HR0bi4pMnxBSENjvx07tyZAgUK8O2337Jo0SIWLFiAu7s73bp1o127dvaOJyLphLurC3EJiZbl+MQkPNzuf8tzdjJRITAP41dseqz7PW0xcXF4e/71B5+zkxPJycmWEfEC/rnoEDYCD3c3qpYuiY+Xl9V9P576JW2bNLZ57rvcXV0sBQxAQlIy7q7Wj2PE7Rgu3TzJqSs3KJ4nB/VKPsMPfxwmm583Xu6uJCQlkz97Jq5Hxdo6PtExMfh4e1uWnZycSEpOxsXZmeiYaKs2L08voqKj8fL05OKlSzR/sw23bkUy6tMwm+e2l+R0NvKT5oofgIYNG9KwYUMiIyPZtGkT4eHhjBgxgvz589s7moikcQ1CAymQPRO5Mvly5upNy/o7RU3Sff2fyZWVE5evWz6o4xOTrD6k/+l+T5uXhwexcfGWZbPZsBQ+x8+dZ8ve/cwZOhhPD3c+nT6Ltdt3UqNsKFeu3+CjyV/wYo2q1C5f1ua5qxUrQEDWDGTP4MOF65GW9W4uzlbFEMDpqzdITEoG4PCFq1QtVoC4xCRW7znGyxWCuR0bz6WbUcTGJ2Jr3l5exMTEWJbNZgOXPx9/by9vomP/KshiYmPw8fFh7oKFlC9Xls7vtOPylSt0er8n3874Anc3N5vnl4dLU4e9Dh06RFjYX5Wyn58fDRo0YMaMGQQHB7Np0yY7phOR9ODHnUeY8tNWBof/QhY/bzzdXHF2MlEwR2ZOXb1xX/9ncmXl0PmrluXLN6PImoL7PW3BhQuyde9+AA4cP0nBgFyWNm9PT9xd3XB3c8XZyYlMfr5ExcRwPTKSXmMm8e4rL/JclYo2zwyw/sBJvl2/i/HLfyOTjyceri44mUzkyZqR8xG3rPo+VzqIIP87k7jzZ8/EpZu3MZlM5M7sxzfrdrLs94Nk8fHi3N/uZwshwcXZtHUbAHsPHKBwwQKWtgL58nL23HluRUaSmJjIzt17KVGsGL6+vpYRIT9fX5KTkjAnJ9s8uz3oVPdUSE5OZvbs2TRu3NjqrC6TyYSvry+ZM9v+uLuIpE9mw2DZ7wd5t245TCYT246eIzImHk83V5pXKsGstX8AkD2DNzuOn3vk/WytSmgIOw4costno8Ew6PXWa/yydTux8fE0rlaZxtUq023YGFxdXMidLSv1K5Vn2ndLuB0Tw+zlPzJ7+Y8AfPZeR7uMPJgNg1/2HOPVKiXBZGLPqYtExSXg4epCwzJFWLRlH2v3naBRmSKULuRPYpKZlX8cwjAMzGaDNrXLkpxsZuvRs8Qm2H7kp0bVKmzb8QftunTDMAwG9u7JT6t/ISY2lpeeb0z3Th14r1cfzGaD559rQPZsWWnZ7BXCho3g3W7dSUpMomO7tnh6eto8uz2kty82NRlppQz7U4cOHdi3bx8ffPABZcuW5ebNm6xevZp58+axdOlSevbsqbO9RBzIh7NW2jvCExv5ZkPOr19l7xhPzL9aPYYu/NXeMVKl7ys1uXnB9pcq+LdkzJ3H3hFSZOnWfSnu+2L54KeYJGXS1MgPwNixY/n888+ZNm0agwYNws3NjXLlyjFnzhxy5sxp73giIiLyN+Z0dqGfNFf8eHh40K1bN7p16/bA9tmzZ9s4kYiIiPyXpLniR0RERNKX9DbnJ02d7SUiIiLpz9M82ysuLo6uXbvSqlUr3nnnHa5fv/7AfmazmXbt2jF37txHblPFj4iIiKRZc+fOJTAwkG+//ZYmTZowefLkB/YbO3Yst26l7LIIKn5EREQkVcxGym+Pa8eOHVStWhWAatWqsXnz5vv6/Pjjj5hMJqpVq5aibWrOj4iIiKTK4xzOCg8PJzw83LLcokULWrRoAcCCBQuYNWuWVf8sWbLg6+sLgLe3N7dv37ZqP3LkCMuXL2f8+PFMmjQpRRlU/IiIiEiqPE7x8+o9xc7fNWvWjGbNmlmt69KlC9HR0QBER0fj52f9RbdLlizh8uXLvPnmm5w/fx5XV1f8/f0fOgqk4kdERETSrNKlS7Nu3TpCQkJYv349ZcqUsWrv1auX5f8TJkwga9asjzz8pTk/IiIikirmx7g9rpYtW3L06FFatmxJeHg4Xbp0AWDmzJn88ssvT5RXIz8iIiKSKk/zm7I8PT0ZP378fevbtGlz37quXbumaJsa+RERERGHopEfERERSZXkdPbdXhr5EREREYeikR8RERFJlac55+dpUPEjIiIiqaIvNhURERFJwzTyIyIiIqli1oRnERERkbRLIz8iIiKSKultzo/JSG9TtEVERCRN+fynrSnu+2798k8xScpo5EdE5CnqMHWJvSM8sakdmnD7xg17x0gV30yZ2H3inL1jPLGSBQPsHeE/ScWPiIiIpEp6O+yl4kdERERSRcWPiIiIOJT0Nn1YxY+IiIikSjqrfXSdHxEREXEsGvkRERGRVNGcHxEREXEoBip+RERExIGks4EfFT8iIiKSOuntsJcmPIuIiIhD0ciPiIiIpEp6u86PRn5ERETEoWjkR0RERFIl2ayRHxEREZE0SyM/IiIikirpbc6Pih8RERFJlad5qntcXBw9e/YkIiICb29vhg0bRubMma36rFu3jkmTJgFQrFgxBg0ahMlk+sdtpunDXkFBQWzatMlq3fbt2wkJCSEsLMxOqURERMRW5s6dS2BgIN9++y1NmjRh8uTJVu1RUVGMGDGCqVOnMn/+fPz9/blx48ZDt5mmi5+/O3ToEB06dKBRo0b079/f3nFERESEO1d4Tuntce3YsYOqVasCUK1aNTZv3mzVvnPnTgIDAxk2bBitWrUia9as940M/V26Oex19uxZ2rVrR6VKlQgLC3vocJaIiIjYzuPM+QkPDyc8PNyy3KJFC1q0aAHAggULmDVrllX/LFmy4OvrC4C3tze3b9+2ar9x4wZbt25lyZIleHl58dprr1GqVCkKFCjwjxnSRfETERHB22+/TVBQECNHjsTZ2dnekUREnroS+XLSqEwQZrPBpsOn2XjwtFV7s0olyJM1AwB+nu7EJCQyfPF68mXLSNNKwZgwERkTx4w1O0hKNts0u9ls5rMRIzh69Ciurq4M7NePPHnyWNrXb9jA9BkzcHZ25oXGjXmpSROSkpL4OCyMixcvkpCYyNtvvUX1atVsmvte27dsYuG33+Dk7ETNes9R57lGD+y3YvFCbt64zmtt3wFgw5rVLF+0ACcnZ2rWa0C9xi/YMrZdPM6cn3uLnb9r1qwZzZo1s1rXpUsXoqOjAYiOjsbPz8+qPWPGjJQoUYJs2bIBULZsWQ4ePJi+i5+oqCjatWvHxYsXmT9/Pm5ubvaOJCLy1Dk5mWhWKZjPFq4jPimJnk2qsefUJSJj4y19Fmzaa+nb88WqfLNuFwCvVw/l81XbuBoZTeUi+cji48XlW1E2zb923ToS4uOZOX06e/ftY8z48YweMQKApKQkRo8bx9czZuDp6cnb775L1apV2bR5MxkzZOCTwYO5eesWr73xht2Kn6SkJGZ9PoWh4ybj4eHBwA/eo2z5imS853BKQnw8U8eN4tjhQ5SvXNWyfvb0aYye9iUeHp70aN+WStVr4vPnyMV/1dM816t06dKsW7eOkJAQ1q9fT5kyZazag4ODOXLkCNevX8fPz4/du3fTvHnzh24zzc/5GTJkCABubm58/vnndk4jImIbuTL6cvVWNDEJiSSbDY5fiqBwriwP7FszuCAHzl3lwvVIcmTwITougdohhXj/hSp4e7javPAB2LV7NxUrVgSgRHAwBw8dsrSdPHmSPAEB+Pn54erqSsmSJdm1axd1atWiw7vvWvq52HGU//zZ0+TM7Y+Pry8urq4EFQ/m4P69Vn0SEhKoXrseL7/aymp9vgIFiYmOJiEhAQxD0zRSqWXLlhw9epSWLVsSHh5Oly5dAJg5cya//PILmTNn5oMPPqBdu3Y0b96cunXrEhgY+NBtpvmRnwwZMjBjxgxWrlxJWFgY9erVo1SpUvaOJSLyVHm4uRCbkGhZjktIwtPN9b5+zk4mqhbLz2eL1gHg4+lGwZyZCf9tD5dvRdH5uQqcvnqTw+ev2Sw73Dk84ePtbVl2cnIiKSkJFxeX+9q8vbyIiorCy8vLct/effvSsX17m2a+V2x0DF73ZPT09CQm2rqI9PH1pWSZsqz9+Uer9Xny56d31454eHjwbOUqePv42CSzPT3NU909PT0ZP378fevbtGlj+X+jRo1o1OjBhyUfJM0XP7179yZTpky0atWK5cuX06dPH5YuXYq7u7u9o4mI/OteKFeUwrmy4J/Zj5NX/jpd9+/F0F1FA7Jz7GIEcQlJAETFJXD1VhQXb9yZFHrg7BXyZcto8+LH29ubmJgYy7JhNuPi4mJpi76nLTomxnJY6NLly/Ts3Zumr7xCg/r1bZoZYN6sGRzav4/TJ0/wTJEilvWxsbF4ez+6iDl98jh/bNvKpK++wcPDk/EjhrJ5wzoqVq3+NGPbXXq7yGGaP+x198ViMpkICwvj/PnzjB071r6hRESeku9/P8jo7zfS8+sfyJ7BGy93V5ydTBTOlZUTl6/f17+Ifzb2nblsWb4WGY27qwvZ/O6MWhTOmYUL12/fd7+nrWRICL/9eZ22vfv2UbhQIUtbgQIFOHv2LLdu3SIxMZGdO3cSEhxMREQEXbp1o2vnzrz4/PM2zwzw6pttGTx8NF/M/Y5LFy4QdTuSpMREDu7bQ2DRYo+8v5eXD25u7ri5uePk7EyGjBmJvm37x9/Wnuap7k9Dmh/5uVehQoVo3749kyZNol69eoSGhto7kojIU2E2GyzYtI9ujSphMpnYdOg0N6Pj8HJ3pXX1UKat2gZAjow+bDly1nK/ZLPB7LU7ebtOWQBOXL5uVRzZSs0aNdj6+++0fecdDMNg0IAB/PjTT8TExvJykyb0eO89unbvjtls5oXnnyd79uyMHD2a27dvM33GDKbPmAHA+DFj8PDwsHl+FxcX3ninA//r3wezYaZmvQZkzpqNqNuRTB07ig8HfvzA+2XLkYM6DRsz8MP3cHFxJUeuXNSoa/sRLFt7moe9ngaTkYbHqoKCgpg5cyaVKlWyrEtMTOSll14iKSmJJUuW2OVFISKSUh2mLrF3hCc2tUMTbj/iSrlpnW+mTOw+cc7eMZ5YyYIB9o6QIgO+XZXivmGt6j3FJCmTpkd+Dh8+fN86V1dXli9fboc0IiIi8iBpeBzlgdJ08SMiIiJpXzqrfVT8iIiISOqktzk/Kn5EREQkVdLbYa80f6q7iIiIyL9JIz8iIiKSKub0NfCj4kdERERSx3iqX23679NhLxEREXEoGvkRERGRVElvE55V/IiIiEiqaM6PiIiIOJT0NvKjOT8iIiLiUDTyIyIiIqmiKzyLiIiIQ0lvh71U/IiIiEiqpLPaR8WPiIiIpE56O+ylCc8iIiLiUDTyIyIiIqmiOT8iIiLiUNJZ7aPiR0RERFInvc35UfEjIvIUTe3QxN4RUsU3UyZ7R0i1kgUD7B3hPy+9fau7ih8REflHn373q70jpEq/pjWJvB5h7xhPzC9zFntH+E9S8SMiIiKp8jS/2DQuLo6ePXsSERGBt7c3w4YNI3PmzFZ9vvzyS1asWIHJZKJDhw7UrVv3odvUqe4iIiKSKoZhpPj2uObOnUtgYCDffvstTZo0YfLkyVbtkZGRzJ49m3nz5jFjxgw+/fTTR25TxY+IiIikWTt27KBq1aoAVKtWjc2bN1u1e3p6kjt3bmJjY4mNjcVkMj1ymzrsJSIiIqmS/BjHvcLDwwkPD7cst2jRghYtWgCwYMECZs2aZdU/S5Ys+Pr6AuDt7c3t27fv22auXLlo1KgRycnJtG/f/pEZVPyIiIiIzdxb7Pxds2bNaNasmdW6Ll26EB0dDUB0dDR+fn5W7evXr+fKlSv88ssvALz99tuULl2akJCQf8ygw14iIiKSKk9zzk/p0qVZt24dcKfQKVOmjFV7hgwZ8PDwwM3NDXd3d3x9fYmMjHzoNjXyIyIiIqnyNL/eomXLlvTu3ZuWLVvi6urKqFGjAJg5cyZ58+aldu3abNq0iebNm+Pk5ETp0qWpXLnyQ7dpMtLbF3KIiIjN6Do/9pVervPTetz8FPed/V7zp5gkZTTyIyIiIqnyOBOe0wLN+RERERGHopEfERERSZX09sWmGvkRERERh6KRHxEREUmV9DXuo+JHREREUim9nTiu4kdERERSJb3N+bF78RMZGcmUKVNYtWoV165dI2fOnLzyyiu0adMGV1dXe8cTERGRR0hntY99i5+bN2/SokULsmTJQlhYGAEBARw4cICwsDCOHDnCyJEj7RlPRERE/oPsWvyMHDkSV1dXZs6cibu7OwB58uQhU6ZMtG7dmtatW1OyZEl7RhQREZFH0JyfFEpISGDFihX06tXLUvjc9eyzzzJr1iwCAwM5fvw4Q4cOZceOHSQlJREcHMyQIUN45pln7JRcRERSonCuLFQpmh+zYbDn1EV2nbxo1e7h6kKHBuW5GnnnG7uPnL/G78fOEZw3BxUC8xKflMSeU5fYferigzb/VJnNZoaNGMnRY0dxdXVjQN++5MkTYGlfv2Ej02fOwMXZmecbN+alF18kOTmZ/w39jNNnzuDs7MRH/fsTEBDwkJ/y35He5vzY7To/Z86cISYmhhIlSjywvUKFCnh4eNCpUydy587N0qVLmTdvHmazmeHDh9s4rYiIPA4nk4k6JQszb8Nuvlm7k1IFcuPt7mbVJ2cmXw6cvcKcdbuYs24Xvx87h6ebK9WLF+SbdTuZvXYnxfPmIIOXh83zr12/nviEBGZ88QVdOnVk7ITxlrakpCTGjBvHxLFjmTZ5MouXLuVaRAQbNm4E4MvPp9H+nXcYM378P23+P+dpfqv702C3kZ+7Xzfv6+v7j31iY2Np2rQprVq1wtvbG4CXXnqJadOm2SSjiIg8mSx+XtyIiiUuMQmAcxG3yJM1A4fOX7X0yZnJlxwZfXm9eijR8Qms2nUUP093Lt+6bbnfxRuR+Gf241ZMnE3z7969m0oVygNQIjiYgwcPWdpOnjpFQEAAfn5+AJQKKcmuXbupU7sWVf78NvGLFy+ROXNmm2a2p3T21V72K34yZcoEwK1bt/6xj5eXF61atWLp0qXs27ePEydOcODAATJmzGijlCIi8iTcXVyI/7OAAUhITMbd1fojJyIyhks3TnLqyg2K58lB/VLPsHLHYbL6eePt7kp8UjL5s2Xi+u1YW8cnOjoGbx8fy7KTszNJSUm4uLgQHR2Nzz1tXl5eREVHAeDi4sLgIZ+wdt06Pvv0fzbPLSljt+Inb968ZMyYkb179xISEnJfe/fu3alZsyZTp04lQ4YM1KlTh8aNG3PixAk+//xzOyQWEZFHqV68AAFZM5A9gw8Xrkda1ru5OlsVQwCnr94gMSkZgMMXrlKteAHiEpNYvfsYL1cM5nZsPJduRhETn2jTfQDw9vYiJjrGsmyYzbi4uPzZ5k1MzF9tMTEx+N5TDA3+aCDXIjrRpl075n/7LZ6enrYLbidp5XBWStltzo+zszONGjXim2++ISEhwapty5Yt/PDDD5w9e5ZLly4xe/Zs2rVrR6VKlbhw4UK6e5BFRBzFuv0nmbNuF+OW/UYmb088XF1wMpnIkzUj5yKsR/oblgmiSEA2APJnz8SlG7cxmUz4Z/Zj9tqdfL/tIFl8ve67ny2UDAnht82bAdi7bx+FChWytBXIn5+zZ89y61YkiYmJ7Ny1ixLBJVj5ww/MnPU1AB4eHjiZnHBycoyv0NScn8fQpUsX1q5dS5s2bejatSu5c+dmx44dDBs2jJdffpnKlSszceJEfv75Z0JCQti8eTNz5szBw8P2k99ERCTlzIbB6j3HeLVqSUwmE3tOXSQqLgEPVxcalS3Cws37+HXvCRqXLULpQv4kJplZseMQhmGQbDZoW7ssSWYz246cJTbB9iM/NapXZ+u232n7zruAwUf9+/PjT6uIiY3h5SZN6N6tG117dMcwGzzfuDHZs2ejZo0aDAn7H+927EhSUhLvd3/vvrOZJW0wGXYuwy5fvsykSZNYt24dN27cICAggKZNm9K6dWtcXV2ZOHEic+bMIT4+nsDAQJo3b07fvn359ddfyZ07tz2ji4j853363a/2jpAq/ZrWJPJ6hL1jPDG/zFnsHSFFGoR9leK+Pw5466nlSCm7Fz8iIpJ2qfixLxU/T4fdv9tLRERE0rf0No6i4kdERERSJb1d58cxpqGLiIiI/EkjPyIiIpIqOuwlIiIiDiW9fbGpih8RERFJlXRW+2jOj4iIiDgWjfyIiIhIqvz68dv2jvBYNPIjIiIiDkXFj4iIiDgUFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY+IiIg4FBU/IiIi4lBMhpHevotVRERE5Mnpi01FROQ/7fqtSHtHeGKZM/jZO8J/kg57iYiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY+IiIg4FBU/IiIi4lBU/IiIiIhDUfEjIiIiDkXFj4iIiDgUFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQXOz5w4OCgqyWM2bMSO3atenXrx8+Pj52SiUiIiL/ZXYf+Rk7diwbN25k/fr1TJs2jX379vHZZ5/ZO5aIiIj8R9l15AcgQ4YMZMuWDYAcOXLwxhtv8OmnnxIWFmbnZCIi4sg2bFjPzOnTcXZ2ofELz/Nik5es2m/evMmggQOIj4sna7asDPhoEB4eHqz66SfC583FycmJwoWfoWfv3jg52X2sQe6R5p6NzJkzWy1HRUXRv39/KlasSHBwMPXr1+enn36yUzoREXEESUlJjBszhrETJjJ52jSWLl5MxLVrVn1mTJ9Ovfr1mfrFFwQGBbF40SLi4uL4fOoUJk2ZyhdfziA6OorfNm6w017IP0lTxc/169eZPXs2L7zwgmXd0KFDOX78ODNmzGD58uWUK1eOgQMHkpCQYMekIiLyX3bq5EkCAgLw8/PD1dWVkJKl2LVrl1Wf3bt3UaFCRQAqVqzE9t+34ebmxufTv8TDwwOApKRk3NzcbR1fHsHuh706dOiAs7MzhmEQGxtLxowZGTBggKW9TJkyvPHGG5bJ0W3btmXBggVcvnyZPHny2Cu2iIj8h0VHR1udeOPl7UV0VNR9fbz/7OPl7UVUVBROTk5kzpIFgAXh4cTGxvBs+fK2Cy4pYvfiZ8iQIYSGhgJw69Ytli1bRosWLViwYAEFChSgSZMmrF69mgULFnDixAn2798PgNlstmdsERH5D5o2ZQq7d+/i2LFjFC9e3LI+JjoGH19fq77e3t7ExMTg4eFxp93nTrvZbGbihPGcPXOGocOGYzKZbLoP8mh2P+yVPXt28uXLR758+QgJCaF///5kyZKF+fPnA9CrVy8+++wzfH19admyJdOmTbNzYhER+a9q37Ejk6dOY+WPP3Hu3Dlu3bpFYmIiu3btJLhECau+ISEl2bzpNwA2b95EqVKlABg29FMSEhIYNmKk5fCXpC12H/n5J8nJyURFRbF8+XLmzp1rGR1at24dAIZh2DOeiIj8h7m4uNCte3d6dOuK2TBo/PzzZM+enVu3bjH0f2F8NnwEbdq2ZcjHg1m6ZAkZM2bk40/COHzoEMu+/56SpUrRpVNHAJq3eJUaNWvaeY/kXnYvfm7dusXVq1cBiIuLY+HChZw+fZoGDRrg7u6Op6cnq1atIlu2bJw6dYohQ4YAaMKziIg8VVWrVqNq1WpW6zJkyMBnw0cAkDlLFsaOn2DVHlSkCJu2brNZRnkyJsOOQyh/v8Kzu7s7RYoU4Z133qFu3boArF69mmHDhnH58mUCAgJo06YNEyZM4P3336dJkyZ2SC0iIunJ9VuR9o7wxDJn8LN3hP8kuxY/IiIiT5uKH/k7u094FhEREbElFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY+IiIg4FBU/IiIi4lBU/IiIiIhDUfEjIiIiDkXFj4iIiDgUFT8iIiLiUFT8iIiIiENR8SMiIiIORcWPiIiIOBQVPyIiIuJQTIZhGPYOISIiImIrGvkRERERh6LiR0RERByKih8RERFxKCp+RERExKGo+BERERGHouJHREREHIqKHxEREXEoKn5ERETEoaj4EREREYei4kdEREQcioofERERcSgqfkRERMShqPgRERERh6LiR0RERByKih8RERFxKCp+RERExKGo+BERERGH4mLvALVq1eL8+fP3rX/mmWdYvny5HRKJiIjIf5ndix+APn360LhxY6t1Li5pIpqIiIj8x6SJCsPHx4ds2bLZO4aIiIg4gDQ/5yc8PJzatWsTGhpKy5Yt2bNnj70jiYiISDqWpoufNWvWMG7cOPr27cvixYupVq0ab775JleuXLF3NBEREUmnTIZhGPYMUKtWLa5evXrfHJ/Vq1fTtWtX6tWrx1tvvWVZ37p1aypUqEDnzp1tnFRERET+C9LEnJ8uXbrQoEEDq3UZM2bk+PHjjB49mnHjxlnWJyQkkDNnTltHFBERkf+INFH8ZM6cmXz58t23Pjk5md69e1OlShWr9V5eXraKJiIiIv8xaXrOT4ECBbh06RL58uWz3GbMmMG2bdvsHU1ERETSqTRd/LRp04bZs2ezePFizpw5w8SJE1m4cCEFCxa0dzQRERFJp9LEYa9/0rBhQyIiIpg4cSJXrlyhYMGCTJo0iaJFi9o7moiIiKRTdj/bS0RERMSW0vRhLxEREZF/m4ofERERcSgqfkRERMShqPgRERERh6LiR0RERByKih8RERFxKCp+RERExKGo+BERERGHouJHREREHIqKHxEREXEoKn5ERETEoaj4EREREYei4kdEREQcioofERERcSgqfkRERMShqPgRERERh6LiR0RERByKih8RERFxKCp+RERExKGo+BERERGHouJHREREHIqKHxEREXEoKn5ERETEoaj4EREREYei4kdEREQcioofERERcSgOVfzUqlWLoKAgy61IkSI8++yzdOzYkYsXL6ZoGxs2bKBKlSqEhoZy5MiRh/adMGECLVu2/DeiP5FatWrRokULDMOwWr9161aCgoJISkqyU7JHu/scnT179r62uXPnEhQUxJgxY+yQLPWWL19OUFAQM2bMsFrfunXrNL1PSUlJTJ48mbp16xIcHEzVqlUZOHAgERERQNrPD3d+rzZt2mS1bvv27YSEhBAWFmanVI8WGRnJsGHDqF27NiVLlqR+/fp8/vnnJCYm2jvaI937nhsUFET58uXp168fUVFR9o6WYn//7Lh7a9y4sb2jyRNysXcAW+vTp4/lF9ZsNnPs2DEGDRpE7969+frrrx95/5EjR1K1alU6d+5Mzpw5H9q3bdu2tG7d+l/J/aR27drF/PnzadGihV1zPAlXV1fWrFnDm2++abV+9erVmEwmO6VKveXLl5MvXz4WL15M27Zt7R0nxUaNGsX69esZPHgw+fPn5+LFi4wYMYJ33nmHhQsXMmHCBFxdXe0d87EcOnSIDh060KhRI/r372/vOA908+ZNWrRoQZYsWQgLCyMgIIADBw4QFhbGkSNHGDlypL0jPtLYsWMpW7YsZrOZixcv8tFHH/HZZ5+l6YLz7+797LjLxcXhPkL/Mxxq5AfAx8eHbNmykS1bNnLkyEHlypXp1q0bW7du5fbt24+8/+3btwkNDSUgIOCRv/je3t5kzJjxX0r+ZPz9/Rk9ejTXr1+3a44nUbZsWdasWWO1Lioqip07d1KsWDE7pUqdW7dusXHjRrp06cKRI0c4cOCAvSOl2KJFi+jatSuVK1fG39+fsmXLMnLkSPbv38/u3bvJmDEj3t7e9o6ZYmfPnqVdu3ZUqlSJsLCwNFtQjxw5EldXV2bOnEnFihXJkycP9evXZ9SoUSxbtozdu3fbO+IjZciQwfKeW6pUKd544w1Wrlxp71iP5d7Pjru3TJky2TuWPCGHK34exM3NDQAnJydu375N7969KVOmDJUrV2bgwIGW4dmgoCDOnz/PwIEDUzSic+9hr549e1KvXj0SEhIAWLlyJSEhIZw4ceIp7dUdb731Ft7e3owYMeKB7cePH6ddu3aEhoZSokQJWrZsydGjR59qppSqXbs2O3bssCpK161bR9myZe/7kP3++++pU6cOJUuW5IMPPuD9999nwoQJto78SD/99BNubm40bNiQ/Pnzs2jRIntHeixbtmwhOTnZspwnTx5WrFhBkSJF0sVhr7siIiJ4++23CQoKYuTIkTg7O9s70gMlJCSwYsUKXnvtNdzd3a3ann32WWbNmkVgYGCafh0/SObMma2Wo6Ki6N+/PxUrViQ4OJj69evz008/2Snd4wsPD6d27dqEhobSsmVL9uzZY+9I8ggOX/ycOnWK8ePHU7VqVby9venXrx83btxgzpw5TJs2jZMnT9K3b18ANm7cSM6cOenTp89jf7D27duXyMhIZsyYwc2bNwkLC6Nbt24ULFjwaeyWhaenJ/369WPx4sXs2LHDqs0wDDp16kTu3LlZunQp8+bNw2w2M3z48KeaKaUKFSqEv78/69evt6z75ZdfqFOnjlW/7du3069fP9q2bcuiRYvw9PRMs39VLlu2jGrVquHi4kLt2rVZvnx5upi3AfDGG28wd+5catasyYABA1ixYgWRkZEULlwYDw8Pe8dLsaioKNq1a8fFixcZNWqU5Y+ftOjMmTPExMRQokSJB7ZXqFABDw+PNP06/rvr168ze/ZsXnjhBcu6oUOHcvz4cWbMmMHy5cspV64cAwcOtPyxmJatWbOGcePG0bdvXxYvXky1atV48803uXLlir2jyUM4XPEzZMgQQkNDLX8hvfzyyxQvXpwRI0Zw5swZfv75Z4YPH06RIkUIDg5m2LBhrFq1iosXL5ItWzacnZ3x8fF57MNZmTNnpm/fvkybNo2+ffsSEBBAmzZtns5O/k2dOnWoUaMGH3/8sdUk57i4OJo2bUrv3r3JmzcvxYsX56WXXuLYsWM2yZUStWrVshz6SkxMZOPGjdSqVcuqz9y5c6lfvz6tWrWiUKFCDB48+JHzsezh8uXLbN++3VK81atXjxs3brBu3To7J0uZzp07M2bMGPLmzcuiRYt4//33qVKlCtOnT7d3tMcyZMgQ4M6I7+eff27nNA8XGRkJgK+v7z/2iY2NTfOv4w4dOhAaGkqpUqWoWLEiBw4csBo9L1OmDB9//DFFixYlf/78tG3bllu3bnH58mU7prZ272fH3VtERATTp0/n3XffpU6dOuTPn5+OHTsSHBzMggUL7B1ZHsLhZmt16dKFBg0aEBMTw8SJE7lw4QI9evQgU6ZM7Nq1C8MwqFmz5n33O3XqFLly5UrVz37xxRdZvHgxa9asYfny5TYdah8wYACNGjVi9uzZlvkynp6etGrViqVLl7Jv3z5OnDjBgQMH7D5P6V61a9emU6dOJCUlsWXLFgoXLkzWrFmt+hw+fJimTZtall1cXAgODrZ11EdauXIlzs7OVK9eHYCSJUuSPXt2lixZct9oVlrVsGFDGjZsSGRkJJs2bSI8PJwRI0aQP39+e0dLsQwZMjBjxgxWrlxJWFgY9erVo1SpUvaO9UB355TcunXrH/t4eXml+dfx3cIB7uzLsmXLaNGiBQsWLKBAgQI0adKE1atXs2DBAk6cOMH+/fuBOyelpBV3PzvulTFjRo4fP87o0aMZN26cZX1CQkKa/ANM/uJwxU/mzJnJly8fAGPGjKFp06Z07tyZ+fPnk5ycjJeXF0uWLLnvftmyZUv1z46JieHMmTM4Ozuzbds2nnnmmVRvM6UCAgLo0KEDEyZMYPDgwQDEx8fTtGlTMmTIQJ06dWjcuDEnTpxIU38Nly5dGmdnZ3bs2MEvv/xC3bp17+vj7Ox83+n8f19OC+4e4ipfvrxlndlsZu3atWl+QvqhQ4f47rvvGDBgAAB+fn40aNCA+vXr07Rp0/tOH0/LevfuTaZMmWjVqhXLly+nT58+LF269L45NWlB3rx5yZgxI3v37iUkJOS+9u7du1OzZk2mTp2apl/H2bNnt7zvAoSEhLB+/Xrmz59P79696dWrF3/88QcvvvgiLVu2JFu2bGnuDNV7PzvulZycTO/evalSpYrVei8vL1tFkyfgcIe97uXm5kZYWBiHDh1i5syZFChQgJiYGJKTk8mXL5/lF33o0KH/yjUpxo0bh6enJ0OGDGHUqFFcunQp1dt8HG+//TbZs2dn7NixAGzbto1Lly4xe/Zsy1kvFy5cSFOFg5OTEzVq1GDNmjX8+uuvDxwhKVy4MPv27bMsJycnc/DgQVvGfKRTp06xb98++vbty5IlSyy3adOmkZiYyIoVK+wd8aGSk5OZPXs2u3btslpvMpnw9fW9bwJrWnb3LE2TyURYWBjnz5+3vCbSGmdnZxo1asQ333xz3/yXLVu28MMPP3D27Nk0/zr+J8nJyURFRbF8+XJGjRrFe++9R926dS0jXelhHwoUKMClS5csnxn58uVjxowZbNu2zd7R5CEcuviBO3+BNG3alClTpuDj40PVqlXp1asXu3fv5tChQ/Tu3ZuIiAiyZ8+eqp+zb98+vvnmGz766CNeeeUVihQpYhmBsRU3NzcGDRrE+fPngTvzCGJjY/n55585d+4cCxYsYM6cOWlukmHt2rVZsGABGTNmJE+ePPe1v/766/z000/Mnz+fkydPMnToUM6fP5+mTl1evnw5fn5+tGzZksDAQMutevXqhIaGsnjxYntHfKjixYtTs2ZNunTpwuLFizl79ix79+5lzJgxHDx40OqwY3pSqFAh2rdvz1dffcXOnTvtHeeBunTpQnx8PG3atGHLli2cOXOGxYsX0717d15++WUqV66c5l/Ht27d4urVq1y9epWzZ88yduxYTp8+TYMGDXB3d8fT05NVq1Zx7tw5Nm7caJmXlZb24Z+0adOG2bNns3jxYs6cOcPEiRNZuHDhUz+ZRVLH4YsfgB49euDq6sqwYcMYPnw4+fLlo23btrz++utkz56dyZMnp2r7ycnJDBw4kAYNGlC+fHlMJhODBg1iw4YNNj8rqWLFipYLdZUqVYouXbrwySef8MILL7Bw4UIGDRrEzZs3uXDhgk1zPUyVKlUwm83/OC8mNDSUQYMGMXnyZJo0aUJkZCSlS5dOUxfcW7FiBY0bN37goZWWLVuyf//+R14x3N7Gjh1L8+bNmTZtGo0aNaJNmzYcOXKEOXPmpOv5De3bt6dQoUL07duXuLg4e8e5T+bMmZk7dy6FChWid+/eNG7cmC+++IJ3333XMpcmrb+Ou3fvTpUqVahSpQqNGjVi06ZNTJgwwfI6HTFiBKtXr6Zhw4Z8+umndOjQgRw5cqSL62A1bNiQDz74gIkTJ9KoUSN+/vlnJk2aRNGiRe0dTR7CZKSHcUWRh9izZw8+Pj5Wf2k1atSIt99+m5dfftmOyUREJC3SyI+kezt37uTdd9/ljz/+4OzZs0ydOpWLFy9StWpVe0cTEZE0yOHO9vq3/PTTT/Tp0+cf24OCgpg3b54NEzmu1157jXPnztG1a1du375N0aJF+eKLL/6VM/REROS/R4e9nlB0dDTXrl37x3Y3N7dUXxdIRERE/n0qfkRERMShaM6PiIiIOBQVPyIiIuJQVPyIiIiIQ1HxIyIiIg5FxY+IiIg4lP8DDOEGgMU/XV0AAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "from ipywidgets import interact\n", "\n", "# There is a warning about setting a copy of a slice from a dataframe: ignore\n", "warnings.filterwarnings('ignore')\n", "\n", "TYPES = ['Window Float', 'Window Non-Float', 'Container']\n", "\n", "def get_dataset(src, glass_type):\n", " #subset the df into a new df\n", " df = src[src.GlassType == glass_type]\n", " df.drop(['ID'], axis=1,inplace=True)\n", " return df\n", "\n", "def make_heatmap(source, title):\n", " cmap = sns.diverging_palette(230, 20, as_cmap=True)\n", " corr=source.corr()\n", " # Generate a mask for the lower triangle\n", " mask = np.tril(np.ones_like(corr, dtype=bool))\n", " sns.heatmap(corr, mask=mask, cmap=cmap, annot=True, fmt=\".2f\",\n", " vmax=.3, center=0,square=True, linewidths=.5, cbar_kws={\"shrink\": .5})\n", " plt.title(title, fontsize=16)\n", " plt.xticks(fontsize=14,rotation=0)\n", " plt.yticks(fontsize=14,rotation=0)\n", " plt.gcf().set_size_inches(15, 8)\n", " return plt, title\n", "\n", "def update_heatmap(glass_type):\n", " src = get_dataset(dataset, glass_type)\n", " title = make_heatmap(src, glass_type + ' glass correlation')\n", " plt.title.text = title\n", " plt.show()\n", "\n", "for kind_of_glass in TYPES:\n", " update_heatmap(kind_of_glass)\n", "\n", "# Instead of using a for-loop, it is also possible to have an interactive exhibit, using this Python syntax:\n", "#interact(update_heatmap, glass_type=TYPES);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Analysis of Correlation Heat Maps\n", "One might expect barium, an ingredient used to increase brilliance, thus increasing refractive index, would be positively correlated with refractive index in these graphs. Barium, however, is frequently measured at 0.0 weight percent in the 1987 dataset or present in only trace amounts. It may be that late 20th century glassmaking does not incorporate barium to the same degree that it was incorporated into glass production of the early twentieth century. For an example of the importance of barium in the early 20th century, see a 1936 Hazel Atlas glass patent [here](https://tighar.org/Projects/Earhart/Archives/Research/ResearchPapers/freckleintime/Document_05_FrancisFlint1936patent.pdf).\n", "\n", "Calcium, in the form of calcium oxide, is highly correlated in the 1987 database with refractive index. This is due to the fact that calcium oxide increases refractive index. See https://www.academia.edu/12254939/Optical_and_mechanical_properties_of_calcium_phosphate_glasses for a study of this effect. There is good evidence that calcium was the element of choice to increase brilliance in 1980s glass production, rather than barium. The toxic effects of barium, which is a heavy metal, were becoming much better understood by the time the U.K. researchers assembled their database. The World Health Organization published a [memorandum](https://inchem.org/documents/hsg/hsg/hsg046.htm#SectionNumber:4.1) in 1991 in which it specifically warned of the dangers of barium in glass production. One prominent U.S. [patent](https://patentimages.storage.googleapis.com/7e/18/e1/18b38abaca0806/US8877663.pdf) from 2013 specifically mentions CaO (calcium oxide) as an ideal substitute for metals such as barium and lead.\n", "\n", "By contrast, in the early twentieth century barium was a favorite ingredient of glassmakers. As Francis Flint described in his patent (cited above), the use of barium sulfate increased brilliance, but the sulfates needed then to be reduced to prevent small seeds forming in the glass mixture, thus reducing the quality of the glass. Flint recommended zinc, magnesium, aluminum or tin as reducing agents. Sodium and calcium have been recommended in more modern literature of the art.[3] In window non-float glass, aluminum is positively correlated with barium.\n", "\n", "It would seem good practice to analyze the correlations of each of these glass types separately, as we have done, since obviously the desired qualities of the glass will differ depending on the uses to which the glass will be put, and thus recipes will differ accordingly. The desired refractive index and brilliance of vehicle float glass will be far different than that of container or tableware glass, for example.\n", "\n", "The elemental correlations in this 1987 database suggest changing priorities between production techniques of the early twentieth century and production techniques of the 1980s, toward more utilitarian styles and techniques. One does not require statistics to observe that container glass with high refractive index has become less common, if it ever was, giving way to containers in which seeing the contents clearly through the glass is the overriding concern. This change will probably affect what specific types of glass that a machine learning model can correctly predict for much earlier samples that it has not seen.\n", "\n", "Now we turn to our last question." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Using machine learning to train a model on the 1987 database, can that model be used to identify the type (container) of one or both of the jars unseen by the model?" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This should be a straightforward machine learning classification problem. The first step is to gather up relevant modules from various sci-kit learn and imbalanced learning libraries, using multiple import statements." ] }, { "cell_type": "code", "execution_count": 88, "metadata": {}, "outputs": [], "source": [ "from sklearn import metrics\n", "from sklearn.metrics import confusion_matrix, mean_absolute_error, cohen_kappa_score, classification_report\n", "from sklearn.model_selection import train_test_split, KFold, cross_val_score, RepeatedStratifiedKFold\n", "from sklearn.model_selection import StratifiedKFold as SKF\n", "from sklearn.linear_model import LinearRegression, LogisticRegression, Lasso, ElasticNet\n", "from sklearn.tree import DecisionTreeRegressor\n", "from sklearn.neighbors import KNeighborsRegressor, KNeighborsClassifier\n", "from sklearn.svm import SVR\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.preprocessing import StandardScaler\n", "from collections import Counter\n", "from imblearn.over_sampling import SMOTE\n", "from imblearn.pipeline import make_pipeline\n", "from sklearn.model_selection import RepeatedStratifiedKFold\n", "from sklearn.feature_selection import SelectKBest, f_classif\n", "from numpy import set_printoptions, where" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Setting the stage for an attractive output\n", "Next, we will address the fact that the output we want to see needs to be prepared for attractive presentation. After we have generated predictions for the identity of the jars, we will generate, additionally, probability reports that outline the likelihood of the predictions. The standard probability scores supplied by the predict_proba function appear rather clumsily, as:\n", " \n", "[[0.42 0.4 0.07 0.04 0. 0.07]]\n", "\n", "This output is a sequential list of probabilities that corresponds to the sequence of integer types supplied by the original 1987 database. The sequence is implied; that is, the 0.42 above is presumed to be the first type, Window Float, the 0.4 is presumed to be the second, and so on.\n", " \n", "In their stead, it would be nice to have an equivalent and more readable report that formats the 2d array to \n", "appear like this:\n", " \n", " The Nikumaroro jar has a probability of:\n", " A% to be a Window Float\n", " B% to be a Window Non-Float\n", " C% to be a Vehicle Float\n", " D% to be a Headlamp\n", " E% to be a Container\n", " F% to be a Tableware\n", " \n", "To achieve this, we will want to re-sort the 2D list by descending order of probability, but when this is done, the implied sequence will no longer be of any use in identifying the glass types.\n", " \n", "To solve for this, we will make an explicit counter variable. This new explicit counter, once created, will correspond with the original integer type values supplied by the 1987 database and will not be subject to the alteration that the implied index would experience when sorting." ] }, { "cell_type": "code", "execution_count": 89, "metadata": {}, "outputs": [], "source": [ "from colorama import Fore, Back, Style\n", "\n", "def make_nicer_probability_output(array_to_convert,title):\n", " \"\"\"\n", " Author: Joe Cerniglia\n", " Date: March 20, 2022\n", " A function to convert the standard probability scores in Python machine learning libraries\n", " to a report that supplies the categories from a dictionary and sorts the list of probabilities in\n", " descending order.\n", " \n", " Parameters:\n", " array_to_convert: an array to convert to a sorted 2d list \n", " title: some text to be placed in the title or headings of the report\n", " Preconditions:\n", " The array must have the exact number of elements and format needed for the dictionary\n", " and must be the output of a call to model.predict_proba\n", " \n", " returns None. The function itself prints the report.\n", " \"\"\"\n", " prob_list = array_to_convert.tolist()\n", " #print(prob_list)\n", " \n", " # The counter can be used as a dictionary key for a \n", " # dictionary we will create in the next step\n", " counter=0\n", " for probability in prob_list[0]:\n", " prob_list[0][counter]=[counter,probability]\n", " counter+=1\n", " #print(prob_list)\n", " # We have a 3d list; get back to the 2d list\n", " prob_list=prob_list[0]\n", " # Sort in descending order the second column of each row in the 2d list\n", " # This allows for a descending order of probability and for the\n", " # predicted type (highest probability) to rise to the top of the list\n", " \n", " # The lambda expression below takes each line (l) of the list as its \n", " # argument. The expression after the colon, l[1], is the function, \n", " # which takes the second variable in the line and sorts the line in \n", " # descending order by that variable\n", " prob_list=sorted(prob_list,key=lambda l:l[1], reverse=True)\n", "\n", " counter=0\n", " for prediction in prob_list:\n", " if counter==0:\n", " print(Fore.BLACK + 'The ' + title + ' has a probability of:')\n", " if prediction[0]==3:\n", " pred=Fore.RED + probdict[prediction[0]]\n", " else:\n", " pred=Fore.BLACK + probdict[prediction[0]]\n", " print(Fore.BLACK + \"{:.0%}\".format(prediction[1]),'to be a', pred)\n", " counter+=1\n", " return None" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "There is one last piece of code that is needed to have this function work properly. We need to define a dictionary that will translate the counter variable above into English. This dictionary is called in the last for-loop of the function above. Notice that this dictionary is different than the one we created at the beginning, since it is now indexed sequentially starting at 0 and explicitly omits Vehicle Non-Float, a category for which there exist no examples in the database. Note also that we will only use this dictionary inside this function. When we need to translate the numerical glass types found in the 1987 database elsewhere in this code, we will still use the dictionary (dict) that we defined earlier." ] }, { "cell_type": "code", "execution_count": 90, "metadata": {}, "outputs": [], "source": [ "# This dictionary orders the possibilities of the 2d\n", "probdict = {0: 'Window Float',\n", " 1: 'Window Non-Float',\n", " 2: 'Vehicle Float',\n", " 3: 'Container',\n", " 4: 'Tableware',\n", " 5: 'Headlamp'}" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Split the data into feature set X and target set Y.