{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": { "collapsed": true }, "outputs": [], "source": [ "# imports\n", "import h2o \n", "import numpy as np\n", "import pandas as pd\n", "from h2o.estimators.xgboost import H2OXGBoostEstimator\n", "from h2o.estimators.random_forest import H2ORandomForestEstimator\n", "from h2o.grid.grid_search import H2OGridSearch" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "collapsed": true }, "outputs": [], "source": [ "# display matplotlib graphics in notebook\n", "%matplotlib inline" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Checking whether there is an H2O instance running at http://localhost:54321..... not found.\n", "Attempting to start a local H2O server...\n", " Java Version: java version \"1.8.0_112\"; Java(TM) SE Runtime Environment (build 1.8.0_112-b16); Java HotSpot(TM) 64-Bit Server VM (build 25.112-b16, mixed mode)\n", " Starting server from /Users/phall/anaconda/lib/python3.5/site-packages/h2o/backend/bin/h2o.jar\n", " Ice root: /var/folders/tc/0ss1l73113j3wdyjsxmy1j2r0000gn/T/tmpxmj_ouok\n", " JVM stdout: /var/folders/tc/0ss1l73113j3wdyjsxmy1j2r0000gn/T/tmpxmj_ouok/h2o_phall_started_from_python.out\n", " JVM stderr: /var/folders/tc/0ss1l73113j3wdyjsxmy1j2r0000gn/T/tmpxmj_ouok/h2o_phall_started_from_python.err\n", " Server is running at http://127.0.0.1:54321\n", "Connecting to H2O server at http://127.0.0.1:54321... successful.\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", "
H2O cluster uptime:01 secs
H2O cluster version:3.14.0.2
H2O cluster version age:1 month and 22 days
H2O cluster name:H2O_from_python_phall_8bk9h3
H2O cluster total nodes:1
H2O cluster free memory:3.556 Gb
H2O cluster total cores:8
H2O cluster allowed cores:8
H2O cluster status:accepting new members, healthy
H2O connection url:http://127.0.0.1:54321
H2O connection proxy:None
H2O internal security:False
H2O API Extensions:XGBoost, Algos, AutoML, Core V3, Core V4
Python version:3.5.2 final
" ], "text/plain": [ "-------------------------- ----------------------------------------\n", "H2O cluster uptime: 01 secs\n", "H2O cluster version: 3.14.0.2\n", "H2O cluster version age: 1 month and 22 days\n", "H2O cluster name: H2O_from_python_phall_8bk9h3\n", "H2O cluster total nodes: 1\n", "H2O cluster free memory: 3.556 Gb\n", "H2O cluster total cores: 8\n", "H2O cluster allowed cores: 8\n", "H2O cluster status: accepting new members, healthy\n", "H2O connection url: http://127.0.0.1:54321\n", "H2O connection proxy:\n", "H2O internal security: False\n", "H2O API Extensions: XGBoost, Algos, AutoML, Core V3, Core V4\n", "Python version: 3.5.2 final\n", "-------------------------- ----------------------------------------" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# start and connect to h2o server\n", "h2o.init()" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "collapsed": true }, "outputs": [], "source": [ "# location of clean data file\n", "path = '/Users/phall/workspace/h2o-tutorials/training/h2o_algos/data/loan.csv'" ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "collapsed": true }, "outputs": [], "source": [ "# define input variable measurement levels \n", "# strings automatically parsed as enums (nominal)\n", "# numbers automatically parsed as numeric\n", "col_types = {'bad_loan': 'enum'}" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Parse progress: |█████████████████████████████████████████████████████████| 100%\n" ] } ], "source": [ "frame = h2o.import_file(path=path, col_types=col_types) # import from url" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Rows:163987\n", "Cols:15\n", "\n", "\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", "
loan_amnt term int_rate emp_length home_ownership annual_inc purpose addr_state dti delinq_2yrs revol_util total_acc bad_loan longest_credit_length verification_status
type int enum real int enum real enum enum real int real int enum int enum
mins 500.0 5.42 0.0 1896.0 0.0 0.0 0.0 1.0 0.0
mean 13074.169141456332 13.7159040655661785.684352932995327 71915.67051974907 15.8815301212900960.2273570060625282654.07917280242255 24.579733834274634 14.854273655448347