\n", "Now with the correct modules imported, and our utility function defined, we can split the database into two arrays. First, we can use the pandas dataframe values method to convert the entire dataframe we set aside earlier to an array. Next, we can slice the array column-wise into X and Y arrays, with X as our features array and Y as our target array. Note that our columns are sliced from the number 1, which is actually the second variable in the array, not from the number 0. The reason is that we need to drop the ID variable. The ID variable is an index. It would improve the accuracy of our machine learning model (to 100% in fact!), but this accuracy would be a mirage. It would not accomplish our goal of training the model in how to classify unseen examples." ] }, { "cell_type": "code", "execution_count": 91, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "array = dataset_ml.values\n", "print(type(array))\n", "X = array[:,1:10]\n", "Y = array[:,10]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Show the list of data features ranked by their power to influence prediction of the target variable.\n", "Earlier, we stated that we would make some educated guesses as to the values of the jars' features that were either unavailable (refractive index) or not stated in the required units of weight percent (Fe, K, Si). We did not do this without some trepidation, since tampering with these features' values would appear to reduce the rigor of our analysis. However, the benefits of having a complete training dataset appeared to outweigh these drawbacks. \n", "\n", "The impact of this decision was unknown, but it was not unknowable. There is a method to assess the power of a given feature to influence the classification of a given sample of glass. This method is known as feature selection. Using feature selection, we may see a list of all the features in the dataset ranked in order of strongest to least strongest influence on the prediction of the class. If the features we modified were ranked highly in this list, we should be concerned about the integrity of the analysis. If the features we modified were not ranked highly in this list, we can proceed with our analysis with the confidence that the algorithm will be untroubled by our expedient modifications to the original data.*\n", "\n", "The results were as favorable as one might have hoped. Using the SelectKBest class from the scikit-learn library, we can see that the variables for which we needed to supply values (highlighted in gold) were ranked toward the bottom of the ranked list. \n", "\n", "The code that appears below takes the Numpy array created by SelectKBest and transforms it into a concise report that lists the relative importance, ranked descending, for all of the features in the glass database.\n", "\n", "#### (*Note that we also evaluated the effect of the features empirically. An additional experiment to run this program with a range of values for K, Si, Fe, and Ref_ix did not materially alter the predictions of the machine learning algorithm.)\n" ] }, { "cell_type": "code", "execution_count": 92, "metadata": { "scrolled": true }, "outputs": [], "source": [ "\n", "names=['Ref_ix','Na','Mg','Al','Si','K','Ca','Ba','Fe']\n", "\n", "# feature extraction using univariate selection\n", "test = SelectKBest(score_func=f_classif, k=4)\n", "fit = test.fit(X, Y)\n", "# summarize scores\n", "\n", "# Convert scores to dataframe\n", "df_scores = pd.DataFrame(fit.scores_)\n", "# Convert names list to dataframe\n", "df_names = pd.DataFrame(names) \n", "# Join two dataframes by indexes\n", "best_features=pd.concat([df_scores, df_names], axis=1)\n", "# change the type of the columns from int to str\n", "best_features.columns = best_features.columns.astype(str)\n", "# rename the columns to have more sensible names\n", "best_features.columns.values[0] = \"Score\"\n", "best_features.columns.values[1] = \"Feature_Name\"\n", "\n", "# sort the rows (features) by rank in descending order\n", "best_features.sort_values(by='Score', ascending=False, inplace=True)\n", "# Add a column for rank to the dataframe\n", "best_features['Rank'] = np.arange(1,len(best_features)+1)\n", "# Re-order the columns\n", "best_features = best_features[['Feature_Name', 'Score','Rank']]\n", "# Format the dataframe to give the score two decimal places\n", "best_features = best_features.style.format({'Score': \"{:.2f}\"})\n", "#print(type(best_features))\n", "best_features=best_features.data\n", "\n", "def color_relevant_gold(row):\n", " \"\"\"\n", " Takes a dataframe row and returns a string with\n", " the background-color property `'background-color: gold'` for relevant\n", " rows, black otherwise.\n", " \"\"\"\n", " if (row.values[0] == 'K' or row.values[0] == 'Ref_ix' or row.values[0] == 'Si'\n", " or row.values[0] == 'Fe'):\n", " color = 'gold'\n", " else:\n", " color = ''\n", " return ['background-color: %s' % color]*len(row.values)\n", "\n", "# There is a problem in rendering the Pandas properly, so I am embedding this code within a function \n", "# and pasting in the image that Jupyter Notebook renders.\n", "def awaiting_resolution():\n", " print(' ----Best features----')\n", " from IPython.display import display, HTML\n", " # There is a deprecation warning here that is unclear how to resolve\n", " warnings.filterwarnings('ignore')\n", " display(HTML(best_features.style.apply(color_relevant_gold, axis=1).render()))\n", " # Turn warnings back on again\n", " warnings.filterwarnings('always')" ] }, { "attachments": { "image.png": { "image/png": "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" } }, "cell_type": "markdown", "metadata": {}, "source": [ "![image.png](attachment:image.png)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Split the data into training and validation sets.\n", "To create the conditions under which we will have the ability to test our future model's effectiveness at classifying unseen data, we can now split both X and Y into validation and training arrays. The \"train\" pair of arrays can then be used for training the model, and the \"validation\" pair of arrays can be used to demonstrate how effective the model is after we have trained it. To do this, we will use the train_test_split function from sklearn's model_selection library. We will create an 80-20 split of the data by entering a test_size parameter of 0.20. The training set will be 80% of the data, and the validation set will be 20% of the data. We will stratify the data so that the relative proportion of glass types in both pairs of training and validation sets is equivalent, despite the fact that the arrays are of different overall sizes." ] }, { "cell_type": "code", "execution_count": 93, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "validation_size = 0.20\n", "seed = 7\n", "X_train, X_validation, Y_train, Y_validation = train_test_split(X, Y,\n", " test_size=validation_size, random_state=seed, stratify=Y)\n", "print(type(X_train))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Add the Nikumaroro jar and clear facsimile into the X and Y validation datasets.\n", "Next, we need to add our two jars (clear facsimile and Nikumaroro jar) to the validation arrays. The features of these two jars will be added to X_validation, and the target (glass type) of these two jars will be added to Y_validation. Having accomplished this, we will then have added two glass samples that the study authors could not possibly have anticipated when they built their database in 1987. This will be a great test of the skill of machine learning algorithms in general and of the completeness of their original dataset in particular.\n", "\n", "To demonstrate that our jars have been added successfully, we will print out the shape of the validation arrays both before and after performing the append operations. Note that the shape returns a tuple, providing the number of rows, followed by the number of columns, separated by a comma.\n", "\n", "The append operations for arrays are a little tricky. To do this, in addition to using numpy's append method, we need to chain to that the reshape method to size the array appropriately prior to appending to it. The chained command executes from right to left, first reshaping the array to which we are appending, and then performing the append operation itself." ] }, { "cell_type": "code", "execution_count": 94, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "shape of X_validation before adding jars: (43, 9)\n", "shape of X_validation after adding jars: (45, 9)\n", "shape of Y_validation before adding jars: (43,)\n", "shape of Y_validation after adding jars: (45,)\n" ] } ], "source": [ "#names=\n", "# ['Ref_ix','Na', 'Mg', 'Al', 'Si', 'K', 'Ca', 'Ba', 'Fe']\n", "\n", "artifact_features=[1.52369, 13.1, 4.3, .74, 72.37, .24, 8.5, .74, .02]\n", "facsimile_features=[1.51316, 11.7, 2.4, .85, 72.37, .12, 3.6, .37, .01]\n", "artifact_features_array=np.array(artifact_features)\n", "facsimile_features_array=np.array(facsimile_features)\n", "print('shape of X_validation before adding jars:',X_validation.shape)\n", "\n", "X_validation=np.append(\n", "X_validation,artifact_features_array).reshape(X_validation.shape[0]+1,9)\n", "X_validation=np.append(\n", "X_validation,facsimile_features_array).reshape(X_validation.shape[0]+1,9)\n", "print('shape of X_validation after adding jars:',X_validation.shape)\n", "\n", "# 5 is equal to a container, the actual identity of the Nikumaroro jar and of the facsimile\n", "artifact_identity=[5.0]\n", "facsimile_identity=[5.0]\n", "artifact_array=np.array(artifact_identity)\n", "facsimile_array=np.array(facsimile_identity)\n", "print('shape of Y_validation before adding jars:',Y_validation.shape)\n", "\n", "Y_validation=np.append(\n", "Y_validation,artifact_array).reshape(Y_validation.shape[0]+1)\n", "Y_validation=np.append(\n", "Y_validation,facsimile_array).reshape(Y_validation.shape[0]+1)\n", "print('shape of Y_validation after adding jars:',Y_validation.shape)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Create a list of models to begin the testing process of choosing the best one.\n", "Now comes the heart of the machine learning program, the creation of models, also known as machine learning algorithms. There are many different kinds of machine learning algorithms. Each has its own strengths and weaknesses. The skill of machine learning algorithms on particular datasets will vary. Some will show very little skill in exposing the structure of the data, resulting in a model that cannot classify unseen glass examples very accurately. Others will show much greater skill. We want to test a variety of machine learning algorithms on the data so that we can select the model that is most likely to succeed in classifying the type of glass for this particular set of data.\n", "\n", "Machine learning algorithms come in two basic varieties: regression and classification. Regression algorithms treat our target variable (glass type) as a continuous variable. This means that they consider the targets as floating point numbers rather than as discrete integers. Classification algorithms treat the target variable as discrete categories. There is every advantage in testing both varieties, rather than trying to anticipate in advance which type of model might perform best." ] }, { "cell_type": "code", "execution_count": 95, "metadata": {}, "outputs": [], "source": [ "models = []\n", "models.append(('LR', LinearRegression()))\n", "models.append(('LASSO', Lasso()))\n", "models.append(('EN', ElasticNet()))\n", "models.append(('KNN', KNeighborsRegressor(n_neighbors=5)))\n", "models.append(('CART Regressor', DecisionTreeRegressor()))\n", "models.append(('SVR', SVR()))\n", "models.append(('KNClass',KNeighborsClassifier(n_neighbors=5)))\n", "models.append(('RandomForest',RandomForestClassifier(\n", "n_estimators=100, max_features=9,class_weight='balanced')))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Score the models.\n", "There are many methods for scoring the effectiveness of different models so that they may be compared side by side. Some of these methods work optimally for regression models; others work optimally for classification models. The negative mean squared error, while not optimal for classification models, is acceptable for use in multi-class datasets such as this one, and it works very well for regression models. Thus, it is a good 'all-purpose' scoring method for our needs.\n", "\n", "What this code snippet does is to evaluate the success of each of the models defined above by repeatedly testing them on different randomized 'folds' of the data. The results of each test will not be the same. Each will vary to a lesser or greater extent, depending on which fold is selected. The score is actually a composite result of repeated tests on many folds. The negative mean square error score is expressed as a negative number, such that the highest negative number, that which is closest to zero, will be considered to have the best score. The standard deviation, given in parentheses, provides a sense of how much variation to expect within a given score." ] }, { "cell_type": "code", "execution_count": 96, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Evaluation of Mean Square Error results of different models\n", "LR: -1.206907 (0.388832)\n", "LASSO: -2.496606 (0.292277)\n", "EN: -2.112353 (0.293161)\n", "KNN: -0.980076 (0.386747)\n", "CART Regressor: -1.397302 (0.709613)\n", "SVR: -4.767184 (0.520934)\n", "KNClass: -1.683810 (0.841334)\n", "RandomForest: -1.193651 (0.683872)\n" ] } ], "source": [ "# Test options and evaluation metric\n", "scoring = 'neg_mean_squared_error'\n", "# evaluate each model in turn\n", "results = []\n", "names = []\n", "print('Evaluation of Mean Square Error results of different models')\n", "for name, model in models:\n", " kfold = RepeatedStratifiedKFold(n_splits=7, n_repeats=3, random_state=7)\n", " cv_results = cross_val_score(\n", " model, X_train, Y_train, cv=kfold, scoring=scoring)\n", " results.append(cv_results)\n", " names.append(name)\n", " msg = \"%s: %f (%f)\" % (name, cv_results.mean(), cv_results.std())\n", " print(msg)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Graph the score results.\n", "The test results above are complete, but they are a little hard to read. It would be far easier to evaluate a graphical representation of the data using a box plot. The box plot below takes each machine learning algorithm and places it alongside its neighbors for easy comparison. Note that there are tuning and scaling operations that might have incrementally increased the accuracy of our models, which have not been employed here.\n", "\n", "Based on the graph, it would appear that the Random Forest Classifier, a model with an excellent reputation among classification models, is an effective model, but KNN and Cart Regressor also seem competitive. In actual practice, however, KNN and CART produced a far less accurate result. When testing these models, the resulting confusion matrix (see below for an explanation of the confusion matrix) showed that these models are actually fairly weak.\n", "\n", "This illustrates an important point: Multiple scoring methods are often required to see which model offers the best results. Still, a box graph such as the one shown below can narrow the number of judgments that must be made." ] }, { "cell_type": "code", "execution_count": 97, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ " # Compare Algorithms\n", "fig = plt.figure()\n", "fig.suptitle('UNTUNED Unscaled Algorithm Comparison',fontsize=20)\n", "ax = fig.add_subplot(111)\n", "plt.boxplot(results)\n", "ax.set_xticklabels(names)\n", "plt.gcf().set_size_inches(15, 7)\n", "plt.xticks(fontsize=16)\n", "plt.yticks(fontsize=16)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Oversampling, explained\n", "Random Forest often pairs well with oversampling techniques, such as SMOTE (Synthetic Minority Over-sampling Technique). What SMOTE will do is to magnify certain classes that are sparsely represented in the training dataset. This magnification will allow the training process to resolve these under-represented classes more effectively, because it will have seen more examples than were initially presented. For example, tableware has only seven examples in the training dataset. By applying SMOTE to the training dataset, this number is magnified to 61. In fact, all of the glass types are 'leveled' to 61 examples after SMOTE is applied. This leveling will make it much easier for the Random Forest Classifier to classify unseen examples to a greater level of accuracy.\n", "\n", "Now we can observe a comparison of counts for each of the glass types before and after the application of SMOTE. This is only a test to see what the counts would be. Later on, we will apply SMOTE to the definition of our model. Our dictionary will come in handy here to translate the numerical glass types found in the 1987 database to their English equivalents." ] }, { "cell_type": "code", "execution_count": 98, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Counts Before SMOTE:\n", "Window Float : 56\n", "Headlamp : 23\n", "Vehicle Float : 14\n", "Window Non-Float : 61\n", "Tableware : 7\n", "Container : 10\n", "\n", "Counts After SMOTE:\n", "Window Float : 61\n", "Headlamp : 61\n", "Vehicle Float : 61\n", "Window Non-Float : 61\n", "Tableware : 61\n", "Container : 61\n" ] } ], "source": [ "counter=Counter(Y_train)\n", "print('Counts Before SMOTE:')\n", "for ele in counter:\n", " print(dict[int(ele)],':',counter[ele])\n", "oversample = SMOTE(random_state=42,k_neighbors=5)\n", "X_trainsm, Y_trainsm = oversample.fit_resample(X_train, Y_train)\n", "countersm=Counter(Y_trainsm)\n", "print()\n", "print('Counts After SMOTE:')\n", "for ele in countersm:\n", " print(dict[int(ele)],':',countersm[ele])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Visualizing SMOTE\n", "Seeing the change in counts before and after the application of SMOTE oversampling is an excellent way of visualizing what SMOTE does. Another way to do this is to create a scatter plot on two elements of the periodic table and compare where the points are plotted before and after the application of SMOTE. The graphs following this function encode each dot with a color. The legend in the upper right corner of the graph shows which glass type each color represents." ] }, { "cell_type": "code", "execution_count": 128, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "names=['Ref_ix','Na','Mg','Al','Si','K','Ca','Ba','Fe']\n", "def smotegraph(title_suffix, var_suffix=None):\n", " \"\"\"\n", " Author: Joe Cerniglia\n", " Date: September 17, 2022\n", " A function to plot two scatter graphs to show the distribution of glass types over two \n", " elements of the periodic table.\n", " \n", " Parameters:\n", " title_suffix: string that provides title text for which training dataset has been \n", " graphed, original or SMOTE-oversampled.\n", " var_suffix: string to indicate which training dataset to use in the plot\n", " \n", " returns None. The function produces graphs but does not return anything outside the function itself.\n", " \"\"\"\n", " # scatter plots of examples by class label: Aluminum (3) x Iron (8)\n", " plt.figure(figsize=(7,5))\n", " for label, _ in counter.items():\n", " # row_ix returns an array containing all the row numbers for a given label (color)\n", " if var_suffix==None:\n", " row_ix = where(Y_train == label)[0] #slicing with zero allows access \n", " #to the array embedded inside the tuple returned by the where\n", " plt.scatter(\n", " X_train[row_ix, 3], X_train[row_ix, 8], label=dict[label], alpha=1, s=30)\n", " else:\n", " row_ix = where(Y_trainsm == label)[0]\n", " plt.scatter(\n", " X_trainsm[row_ix, 3], X_trainsm[row_ix, 8], label=dict[label], alpha=1, s=30)\n", " plt.legend(markerscale=1.5, fontsize=10)\n", " plt.title(\"Glass of the UK database by Type\" + title_suffix, fontsize=20)\n", " plt.xlabel(\"Al weight %\",fontsize=16)\n", " plt.ylabel(\"Fe \\n weight %\",fontsize=16, rotation=0, labelpad=50)\n", " plt.xticks(fontsize=14)\n", " plt.yticks(fontsize=14)\n", " plt.show()\n", " return None\n", "smotegraph(' - before SMOTE oversampling')\n", "smotegraph(' - after SMOTE oversampling','sm')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The second graph clearly shows the number of dots on the graph increased. SMOTE creates observations that did not exist in the original U.K. database but which are nearest neighbors. Note especially how the upper right quadrant now has a number of containers (brown dots) where only one existed before! These observations have moderate amounts of iron in them. Iron increases refractive index, which generally means that glass with high iron content is less transparent. Since containers, especially for products, once equated translucency with luxury, these synthetic observations created by SMOTE seem plausible. Note also the row of purple dots at the bottom of the graph. These are tableware samples created by SMOTE that have low iron content, and thus are highly transparent, which seems quite correct for tableware." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Build data pipelines that can be used to input the training data.\n", "Now that we have seen what the counts would be for an oversampled training dataset, we can create a data pipeline that combines SMOTE with our previously selected model, Random Forest Classifier. To test the efficacy of SMOTE, we should create two models. The first model we create will have the ability to apply a Random Forest Classifier to a non-oversampled dataset. The second will have the ability to apply Random Forest Classifier to an oversampled dataset. We will also want to add in a standard scaler to both model pipelines. The standard scaler takes features of different scale and makes them more similar in scale. This technique is very useful in further enhancing the model's ability to classify. Note that even after we have defined our model pipelines, we have not yet supplied any data to them." ] }, { "cell_type": "code", "execution_count": 100, "metadata": {}, "outputs": [], "source": [ "modelnoSMOTE=make_pipeline(\n", "StandardScaler(),RandomForestClassifier(\n", "n_estimators=100,max_features=9,class_weight='balanced'))\n", "\n", "modelwithSMOTE=make_pipeline(\n", "StandardScaler(),SMOTE(\n", "random_state=42, k_neighbors=5),RandomForestClassifier(\n", "n_estimators=100,max_features=9,class_weight='balanced'))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Fit the non-SMOTE model to the training data to train the model.\n", "First, we will use and evaluate the non-SMOTE model. We will apply the non-SMOTE model to actual data by supplying it with our training data. The fit method applies the model by using the training features and targets to learn how to generalize to unseen validation data. " ] }, { "cell_type": "code", "execution_count": 101, "metadata": {}, "outputs": [], "source": [ "modelnoSMOTE.fit(X_train, Y_train);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Apply the trained non-SMOTE model to unseen data.\n", "Now we can test the model's performance by applying it to unseen data (X_validation) and generating a report on the accuracy of its performance.\n", "\n", "The variable called predictions takes our validation features and attempts to classify what type of glass the features for each example represents. Recall that we added in the two jars to our validation dataset, and so these predictions will include them." ] }, { "cell_type": "code", "execution_count": 102, "metadata": {}, "outputs": [], "source": [ "predictions = modelnoSMOTE.predict(X_validation)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### How to Read a Confusion Matrix\n", "The confusion matrix need not be confusing. It provides all the predictions in a matrix. Think of the columns and rows of the matrix as having been labeled with the names of the integer glass types, ordered left to right and top to bottom, in sequence. Row labels represent the true values, and column labels represent the predicted values. The intersection of true and predicted represents the status of any given cell in the matrix. Using this matrix, we can tell how many predictions were correct, how many were over-predicted, and how many were under-predicted. The diagonal line that can be drawn from the top left of the matrix to the bottom right, slicing the matrix into two equal diagonal halves, represents correct predictions (true=predicted). Each number in the same row as a correct prediction represents an under-prediction. In other words, the model under-predicts when the number of correct predictions for a category is less than the actual number in that category. By summing all of the numerals in any given row, we obtain the true number of glass types for each type. Each number in the same column as a correct prediction represents an over-prediction. In other words, the model over-predicts when the number of total predictions for a category in a given column is more than the correct predictions in that category. " ] }, { "cell_type": "code", "execution_count": 103, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Random Forest without SMOTE oversampling\n", "[[13 1 0 0 0 0]\n", " [ 1 13 0 1 0 0]\n", " [ 2 0 1 0 0 0]\n", " [ 0 1 0 3 0 1]\n", " [ 0 1 0 0 1 0]\n", " [ 2 0 0 1 0 3]]\n" ] } ], "source": [ "print('Random Forest without SMOTE oversampling')\n", "print(confusion_matrix(Y_validation, predictions))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Cohen's Kappa\n", "The Cohen's Kappa statistic is a measure of inter-rater reliability. The \"raters,\" in this machine learning example, are the actual glass types and the predictions of those types. In addition to considering the extent of agreement between these two, it also includes statistical consideration of the probability that this agreement could have occurred by chance.