maxs 35000.0 26.06000000000000210.0 7141778.0 39.99 29.0 150.70000000000002118.0 65.0
sigma 7993.556188734652 4.391939870545798 3.610663731100237 59070.91565491826 7.587668224192545 0.6941679229284183 25.28536676677049 11.685190365910657 6.947732922546695
zeros 0 0 14248 0 270 139459 1562 0 11
missing0 0 0 5804 0 4 0 0 0 29 193 29 0 29 0
0 5000.0 36 months10.65 10.0 RENT 24000.0 credit_card AZ 27.6500000000000020.0 83.7 9.0 0 26.0 verified
1 2500.0 60 months15.27 0.0 RENT 30000.0 car GA 1.0 0.0 9.4 4.0 1 12.0 verified
2 2400.0 36 months15.96 10.0 RENT 12252.0 small_business IL 8.72 0.0 98.5 10.0 0 10.0 not verified
3 10000.0 36 months13.49 10.0 RENT 49200.0 other CA 20.0 0.0 21.0 37.0 0 15.0 verified
4 5000.0 36 months7.9 3.0 RENT 36000.0 wedding AZ 11.2000000000000010.0 28.3 12.0 0 7.0 verified
5 3000.0 36 months18.64 9.0 RENT 48000.0 car CA 5.35000000000000050.0 87.5 4.0 0 4.0 verified
6 5600.0 60 months21.28 4.0 OWN 40000.0 small_business CA 5.55 0.0 32.6 13.0 1 7.0 verified
7 5375.0 60 months12.69 0.0 RENT 15000.0 other TX 18.0800000000000020.0 36.5 3.0 1 7.0 verified
8 6500.0 60 months14.65 5.0 OWN 72000.0 debt_consolidationAZ 16.12 0.0 20.6 23.0 0 13.0 not verified
9 12000.0 36 months12.69 10.0 OWN 75000.0 debt_consolidationCA 10.78 0.0 67.10000000000001 34.0 0 22.0 verified
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "frame.describe() # summarize data" ] }, { "cell_type": "code", "execution_count": 8, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# split into 40% training, 30% validation, and 30% test\n", "train, valid, test = frame.split_frame([0.4, 0.3])" ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "bad_loan\n", "['loan_amnt', 'term', 'int_rate', 'emp_length', 'home_ownership', 'annual_inc', 'purpose', 'addr_state', 'dti', 'delinq_2yrs', 'revol_util', 'total_acc', 'longest_credit_length', 'verification_status']\n" ] } ], "source": [ "# assign target and inputs\n", "y = 'bad_loan'\n", "X = [name for name in frame.columns if name != y]\n", "print(y)\n", "print(X)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# set target to factor - for binary classification\n", "# just to be safe ... \n", "train[y] = train[y].asfactor()\n", "valid[y] = valid[y].asfactor()\n", "test[y] = test[y].asfactor()" ] }, { "cell_type": "code", "execution_count": 11, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "drf Model Build progress: |███████████████████████████████████████████████| 100%\n" ] } ], "source": [ "# random forest\n", "# random forest is often best guess model with little tuning\n", "\n", "# initialize rf model\n", "rf_model = H2ORandomForestEstimator(\n", " ntrees=50, # Up to 500 decision trees in the forest \n", " max_depth=30, # trees can grow to depth of 30\n", " stopping_rounds=5, # stop after validation error does not decrease for 5 iterations/new trees\n", " score_each_iteration=True, # score validation error on every iteration/new tree\n", " model_id='rf_model', # for easy lookup in flow\n", " seed=12345) \n", "\n", "# train rf model\n", "rf_model.train(\n", " x=X,\n", " y=y,\n", " training_frame=train,\n", " validation_frame=valid)\n", "\n", "# print model information\n", "# rf_model\n", "\n", "# view detailed results at http://ip:port/flow/index.html" ] }, { "cell_type": "code", "execution_count": 12, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.6555005185496887\n", "0.6787530806751643\n", "0.680813561999385\n" ] } ], "source": [ "# measure rf AUC\n", "print(rf_model.auc(train=True))\n", "print(rf_model.auc(valid=True))\n", "print(rf_model.model_performance(test_data=test).auc())" ] }, { "cell_type": "code", "execution_count": 13, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "xgboost Grid Build progress: |████████████████████████████████████████████| 100%\n" ] } ], "source": [ "# GBM is often more accurate the RF, but requires more tuning\n", "# GBM with random hyperparameter search\n", "# train many different GBM models with random hyperparameters\n", "# and select best model based on validation error\n", "\n", "# define random grid search parameters\n", "hyper_parameters = {'ntrees':list(range(50, 500, 50)),\n", " 'max_depth':list(range(2, 20, 2)),\n", " 'sample_rate':[s/float(10) for s in range(1, 11)],\n", " 'col_sample_rate':[s/float(10) for s in range(1, 11)]}\n", "\n", "# define search strategy\n", "search_criteria = {'strategy':'RandomDiscrete',\n", " 'max_models':10,\n", " 'max_runtime_secs':300}\n", "\n", "# initialize grid search\n", "gsearch = H2OGridSearch(H2OXGBoostEstimator,\n", " hyper_params=hyper_parameters,\n", " search_criteria=search_criteria)\n", "\n", "# execute training w/ grid search\n", "gsearch.train(x=X,\n", " y=y,\n", " training_frame=train,\n", " validation_frame=valid)\n", "\n", "# view detailed results at http://ip:port/flow/index.html" ] }, { "cell_type": "code", "execution_count": 14, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " col_sample_rate max_depth ntrees sample_rate \\\n", "0 0.7 2 250 0.6 \n", "1 0.7 10 450 0.7 \n", "2 0.3 8 250 0.9 \n", "3 0.3 8 200 0.6 \n", "4 0.5 4 450 0.1 \n", "5 0.2 16 100 1.0 \n", "6 0.9 12 150 0.8 \n", "7 0.5 16 350 0.8 \n", "8 0.5 12 300 0.4 \n", "9 0.8 16 300 0.5 \n", "\n", " model_ids \\\n", "0 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_8 \n", "1 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_9 \n", "2 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_0 \n", "3 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_3 \n", "4 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_1 \n", "5 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_2 \n", "6 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_4 \n", "7 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_5 \n", "8 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_7 \n", "9 Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_6 \n", "\n", " logloss \n", "0 0.43700082401861967 \n", "1 0.46780467370872925 \n", "2 0.5090760423497354 \n", "3 0.5176778402024944 \n", "4 0.5640265467535642 \n", "5 0.5842232443528029 \n", "6 0.6135183619936883 \n", "7 0.7792439894090176 \n", "8 0.8144849310050937 \n", "9 0.8192994507825724 \n", "Model Details\n", "=============\n", "H2OXGBoostEstimator : XGBoost\n", "Model Key: Grid_XGBoost_py_6_sid_ae78_model_python_1507998691868_205_model_8\n", "\n", "\n", "ModelMetricsBinomial: xgboost\n", "** Reported on train data. **\n", "\n", "MSE: 0.1338834552544421\n", "RMSE: 0.3659008817349886\n", "LogLoss: 0.4245879659541619\n", "Mean Per-Class Error: 0.33333585720921455\n", "AUC: 0.7299580222631977\n", "Gini: 0.45991604452639545\n", "Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.21067449373898106: \n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "
01ErrorRate
038741.014993.00.279 (14993.0/53734.0)
14736.07374.00.3911 (4736.0/12110.0)
Total43477.022367.00.2996 (19729.0/65844.0)
" ], "text/plain": [ " 0 1 Error Rate\n", "----- ----- ----- ------- -----------------\n", "0 38741 14993 0.279 (14993.0/53734.0)\n", "1 4736 7374 0.3911 (4736.0/12110.0)\n", "Total 43477 22367 0.2996 (19729.0/65844.0)" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Maximum Metrics: Maximum metrics at their respective thresholds\n", "\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", "
metricthresholdvalueidx
max f10.21067450.4277634202.0
max f20.12476200.5801730286.0
max f0point50.30280280.4039439136.0
max accuracy0.48833250.819831753.0
max precision0.86339681.00.0
max recall0.01192791.0396.0
max specificity0.86339681.00.0
max absolute_mcc0.21067450.2698607202.0
max min_per_class_accuracy0.19119870.6651096218.0
max mean_per_class_accuracy0.18654570.6666641222.0
" ], "text/plain": [ "metric threshold value idx\n", "--------------------------- ----------- -------- -----\n", "max f1 0.210674 0.427763 202\n", "max f2 0.124762 0.580173 286\n", "max f0point5 0.302803 0.403944 136\n", "max accuracy 0.488333 0.819832 53\n", "max precision 0.863397 1 0\n", "max recall 0.0119279 1 396\n", "max specificity 0.863397 1 0\n", "max absolute_mcc 0.210674 0.269861 202\n", "max min_per_class_accuracy 0.191199 0.66511 218\n", "max mean_per_class_accuracy 0.186546 0.666664 222" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Gains/Lift Table: Avg response rate: 18.39 %\n", "\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", "\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", "