\n", "\n", "The kappa score for the non-SMOTE Random Forest model is considered reasonably good." ] }, { "cell_type": "code", "execution_count": 104, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Kappa (1=perfect;0=chance;<0=worse than chance): 0.665314\n" ] } ], "source": [ "kappa=cohen_kappa_score(Y_validation, np.round(predictions,0))\n", "print('Kappa (1=perfect;0=chance;<0=worse than chance): %f' % kappa)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Mean absolute deviation\n", "The mean absolute deviation, or MAD, is a measure of the average distance between each data value and the average value of the dataset. In terms of the confusion matrix above, it provides a sense of how far the values are distributed away from the central diagonal line." ] }, { "cell_type": "code", "execution_count": 105, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "MAD (Mean Absolute Dev): 0.711111\n" ] } ], "source": [ "MAD = mean_absolute_error(Y_validation, predictions)\n", "print('MAD (Mean Absolute Dev): %f' % MAD)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### The classification report\n", "The classification report calculates precision and recall. Precision is the number of correct guesses in a class divided by the total guesses in a class. It can be obtained by observing the columns in the confusion matrix. Recall is the number of correct guesses in a class divided by the total number of actual members of the class. It can be obtained by observing the rows in the confusion matrix. The f1-score is the harmonic mean of precision and recall and is a good overall indicator of model effectiveness. Support is the true count for each type." ] }, { "cell_type": "code", "execution_count": 106, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", " 1.0 0.72 0.93 0.81 14\n", " 2.0 0.81 0.87 0.84 15\n", " 3.0 1.00 0.33 0.50 3\n", " 5.0 0.60 0.60 0.60 5\n", " 6.0 1.00 0.50 0.67 2\n", " 7.0 0.75 0.50 0.60 6\n", "\n", " accuracy 0.76 45\n", " macro avg 0.81 0.62 0.67 45\n", "weighted avg 0.77 0.76 0.74 45\n", "\n" ] } ], "source": [ "print(classification_report(Y_validation, predictions, zero_division=1))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Observe the predictions for the Nikumaroro jar and the clear facsimile.\n", "Since we already added the clear facsimile and Nikumaroro jar to the validation dataset, the model has already made its predictions for these two samples. However, the evaluation thus far has not directly revealed what those predictions actually were. Now we will want to observe for ourselves how the model behaves toward these particular instances of the data. To do this, we can simply enter the array for each sample that we created above into the predict method of the noSMOTE model. We may obtain the underlying probabilities for these predictions by using the predict_proba method of the noSMOTE model." ] }, { "cell_type": "code", "execution_count": 107, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Nikumaroro jar prediction: Window Non-Float\n", "clear facsimile prediction: Headlamp\n" ] } ], "source": [ "yhat = modelnoSMOTE.predict([artifact_features])\n", "print('Nikumaroro jar prediction:',dict[int(yhat)])\n", "yhat2 = modelnoSMOTE.predict([facsimile_features])\n", "print('clear facsimile prediction:',dict[int(yhat2)])\n", "yhat_probability = modelnoSMOTE.predict_proba([artifact_features])\n", "yhat2_probability = modelnoSMOTE.predict_proba([facsimile_features])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Display the details of the jar prediction probabilities for the non-SMOTE model.\n", "Now we have the opportunity to use the utility display function we created above. Nesting calls to this function inside another function will allow the report to be generated with a single call when we next run a report for the SMOTE version of our model. We can also here include extra niceties such as colored text to highlight probabilities for the Container class and check marks to indicate where the prediction was correct." ] }, { "cell_type": "code", "execution_count": 108, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[30mNikumaroro jar prediction: \u001b[30mWindow Non-Float\n", "\u001b[30mThe artifact has a probability of:\n", "\u001b[30m41% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m38% to be a \u001b[30mWindow Float\n", "\u001b[30m17% to be a \u001b[30mHeadlamp\n", "\u001b[30m4% to be a \u001b[30mVehicle Float\n", "\u001b[30m0% to be a \u001b[31mContainer\n", "\u001b[30m0% to be a \u001b[30mTableware\n", "\n", "\u001b[30mclear facsimile prediction: \u001b[30mHeadlamp\n", "\u001b[30mThe clear facsimile has a probability of:\n", "\u001b[30m42% to be a \u001b[30mHeadlamp\n", "\u001b[30m22% to be a \u001b[31mContainer\n", "\u001b[30m17% to be a \u001b[30mWindow Float\n", "\u001b[30m13% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m4% to be a \u001b[30mTableware\n", "\u001b[30m2% to be a \u001b[30mVehicle Float\n" ] } ], "source": [ "def the_report():\n", " if dict[yhat[0]]=='Container':\n", " print(Fore.BLACK + 'Nikumaroro jar prediction:',Fore.RED + dict[yhat[0]]+' ' + u'\\u2713')\n", " else:\n", " print(Fore.BLACK + 'Nikumaroro jar prediction:',Fore.BLACK + dict[yhat[0]])\n", " make_nicer_probability_output(yhat_probability,'artifact')\n", " print()\n", " if dict[yhat2[0]]=='Container':\n", " print(Fore.BLACK + 'clear facsimile prediction:',Fore.RED + dict[yhat2[0]]+' ' + u'\\u2713')\n", " else:\n", " print(Fore.BLACK + 'clear facsimile prediction:',Fore.BLACK + dict[yhat2[0]])\n", " make_nicer_probability_output(yhat2_probability,'clear facsimile')\n", " return\n", "\n", "the_report()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Run the report again for the SMOTE model.\n", "Having proceeded through a step-by-step analysis of the machine learning classification report we have built, we can now run the same report again, this time using the second SMOTE model we created. This model will, again: \n", "1. oversample the data with SMOTE\n", "2. scale the data with StandardScaler \n", "3. classify each glass example with RandomForestClassifier and\n", "4. output the numerical prediction and probabilities of glass type for the Nikumaroro jar and for the clear facsimile" ] }, { "cell_type": "code", "execution_count": 109, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "Random Forest with SMOTE oversampling\n", "[[10 4 0 0 0 0]\n", " [ 1 13 0 1 0 0]\n", " [ 1 0 2 0 0 0]\n", " [ 0 1 0 4 0 0]\n", " [ 0 0 0 0 2 0]\n", " [ 1 0 0 1 0 4]]\n", "Kappa (1=perfect;0=chance;<0=worse than chance): 0.701789\n", "MAD (Mean Absolute Dev): 0.466667\n", " precision recall f1-score support\n", "\n", " 1.0 0.77 0.71 0.74 14\n", " 2.0 0.72 0.87 0.79 15\n", " 3.0 1.00 0.67 0.80 3\n", " 5.0 0.67 0.80 0.73 5\n", " 6.0 1.00 1.00 1.00 2\n", " 7.0 1.00 0.67 0.80 6\n", "\n", " accuracy 0.78 45\n", " macro avg 0.86 0.79 0.81 45\n", "weighted avg 0.80 0.78 0.78 45\n", "\n", "Nikumaroro jar prediction: Window Non-Float\n", "clear facsimile prediction: Container\n" ] } ], "source": [ "modelwithSMOTE.fit(X_train, Y_train)\n", "\n", "print()\n", "print('Random Forest with SMOTE oversampling')\n", "predictions = modelwithSMOTE.predict(X_validation)\n", "ytrue=5\n", "predictions=np.round(predictions, 0)\n", "\n", "print(confusion_matrix(Y_validation, predictions))\n", "kappa=cohen_kappa_score(Y_validation, np.round(predictions,0))\n", "MAD = mean_absolute_error(Y_validation, predictions)\n", "print('Kappa (1=perfect;0=chance;<0=worse than chance): %f' % kappa)\n", "print('MAD (Mean Absolute Dev): %f' % MAD)\n", "print(classification_report(Y_validation, predictions, zero_division=1))\n", "yhat = modelwithSMOTE.predict([artifact_features])\n", "print('Nikumaroro jar prediction:',dict[int(yhat)])\n", "yhat2 = modelwithSMOTE.predict([facsimile_features])\n", "print('clear facsimile prediction:',dict[int(yhat2)])\n", "yhat_probability = modelwithSMOTE.predict_proba([artifact_features])\n", "yhat2_probability = modelwithSMOTE.predict_proba([facsimile_features]) " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Display the detail of jar prediction probabilities for the SMOTE model.\n", "Our final coding step will be to re-run the report of probabilities using our SMOTE-enhanced Random Forest Classifier model." ] }, { "cell_type": "code", "execution_count": 110, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[30mNikumaroro jar prediction: \u001b[30mWindow Non-Float\n", "\u001b[30mThe artifact has a probability of:\n", "\u001b[30m44% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m33% to be a \u001b[30mWindow Float\n", "\u001b[30m9% to be a \u001b[31mContainer\n", "\u001b[30m8% to be a \u001b[30mHeadlamp\n", "\u001b[30m6% to be a \u001b[30mVehicle Float\n", "\u001b[30m0% to be a \u001b[30mTableware\n", "\n", "\u001b[30mclear facsimile prediction: \u001b[31mContainer ✓\n", "\u001b[30mThe clear facsimile has a probability of:\n", "\u001b[30m44% to be a \u001b[31mContainer\n", "\u001b[30m28% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m16% to be a \u001b[30mHeadlamp\n", "\u001b[30m9% to be a \u001b[30mWindow Float\n", "\u001b[30m3% to be a \u001b[30mTableware\n", "\u001b[30m0% to be a \u001b[30mVehicle Float\n" ] } ], "source": [ "the_report()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Analysis Procedure\n", "Let us review the machine learning process that we have just created and executed. The following sequence of operations was performed:
\n", "1) Create two arrays. The first contains the database features, all of the predictive columns of the original dataset. The second contains the target column, GlassType.
\n", "2) Create a train-test, stratified split so that the model can be trained on 80% of the data and validated on a subsample of 20%.
\n", "3) Add the two external glass sample jars to the validation (unseen) data, both for features (X) and for target (Y).
\n", "4) Test a set of regression and classification models on the training dataset and report the MSE (mean squared error) results for each one, as well as the standard deviation of the MSE.
\n", "5) Plot the preliminary effectiveness of each model on a box plot.
\n", "6) Select the model most likely to succeed. In this case, Random Forest Classifier was selected.
\n", "7) Build two models of Random Forest Classifier, one without oversampling and the other with oversampling. Use the standard scaler in both models to make the scale of the features more uniform with one another.
\n", "8) Fit the models to the training dataset.
\n", "9) Print a report on the performance of the models.
\n", "10) For each model, print a report on the predicted result for both of the two unseen jars and the probability of the prediction for each glass type.
\n", "\n", "### Findings\n", "One of the most interesting findings was that when the SMOTE oversampling algorithm was used with the Random Forest Classifier, the resolving power of the model increased to the point that it was consistently able to identify the clear facsimile jar correctly as a container. Without the SMOTE oversampling algorithm applied, the model predicted the clear facsimile jar to be a headlamp. The prediction of container scored second place, or lower.\n", "\n", "The non-SMOTE model predicted the Nikumaroro jar to be a window, usually of the float variety*. This is very close to the prediction made by the model when SMOTE was used. Following application of SMOTE oversampling, the model predicted the Nikumaroro jar to be a non-float window.\n", "\n", "These machine learning model predictions would seem to indicate that the clear facsimile jar can be correctly identified through prudent enhancement of the model with oversampling techniques. These techniques, however, are ineffective in increasing the ability of the model to predict the Nikumaroro jar to be what it is, a container.\n", "\n", "Using a well-tuned model, the fact that a non-float window was predicted for the Nikumaroro jar demonstrates that this database of late 20th century glassware lacks enough relevant examples that would allow it to predict the Nikumaroro jar correctly.\n", "\n", "\\* Because the model is stochastic in nature, results will not be exactly the same each time.\n", "\n", "### Covariate Drift and Why it Matters\n", "This lack of relevant examples in the training data is a common phenomenon in machine learning, so common in fact that it has a name: covariate drift. Covariate drift is defined as a mismatch between the relative sizes or attributes of categories between the training and the validation data. There are many possible reasons for why covariate drift happens. The passage of time may have changed the population, or perhaps those assembling the training data did not obtain enough available samples from each population, or for whatever reason overlooked certain segments of a population altogether.\n", "\n", "When extrapolating information about different populations, whether those populations be glass types, people, or some other collection, we want the categories we survey to be representative of the populations upon which we wish to generalize. If our validation data is not representative of the data we have used to train our model, that is a problem that needs to be addressed.\n", "\n", "Before we can address it, however, we need to check whether it really exists, and, if so, how much it exists. We need to measure it." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Measuring Covariate Drift\n", "Shikhar Gupta, a data scientist at Amazon, has provided an excellent program/algorithm for obtaining quantitative measurement of covariate drift. His paper is here:\n", "https://towardsdatascience.com/how-dis-similar-are-my-train-and-test-data-56af3923de9b.\n", "I have modified his work by placing it inside a function that can generate multiple reports for different testing situations, and I have adjusted it for the particular situation of testing the U.K. database. I have also added in an ROC curve graph and a short report detailing the various data transformations that are happening as the function progresses.\n", "\n", "The basic idea behind this function is to combine the training features (X_train) and validation features (Y_train) dataset, stacking one on top of the other, into a new features dataset, x. We can then create a new target dataset, y, by creating, for every observation in x, a row in y with a new variable, is_train. The is_train variable is a boolean value indicating whether or not each observation came from the original training dataset (prior to combining training and validation). The values in y may be used to compare predicted results with the actual results from the new target dataset.\n", "\n", "Having combined training and validation into a new features dataset, x, and created a corresponding target dataset, y, we need a way to separate x and y horizontally into new training and validation datasets. We can create our new training and validation datasets by using stratified Kfold. Stratified Kfold will ensure that each subsection of the data that is produced will have the same ratio of training values (original training data) to validation values (original validation data). By the time all the folds have been produced, we will have created an array of predictions for every single observation in x. Each of these observations in x will have generated a prediction of whether it came from the original training or from the original validation. In essence, we are creating a 'blind test' in which the random probability that any given observation from x came from training or from validation is roughly equal.\n", "\n", "This procedure is somewhat different than the traditional one of creating a single validation \"holdout\" dataset we would then use to test the skill of a given model. Instead, what we are doing is fitting each component Kfold to our Random Forest model, assigning a prediction to each observation in the training set created, and adding those predictions to a master array containing a prediction for every observation in the original x.\n", "\n", "Each of the predictions that we generate will also have a probability associated with it. The probability calculates how likely is it that that the observation came from the training dataset. This probability can be used to create weighting factors that measure quantitatively a unit-less \"distance\" between each observation in the final Kfold and the original training data.\n", "\n", "To do this weighting, we filter the predictions on the row indexes of the training features dataset created from that last Kfold. This dataset of predictions contains the same ratio of training to validation as existed in the original x dataset. We can then plot the weights from this data onto a distribution plot. If the distribution of the weights in this Kfold shows a clear separation from the original training data, and if the number contained in that separated distribution is roughly equivalent to the number of observations in the Kfold that came from the original validation dataset, we will know then that the validation data we originally selected to test our model has caused covariate drift. If, on the other hand, the distribution shows a central tendency around the value of 1, we will know then that the validation data does not have covariate drift, that the origin of the data itself is not a predictive variable (as we would hope it would not be), and that in fact our validation data is well-enough matched to our training data that it may be used to generate valid, if not always accurate, predictions.\n", "\n", "Here is the modification of Shikhar Gupta's original program to measure covariate drift:" ] }, { "cell_type": "code", "execution_count": 111, "metadata": {}, "outputs": [], "source": [ "def covariate_drift(validation, title, splits):\n", " \"\"\"\n", " Authors: Shikhar Gupta and Joe Cerniglia\n", " Date: September 17, 2022\n", " A function to create metadata that can be used to measure covariate drift.\n", " This function generates two graphs:\n", " 1. a ROC curve showing the skill of the model at determining whether the data was\n", " from training or from validation\n", " 2. a distribution curve showing the degree to which training and validation differ.\n", " A report is also generated, including a confusion matrix.\n", " Parameters:\n", " validation: an array containing all of the validation examples we wish to compare\n", " against the training data.\n", " title: string to provide a title for the distribution curve\n", " splits: integer number of splits for the stratifiedKfold. The reciprocal of this number will \n", " determine the ratio of validation samples to training samples.\n", " \n", " returns None. The function produces graphs but does not return anything outside the function itself.\n", " \"\"\"\n", " # change validation to a pandas dataframe\n", " names=['Ref_ix','Na', 'Mg', 'Al', 'Si', 'K', 'Ca', 'Ba', 'Fe', 'GlassType']\n", " validation=pd.DataFrame(validation, columns=names)\n", " \n", " print('# training rows:',len(train))\n", " print('# validation rows:',len(validation))\n", " \n", " #adding a column to identify whether a row comes from the training set or not\n", " validation['is_train'] = 0\n", " train['is_train'] = 1\n", "\n", " #combining test and train data\n", " df_combine=pd.concat([train, validation], axis=0, ignore_index=True)\n", "\n", " #dropping 'target' column. We are creating a new target from the origin of the data: test or train\n", " df_combine = df_combine.drop('GlassType', axis=1)\n", "\n", " x = df_combine.drop('is_train', axis=1).values #leaves only independent variables\n", " y = df_combine['is_train'].values #labels: the new target variable is is_train\n", "\n", " print('Combined validation and train - feature variables (rows, columns):',x.shape)\n", " print('Combined validation and train - target variable (rows, columns):',y.shape)\n", "\n", " m = RandomForestClassifier(n_jobs=-1, max_depth=5, min_samples_leaf=5)\n", " predictions = np.zeros(y.shape) #creating an empty prediction array that is as large as the original\n", " #dataframe of combined values\n", "\n", " #SKF is stratified kfold\n", " skf=SKF(n_splits=splits, shuffle=True, random_state=100)\n", " #Stratified kfold cross validation is an extension of regular kfold cross validation but \n", " #specifically for classification problems where rather than completely random splits, \n", " #the ratio between the target classes (in this case, train/test) is the same in each fold as it \n", " #is in the full dataset.\n", " for fold, (train_idx, validation_idx) in enumerate(skf.split(x,y)):\n", " X_train, X_validation = x[train_idx], x[validation_idx] \n", " y_train, y_validation = y[train_idx], y[validation_idx]\n", " m.fit(X_train, y_train)\n", " probs = m.predict_proba(X_validation)[:, 1] #calculating the probability that is_train = 1\n", " # Each time through the loop a different set of rows in the training arrays is selected, fit to the\n", " # model, and a different (smaller) set of rows in the validation array is assigned a prediction,\n", " # based on the fitted model.\n", " # By the time the loop of 20 is finished, every observation in the combined training\n", " # and validation data has a prediction value.\n", " # The last fold will be the one selected for X_train and X_validation,\n", " # y_train and y_validation, but these arrays will not be needed later, except for counts.\n", " # Also, the training indexes of the final fold will be needed to create the distribution plot.\n", " # This last fold's weights, derived from predictions from the last training fold, will be the ones \n", " # used in the distribution plot.\n", " predictions[validation_idx] = probs\n", " \n", " # Short report about the results\n", " print('Confusion matrix for combined training + validation predictions:')\n", " print(confusion_matrix(y.astype(float), np.round(predictions,0)))\n", " print('# training rows after stratified kfold:',len(X_train))\n", " print('# validation rows after stratified kfold:',len(X_validation))\n", " print('Case mix for X_train after Kfold:')\n", " ylist = [str(x) for x in y_train.tolist()]\n", " ylist = map(lambda x: x.replace('0','from validation'),ylist)\n", " ylist = map(lambda x: x.replace('1','from training'),ylist)\n", " print(Counter(list(ylist)))\n", " \n", " # ROC curve of the entire set of predictions\n", " fpr, tpr, thresholds = metrics.roc_curve(y, predictions, pos_label=1)\n", " print('ROC-Area Under Curve score for train and validation distributions:',metrics.auc(fpr, tpr))\n", " sns.set_style(\"white\")\n", " plt.plot(fpr,tpr)\n", " plt.ylabel('True Positive Rate',fontsize=15)\n", " plt.xlabel('False Positive Rate',fontsize=15)\n", " plt.show()\n", " \n", " sns.set_style(\"whitegrid\")\n", " # Distribution plot of the training weights\n", " plt.figure(figsize=(10,5))\n", " predictions_train = predictions[train_idx] # Filter the predictions on the row indexes of the \n", " # training features dataset created from that last Kfold.\n", " print()\n", " print('# of total training samples in the graph:',len(predictions_train))\n", " weights = (1./predictions_train) - 1. \n", " weights /= np.mean(weights) # Normalizing the weights\n", " plt.xlabel('Computed sample weight \\n \\n The higher the weight, the more dissimilar ' + \n", " 'each kfold sample is to samples from the original training dataset',fontsize=18)\n", " plt.ylabel('# Samples',fontsize=18)\n", " plt.xlim([min(weights), max(weights)])\n", " plt.figtext(0.5, 1, title, wrap=True, horizontalalignment='center', fontsize=18)\n", " g=sns.histplot(weights, kde=False, color='mediumpurple')\n", " g.set_xticks(g.get_xticks()[1:])\n", " g.set_xticklabels(g.get_xticks(), size = 15)\n", " g.set_xticks(range(0,int(round(max(weights)+1,0)),1));\n", " plt.show()\n", " \n", " return " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Selecting the validation dataset to compare with the training set for covariate drift\n", "An important consideration in testing for coviariate drift will be how to build the validation dataset used to compare with the training dataset. The following function handles a variety of testing scenarios. One apparent problem in this testing is that we only have two jars with which to compare. They are the only validation examples that are of interest, since they are the ones that we know did not come from the original U.K. database.\n", "\n", "To make this small validation set comparable in size to the original U.K. database, we will want to copy the data obtained from these jars over and over again, and append each copy to the validation, until it reaches the size of the training dataset. This function repeats the copy process a set number of times until that desired size is reached." ] }, { "cell_type": "code", "execution_count": 112, "metadata": {}, "outputs": [], "source": [ "def set_validation(validation, repeat, run_repeat, type_repeat=None):\n", " '''Modifies a validation dataframe by appending to it repeated artifacts'''\n", " if not run_repeat:\n", " return\n", " else:\n", " for n in range(repeat):\n", " if type_repeat=='mix':\n", " validation.append(artifact_features+artifact_identity)\n", " validation.append(facsimile)\n", " elif type_repeat=='facsimiles':\n", " validation.append(facsimile_features+facsimile_identity)\n", " validation.append(facsimile)\n", " elif type_repeat=='artifacts':\n", " validation.append(artifact_features+artifact_identity)\n", " validation.append(artifact_features+artifact_identity)\n", " return" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Testing how the jar from Nikumaroro and the jar from eBay stack up to the U.K. database\n", "#### Elements of the Report\n", "Now we can run our first test for covariate drift. After using the set_validation function to create a validation dataset of 215 observations comprised of data from the Nikumaroro jar and the clear facsimile, we will then run that validation data through the main program. \n", "\n", "The first two lines returned by the report display the count of original U.K. database glass samples, which is 214. The number of validation rows created by the set_validation function is 215.