groupcumulative_data_fractionlower_thresholdliftcumulative_liftresponse_ratecumulative_response_ratecapture_ratecumulative_capture_rategaincumulative_gain
10.01000850.54290353.54776713.54776710.65250380.65250380.03550780.0355078254.7767117254.7767117
20.02000180.49130482.88384293.21605710.53039510.59149580.02881920.0643270188.3842889221.6057062
30.03001030.46011422.57419383.00199440.47344460.55212550.02576380.0900908157.4193815200.1994370
40.04000360.43544802.53679022.88578160.46656530.53075170.02535090.1154418153.6790163188.5781626
50.05001210.41692322.65670002.83993750.48861910.52232010.02658960.1420314165.6700027183.9937480
60.10000910.35297992.20657502.52330430.40583230.46408500.11032200.2523534120.6575007152.3304334
70.15000610.30840951.83165552.29277800.33687730.42168670.09157720.343930683.1655451129.2778049
80.20000300.27496541.58721452.11640060.29191980.38924750.07935590.423286558.7214507111.6400558
90.29999700.22661681.35928981.86404310.250.34283400.13592070.559207335.928984386.4043096
100.40000610.18921381.11055121.67565580.20425210.30818590.11106520.670272511.055115567.5655806
110.50.15876280.93482141.52749790.17193200.28093680.09347650.7637490-6.517855352.7497936
120.59999390.13238330.75231661.39830760.13836570.25717610.07522710.8389761-24.768344639.8307579
130.70000300.10856170.63495451.28924770.11678060.23711790.06350120.9024773-36.504547328.9247679
140.79999700.08326260.51448211.19240750.09462330.21930710.05144510.9539224-48.551787819.2407500
150.89999090.05236890.31463511.09488220.05786760.20137030.03146160.9853840-68.53648669.4882175
161.00.00320930.14614691.00.02687930.18391960.01461601.0-85.38531190.0
" ], "text/plain": [ " group cumulative_data_fraction lower_threshold lift cumulative_lift response_rate cumulative_response_rate capture_rate cumulative_capture_rate gain cumulative_gain\n", "-- ------- -------------------------- ----------------- -------- ----------------- --------------- -------------------------- -------------- ------------------------- -------- -----------------\n", " 1 0.0100085 0.542903 3.54777 3.54777 0.652504 0.652504 0.0355078 0.0355078 254.777 254.777\n", " 2 0.0200018 0.491305 2.88384 3.21606 0.530395 0.591496 0.0288192 0.064327 188.384 221.606\n", " 3 0.0300103 0.460114 2.57419 3.00199 0.473445 0.552126 0.0257638 0.0900908 157.419 200.199\n", " 4 0.0400036 0.435448 2.53679 2.88578 0.466565 0.530752 0.0253509 0.115442 153.679 188.578\n", " 5 0.0500121 0.416923 2.6567 2.83994 0.488619 0.52232 0.0265896 0.142031 165.67 183.994\n", " 6 0.100009 0.35298 2.20658 2.5233 0.405832 0.464085 0.110322 0.252353 120.658 152.33\n", " 7 0.150006 0.30841 1.83166 2.29278 0.336877 0.421687 0.0915772 0.343931 83.1655 129.278\n", " 8 0.200003 0.274965 1.58721 2.1164 0.29192 0.389247 0.0793559 0.423287 58.7215 111.64\n", " 9 0.299997 0.226617 1.35929 1.86404 0.25 0.342834 0.135921 0.559207 35.929 86.4043\n", " 10 0.400006 0.189214 1.11055 1.67566 0.204252 0.308186 0.111065 0.670273 11.0551 67.5656\n", " 11 0.5 0.158763 0.934821 1.5275 0.171932 0.280937 0.0934765 0.763749 -6.51786 52.7498\n", " 12 0.599994 0.132383 0.752317 1.39831 0.138366 0.257176 0.0752271 0.838976 -24.7683 39.8308\n", " 13 0.700003 0.108562 0.634955 1.28925 0.116781 0.237118 0.0635012 0.902477 -36.5045 28.9248\n", " 14 0.799997 0.0832626 0.514482 1.19241 0.0946233 0.219307 0.0514451 0.953922 -48.5518 19.2408\n", " 15 0.899991 0.0523689 0.314635 1.09488 0.0578676 0.20137 0.0314616 0.985384 -68.5365 9.48822\n", " 16 1 0.00320933 0.146147 1 0.0268793 0.18392 0.014616 1 -85.3853 0" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "\n", "ModelMetricsBinomial: xgboost\n", "** Reported on validation data. **\n", "\n", "MSE: 0.1378801676837094\n", "RMSE: 0.3713221885151888\n", "LogLoss: 0.43700082401861967\n", "Mean Per-Class Error: 0.3511826644261655\n", "AUC: 0.7032413981615196\n", "Gini: 0.4064827963230393\n", "Confusion Matrix (Act/Pred) for max f1 @ threshold = 0.19611620596226523: \n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "\n", "