\n", "\n", "Our next two lines provide verification that x (training) and y (validation) have been combined, as well as the resulting count of that combination.\n", "\n", "The next element in the report is the confusion matrix. This matrix can tell whether the model has been successful in separating training data from validation data. (See the section, How to Read a Confusion Matrix, for a better appreciation of how to interpret them.) If the validation and training row counts appear in the top left and lower right quadrants, respectively, then we know that covariate drift is present in the data.\n", "\n", "After this, we see the counts obtained from the last stratified Kfold. The Kfold subsets a training dataset (95%) from the x. The remaining observations (5%) from x belong to the validation dataset.\n", "\n", "We then want to verify that the ratio of training to validation in the original x and y (50:50) has remained in the stratified Kfold training dataset. The line that begins with the word 'Counter' will show that it has.\n", "\n", "The ROC-Area Under Curve score visually represents how skillful the model was in separating training (U.K. database) from validation (Nikumaroro jar and clear facsimile). If the score is 1.0 and the curve is a right angle that intersects in the top left quadrant, we know then that the model has been completely skillful in this task.\n", "\n", "The last exhibit is the distribution plot, which takes the training set from the last Kfold and bins the weights derived from that training set, which tells us how the observations in the training set compare with the original validation set. Those bins that center on 1.0 represent training examples that stand an equal chance of having come from original validation or training sets. (Ideally, all rows would hover at or near this point.) Those bins that center on 0.0 have a 100% probability of having come from the training set. The greater the positive distance the bins are from 1.0, the more unlike the original training set these observations are." ] }, { "cell_type": "code", "execution_count": 113, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "# training rows: 214\n", "# validation rows: 215\n", "Combined validation and train - feature variables (rows, columns): (429, 9)\n", "Combined validation and train - target variable (rows, columns): (429,)\n", "Confusion matrix for combined training + validation predictions:\n", "[[215 0]\n", " [ 0 214]]\n", "# training rows after stratified kfold: 408\n", "# validation rows after stratified kfold: 21\n", "Case mix for X_train after Kfold:\n", "Counter({'from training': 204, 'from validation': 204})\n", "ROC-Area Under Curve score for train and validation distributions: 1.0\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "# of total training samples in the graph: 408\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "train = dataset_ml.drop(['ID'], axis=1)\n", "# The artifact data are given as lists\n", "validation = [artifact_features+artifact_identity]\n", "facsimile = facsimile_features+facsimile_identity\n", "set_validation(validation, 107, True,'mix')\n", "covariate_drift(validation,'Test of whether adding repeated Nikumaroro jars and clear facsimiles \\n ' + \n", " 'as validation samples causes a covariate drift',20)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Analysis of covariate drift: Effect of Nikumaroro jar and clear facsimile\n", "The exhibits returned by the first test show unambiguously that the jar from Nikumaroro and the clear facsimile cause covariate drift in the analysis. The distribution plot shows two distinct groups, representing a clean division of samples between the U.K. database and the jars. The spread of the jar group is rather wide, indicating that some examples from the U.K. database are more similar to the jars than others." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Analysis of second test of covariate drift: Effect of the training data on itself\n", "Our last test is a control experiment. We want to know whether the training data itself contains some inherent covariate drift of its own. Does the training data from the U.K. database show internal consistency with itself? This is also a good way to check for how a dataset potentially without covariate drift should appear.\n", "\n", "This last test shows that the U.K. database is not perfectly consistent. Rather than the jagged bell curve that would appear in a highly normal distribution, our plot shows a bimodal distribution. This is most likely the result of two large categories of data within the U.K. database. Window and Non-window are the most likely separations that this plot is showing, with the smaller peaks toward the right possibly representing the less populated categories of tableware, headlamps, and containers.\n", "\n", "The overall picture drawn by this plot, however, is of a database without covariate drift." ] }, { "cell_type": "code", "execution_count": 114, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "# training rows: 214\n", "# validation rows: 214\n", "Combined validation and train - feature variables (rows, columns): (428, 9)\n", "Combined validation and train - target variable (rows, columns): (428,)\n", "Confusion matrix for combined training + validation predictions:\n", "[[ 7 207]\n", " [208 6]]\n", "# training rows after stratified kfold: 407\n", "# validation rows after stratified kfold: 21\n", "Case mix for X_train after Kfold:\n", "Counter({'from training': 204, 'from validation': 203})\n", "ROC-Area Under Curve score for train and validation distributions: 0.005983055288671494\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "# of total training samples in the graph: 407\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "train = dataset_ml.drop(['ID'], axis=1)\n", "## Test of whether the training dataset itself has covariate drift\n", "validation2 = train.copy()\n", "comparison=covariate_drift(validation2,'Test of whether training data itself has covariate drift',20)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Treating Covariate Drift\n", "Many articles offer treatment options for addressing covariate drift. The only true measure of the success of treatment is the success of the machine learning models. Was a given model able to identify the jars correctly? We have already seen that, using SMOTE oversampling, we were able to identify the clear facsimile correctly, despite the covariate drift that was in the validation samples. Would additional techniques be able to identify the Nikumaroro jar as a container? \n", "\n", "Albert Um, in an article for Medium, https://albertum.medium.com/covariate-shift-in-machine-learning-adf8d0077f79, has written a program that uses logistic regression to calculate weighting factors that may be used to weight the model to favor those training examples that are most like the validation set. The weights may then be used in the model to fit the original training examples and then predictions can be made on the jars that would presumably be more accurate than they would have been without the weights.\n", "\n", "### Setting up the data\n", "Here is how it works:\n", "The setup of the training dataset is identical to Shikhar Gupta's algorithm, albeit with slightly different syntax. We take the original dataframe of the U.K. database and drop the 'ID' and 'GlassType' variables, since these will not be needed in the analysis. Then we add our 'is_train' column to the data, populating it with '1', indicating that these observations came from training. We convert the training dataset to an array and split it vertically into features (X_train) and the target (Y_train), which is comprised solely of the is_train variable.\n", "\n", "We then build our validation dataset exactly as we did previously, appending multiple copies to itself until we have a dataframe that is the same size as the training dataset. Then we add the is_train column to the validation dataset, populating it with a value of '0'. We then convert the validation dataset to an array and split that array vertically into features (X_validation) and the target (Y_validation), which is comprised solely of the is_train variable." ] }, { "cell_type": "code", "execution_count": 115, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "# training rows: 214\n", "# validation rows: 213\n" ] } ], "source": [ "# Recover original training dataframe\n", "train = dataset_ml.drop(['ID','GlassType'], axis=1)\n", "# Assign new target\n", "train['is_train'] = 1\n", "\n", "# Convert training dataframe to an array\n", "array = train.values\n", "\n", "# Split training into X and Y\n", "X_train = array[:,0:9]\n", "Y_train = array[:,9]\n", "\n", "# Re-create validation consisting of artifacts\n", "validation = [artifact_features + artifact_identity]\n", "facsimile = facsimile_features + facsimile_identity\n", "\n", "# Enlarge the validation set by appending copies\n", "set_validation(validation, 106, True, 'mix')\n", "\n", "# Convert the list to a dataframe\n", "names=['Ref_ix','Na', 'Mg', 'Al', 'Si', 'K', 'Ca', 'Ba', 'Fe', 'GlassType']\n", "validation=pd.DataFrame(validation, columns=names)\n", " \n", "print('# training rows:',len(train))\n", "print('# validation rows:',len(validation))\n", "\n", "# Add the column to identify whether a row comes from train or not\n", "validation['is_train'] = 0\n", "\n", "# Convert validation dataframe to an array\n", "val_array=validation.values\n", "# Split validation into X and Y\n", "X_validation = val_array[:,0:9]\n", "Y_validation = val_array[:,10]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Creating the weights\n", "Now Mr. Um takes the two X and two Y arrays and stacks them using np.vstack and np.hstack. He creates two new arrays, named X_cov and y_cov, which combine, as previously, the training features and validation features (X_cov), and then, in a separate array, the training target and the validation target (y_cov). These functions create arrays in the format that logistic regression will be able to use in fitting the model. \n", "\n", "Next, he fits the stacked arrays, X_cov and y_cov to a logistic regression model. This model uses both training (U.K. database) and validation (the jars) to create a new model that encompasses our \"drifted\" examples. With the new model in hand, he uses the decision_function of logistic regression to calculate confidence scores on the original training sample.\n", "\n", "To reduce what Mr. Um calls \"outrageous weight assignments,\" he uses a cutoff value of 3 (or -3), which he applies conditionally to the array of confidence scores. All scores less than -3 are assigned a -3. All scores greater than 3 are assigned a 3. Mr. Um then exponentiates (raises Euler's number, 2.71828, to the power of each weight), using the np.exp function.\n", "\n", "We then have an array of weights that correspond with each row of the original training sample." ] }, { "cell_type": "code", "execution_count": 116, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "# of weights: 214\n" ] } ], "source": [ "# Albert Um\n", "# Fitting Logistic Regression to obtain weights similar to test\n", "\n", "# vertical stack feature matrix\n", "X_cov = np.vstack([X_train, X_validation])\n", "\n", "# horizontal stack label array\n", "y_cov = np.hstack([Y_train, Y_validation])\n", "\n", "# Instantiate a binary classifier for confidence scores\n", "weights_logreg = LogisticRegression(max_iter=2000000)\n", "weights_logreg.fit(X_cov, y_cov)\n", "\n", "# Calculate the weights for X_train, not to X_train + X_validation,\n", "# but based on the fit of the covariate stacks, which include training + validation\n", "# This brings the count to the original 214.\n", "# confidence scores\n", "conf_scores = weights_logreg.decision_function(X_train)\n", "#print(conf_scores)\n", "# Clipping (C = Cutoff, originally set at 3)\n", "C = 3\n", "conditions = [\n", " conf_scores < -C,\n", " conf_scores > C\n", "]\n", "\n", "choices = [\n", " -C,\n", " C\n", "]\n", "conf_scores = np.select(conditions, choices, conf_scores)\n", "# exponentiate the confidence scores\n", "weights = np.exp(conf_scores)\n", "weights = weights.flatten()\n", "print('# of weights:',len(weights))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Fitting the model with the weights to the training sample\n", "We can now bring back our original U.K. database as an array. Split the array into features (x) and target (y), using the original target variable, GlassType.\n", "\n", "We then create stratified Kfolds of X_train (training features), X_validation (validation features), Y_train (training target), and Y_validation (validation target). We then fit our favorite model, Random Forests, to the training features and training target, using the sample weights we derived from the earlier fitting of the logistic regression model to the data that included both the U.K. dataset and the jars." ] }, { "cell_type": "code", "execution_count": 117, "metadata": {}, "outputs": [], "source": [ "# Bring back the original dataset with the original target column: GlassType\n", "# so that we can apply the weights above to test their predictive ability\n", "array = dataset_ml.values\n", "x = array[:,1:10]\n", "\n", "y = array[:,10]\n", "\n", "#print('length X_train:',len(X_train))\n", "#print('length weights:',len(weights))\n", "\n", "# We use a folded split to fit the model on the training data only, \n", "# thus avoiding data leakage in the fitting of the model.\n", "# The validation datasets must be naive to the training datasets to prevent data leakage.\n", "\n", "# The splits=4 allows the validation set to be about 25% of the size of the training set.\n", "skf=SKF(n_splits=4, shuffle=True, random_state=100)\n", "for fold, (train_idx, validation_idx) in enumerate(skf.split(x,y)):\n", " X_train, X_validation = x[train_idx], x[validation_idx] \n", " Y_train, Y_validation = y[train_idx], y[validation_idx]\n", "\n", "#print(len(X_train),len(X_validation),len(Y_train),len(Y_validation))\n", "#print(Y_train)\n", "# The idea is to weight more heavily those training examples that are \n", "# more similar to the artifacts\n", "# Fit a new model with the logistic regression's custom weights\n", "model = RandomForestClassifier(n_jobs=-1,max_depth=5)\n", "model.fit(X_train, Y_train, sample_weight = weights[train_idx]);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Assessing the skill of the weighted model\n", "Now it is time to see how this new model with weights performs against covariate drift. This next step is rather simple. We can use the model fitted in the prior step to make predictions on the validation features (jars), then print a confusion matrix and a classification report to assess the results." ] }, { "cell_type": "code", "execution_count": 118, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Random Forest with sample_weight\n", "[[12 5 0 0 0 0]\n", " [ 4 13 0 1 0 1]\n", " [ 2 2 0 0 0 0]\n", " [ 0 3 0 1 0 0]\n", " [ 0 1 0 0 1 0]\n", " [ 0 0 0 0 0 7]]\n", " precision recall f1-score support\n", "\n", " 1.0 0.67 0.71 0.69 17\n", " 2.0 0.54 0.68 0.60 19\n", " 3.0 1.00 0.00 0.00 4\n", " 5.0 0.50 0.25 0.33 4\n", " 6.0 1.00 0.50 0.67 2\n", " 7.0 0.88 1.00 0.93 7\n", "\n", " accuracy 0.64 53\n", " macro avg 0.76 0.52 0.54 53\n", "weighted avg 0.67 0.64 0.61 53\n", "\n" ] } ], "source": [ "yhat = model.predict(X_validation)\n", "print('Random Forest with sample_weight')\n", "print(confusion_matrix(Y_validation, yhat))\n", "print(classification_report(Y_validation, yhat, zero_division=1))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Unfortunately, our model performs a good deal worse than both our SMOTE and non-SMOTE models used earlier. On the Nikumaroro jar, it computes a minimal probability that the Nikumaroro jar is a container. On the clear facsimile, the results are very similar." ] }, { "cell_type": "code", "execution_count": 119, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[30mNikumaroro jar prediction: \u001b[30mWindow Float\n", "\u001b[30mThe artifact has a probability of:\n", "\u001b[30m50% to be a \u001b[30mWindow Float\n", "\u001b[30m25% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m16% to be a \u001b[30mHeadlamp\n", "\u001b[30m9% to be a \u001b[30mVehicle Float\n", "\u001b[30m0% to be a \u001b[31mContainer\n", "\u001b[30m0% to be a \u001b[30mTableware\n", "\n", "\u001b[30mclear facsimile prediction: \u001b[30mWindow Non-Float\n", "\u001b[30mThe clear facsimile has a probability of:\n", "\u001b[30m29% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m22% to be a \u001b[30mWindow Float\n", "\u001b[30m18% to be a \u001b[30mHeadlamp\n", "\u001b[30m15% to be a \u001b[30mTableware\n", "\u001b[30m10% to be a \u001b[31mContainer\n", "\u001b[30m7% to be a \u001b[30mVehicle Float\n" ] } ], "source": [ "yhat = model.predict([artifact_features])\n", "yhat2 = model.predict([facsimile_features])\n", "yhat_probability = model.predict_proba([artifact_features])\n", "yhat2_probability = model.predict_proba([facsimile_features])\n", "the_report()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Using the same Random Forests model without weights as a control, we find that the skill of this model is about the same as that of the model with weights, and the clear facsimile and Nikumaroro jar are still assigned a very low probability as containers." ] }, { "cell_type": "code", "execution_count": 120, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Random Forest without sample_weight\n", "[[14 3 0 0 0 0]\n", " [ 4 14 0 0 0 1]\n", " [ 2 1 1 0 0 0]\n", " [ 0 3 0 1 0 0]\n", " [ 0 1 0 0 1 0]\n", " [ 0 0 0 0 0 7]]\n", " precision recall f1-score support\n", "\n", " 1.0 0.70 0.82 0.76 17\n", " 2.0 0.64 0.74 0.68 19\n", " 3.0 1.00 0.25 0.40 4\n", " 5.0 1.00 0.25 0.40 4\n", " 6.0 1.00 0.50 0.67 2\n", " 7.0 0.88 1.00 0.93 7\n", "\n", " accuracy 0.72 53\n", " macro avg 0.87 0.59 0.64 53\n", "weighted avg 0.76 0.72 0.70 53\n", "\n" ] } ], "source": [ "modelnoWeight = RandomForestClassifier(n_jobs=-1,max_depth=5)\n", "modelnoWeight.fit(X_train, Y_train)\n", "yhat = modelnoWeight.predict(X_validation)\n", "\n", "print('Random Forest without sample_weight')\n", "print(confusion_matrix(Y_validation,yhat))\n", "print(classification_report(Y_validation, yhat, zero_division=1))" ] }, { "cell_type": "code", "execution_count": 121, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[30mNikumaroro jar prediction: \u001b[30mWindow Float\n", "\u001b[30mThe artifact has a probability of:\n", "\u001b[30m50% to be a \u001b[30mWindow Float\n", "\u001b[30m22% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m17% to be a \u001b[30mHeadlamp\n", "\u001b[30m11% to be a \u001b[30mVehicle Float\n", "\u001b[30m0% to be a \u001b[31mContainer\n", "\u001b[30m0% to be a \u001b[30mTableware\n", "\n", "\u001b[30mclear facsimile prediction: \u001b[30mWindow Non-Float\n", "\u001b[30mThe clear facsimile has a probability of:\n", "\u001b[30m28% to be a \u001b[30mWindow Non-Float\n", "\u001b[30m19% to be a \u001b[30mHeadlamp\n", "\u001b[30m18% to be a \u001b[30mWindow Float\n", "\u001b[30m14% to be a \u001b[30mVehicle Float\n", "\u001b[30m11% to be a \u001b[31mContainer\n", "\u001b[30m11% to be a \u001b[30mTableware\n" ] } ], "source": [ "yhat = modelnoWeight.predict([artifact_features])\n", "yhat2 = modelnoWeight.predict([facsimile_features])\n", "yhat_probability = modelnoWeight.predict_proba([artifact_features])\n", "yhat2_probability = modelnoWeight.predict_proba([facsimile_features])\n", "the_report()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Why Did the Weighting Fail to Improve the Model?\n", "On an intuitive level, the low skill of this weighted model can be understood by looking back at the scatter plot we used to demonstrate SMOTE. Let's recreate this scatter plot now." ] }, { "cell_type": "code", "execution_count": 126, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sns.set_style(\"white\")\n", "smotegraph(' - after SMOTE oversampling','sm')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For the weighting model to be successful, two things need to happen. \n", "\n", "First, samples in the original training set (U.K. database) that have somewhat similar features to the validation set (jars) must exist. Notice the upper left quadrant of the SMOTE graph above. There are only a few training examples in that region. If, for example, our jars have features that place them in the upper left corner of the graph, there will be no training examples there to weight in favor of the validation examples. \n", "\n", "Second, at least a few of the samples in the original training set (U.K. database) that share similar features to the validation set (jars) must actually **BE** jars. If they are not jars, then weighting these examples will only cause the model to reinforce and amplify its misinterpretations. \n", "\n", "Judging by the number of proposed treatment options, it would seem that there is no single magic bullet to the problem of covariate drift. The best approach is to experiment, with the realization that some validation samples may have drifted beyond the reach of rescue by a skillful model. In essence, the Nikumaroro jar might be regarded not as an example of covariate drift, but rather as a paradigm shift, a combination of features too unique to be generalizable to the world of glass types available in 1987.\n", "\n", "The most basic treatment option, and the preferred one, for covariate drift is to find new data examples to add to the training set that will diversify it and make it more adaptable to training it for unanticipated validation examples." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Conclusions\n", "Our task, then, would seem to be to supplement the 1987 U.K. database with chemical profiles of jars from the early twentieth century that are similar to the Nikumaroro jar. \n", "\n", "But could these examples be found? Those of us who analyzed the Nikumaroro jar have searched for more than ten years and have not so far found any siblings that match that jar's chemical profile. The search, however, continues.\n", "\n", "I derive two lessons from this machine learning project:\n", "1. Machine learning, at least of the type developed here, is not without failure or flaw. It can sometimes fail spectacularly.\n", "\n", "2. Flawlessness, however, is neither very interesting nor very attainable, and in any case, it may not even be necessary. We can learn more sometimes by a model's failure to learn than by its success.\n", "\n", "\n", "We have learned that our machine learning model can correctly characterize the clear facsimile, a jar of the same size and shape as the Nikumaroro jar, and possibly from the same era. The model cannot correctly characterize the semi-opaque Nikumaroro jar. One might suppose that this jar has a rare and original recipe that, in the absence of any information other than the data nourishing the model, is not readily identifiable as to the glass type, even with the powerful machine learning tools, such as random forests, available to us in 2022. While this may not be confirmative as to whether or not the jar was owned and brought to Nikumaroro Island by Amelia Earhart, this fact is further confirmation of its originality and rarity, perhaps even in its own time." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Notes\n", "[1] Wallenberger, Frederick T. and Bingham, Paul A., ed. Fiberglass and Glass Technology: Energy-Friendly Compositions and Applications. New York: Springer Science and Business Media, 2010, p. 323.
\n", "[2] Caddy, Brian, ed. Forensic Examination of Glass and Paint. London: Taylor & Francis, 2001, p. 61.
\n", "[3] Kaur, Gurbinder. Bioactive Glasses: Potential Biomaterials for Future Therapy. Germany: Springer International Publishing, 2017, p. 107." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Acknowledgements\n", "I owe much of what I have learned to date in machine learning to Dr. Jason Brownlee:\n", "https://machinelearningmastery.com/machine-learning-in-python-step-by-step/\n", "\n", "I owe a debt of gratitude to reviewers from The Alan Turing Institute. Jennifer Ding, Research Application Manager and Eirini Zormpa, Community Manager of Open Collaboration, kindly offered numerous suggestions for clarifying both the code and the writing of this paper.\n", "\n", "Thank you to all of the members of the Turing Data Stories Wednesday Meeting Group, who also offered many insights, camaraderie and support through many months of the process of creating new and interesting Turing Data Stories, and the work continues.\n", "\n", "Thank you to Ric Gillespie of The International Group for Historic Aircraft Recovery for graciously allowing the sharing of this research with The Alan Turing Institute.\n", "\n", "Thank you to Ian W. Evett and Ernest J. Spiehler, for making the data from their 1987 glass study available for students and experienced data scientists alike to use.\n", "\n", "Thank you to my co-authors of the paper \"A Freckle In Time,\" which was written from 2010-2013: Greg George, Senior Staff Chemist at Persedo Spirits, Bill Lockhart, former assistant professor of sociology at New Mexico State University at Alamogordo, and Thomas Fulling King, former director of the United States Advisory Council on Historic Preservation." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.9.12" } }, "nbformat": 4, "nbformat_minor": 2 }