01ErrorRate
026876.013057.00.327 (13057.0/39933.0)
13425.05555.00.3814 (3425.0/8980.0)
Total30301.018612.00.337 (16482.0/48913.0)
" ], "text/plain": [ " 0 1 Error Rate\n", "----- ----- ----- ------- -----------------\n", "0 26876 13057 0.327 (13057.0/39933.0)\n", "1 3425 5555 0.3814 (3425.0/8980.0)\n", "Total 30301 18612 0.337 (16482.0/48913.0)" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Maximum Metrics: Maximum metrics at their respective thresholds\n", "\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", "
metricthresholdvalueidx
max f10.19611620.4026529215.0
max f20.11351290.5664545297.0
max f0point50.29701800.3685852139.0
max accuracy0.58800550.817451425.0
max precision0.93703011.00.0
max recall0.01088051.0396.0
max specificity0.93703011.00.0
max absolute_mcc0.24976860.2359822170.0
max min_per_class_accuracy0.18736630.6451002223.0
max mean_per_class_accuracy0.17906700.6488173230.0
" ], "text/plain": [ "metric threshold value idx\n", "--------------------------- ----------- -------- -----\n", "max f1 0.196116 0.402653 215\n", "max f2 0.113513 0.566454 297\n", "max f0point5 0.297018 0.368585 139\n", "max accuracy 0.588005 0.817451 25\n", "max precision 0.93703 1 0\n", "max recall 0.0108805 1 396\n", "max specificity 0.93703 1 0\n", "max absolute_mcc 0.249769 0.235982 170\n", "max min_per_class_accuracy 0.187366 0.6451 223\n", "max mean_per_class_accuracy 0.179067 0.648817 230" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Gains/Lift Table: Avg response rate: 18.36 %\n", "\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", "\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", "
groupcumulative_data_fractionlower_thresholdliftcumulative_liftresponse_ratecumulative_response_ratecapture_ratecumulative_capture_rategaincumulative_gain
10.01001780.53979312.84571792.84571790.52244900.52244900.02850780.0285078184.5717922184.5717922
20.02001510.49079602.57306692.70953170.47239260.49744640.02572380.0542316157.3066938170.9531680
30.03001250.46132542.35029062.58986620.43149280.47547680.02349670.0777283135.0290580158.9866218
40.04000980.43947012.23890242.50217010.41104290.45937660.02238310.1001114123.8902401150.2170098
50.05000720.41995942.20548592.44285750.40490800.44848730.02204900.1221604120.5485947144.2857527
60.10001430.35564362.07320002.25802870.38062140.41455440.10367480.2258352107.3199961125.8028744
70.15000100.31154241.73542782.08387590.31860940.38258140.08674830.312583573.5427831108.3875925
80.20000820.27760151.61001461.96539850.29558460.36083000.08051220.393095861.001457896.5398479
90.30000200.22627101.32747561.75277200.24371290.32179360.13273940.525835232.747562875.2772020
100.39999590.18880341.13035891.59717670.20752400.29322770.11302900.638864113.035886159.7176683
110.50001020.15874700.97981521.47368920.17988550.27055650.09799550.7368597-2.018476647.3689248
120.60000410.13293760.81630841.36413320.14986710.25044300.08162580.8184855-18.369158136.4133243
130.69999800.10887040.65594221.26296900.12042530.23187010.06559020.8840757-34.405776426.2969008
140.79999180.08361970.50782621.16858100.09323250.21454130.05077950.9348552-49.217375316.8580988
150.89998570.05275280.43098411.08662950.07912490.19949570.04309580.9779510-56.90158838.6629503
161.00.00131630.22045841.00.04047420.18359130.02204901.0-77.95415720.0
" ], "text/plain": [ " group cumulative_data_fraction lower_threshold lift cumulative_lift response_rate cumulative_response_rate capture_rate cumulative_capture_rate gain cumulative_gain\n", "-- ------- -------------------------- ----------------- -------- ----------------- --------------- -------------------------- -------------- ------------------------- -------- -----------------\n", " 1 0.0100178 0.539793 2.84572 2.84572 0.522449 0.522449 0.0285078 0.0285078 184.572 184.572\n", " 2 0.0200151 0.490796 2.57307 2.70953 0.472393 0.497446 0.0257238 0.0542316 157.307 170.953\n", " 3 0.0300125 0.461325 2.35029 2.58987 0.431493 0.475477 0.0234967 0.0777283 135.029 158.987\n", " 4 0.0400098 0.43947 2.2389 2.50217 0.411043 0.459377 0.0223831 0.100111 123.89 150.217\n", " 5 0.0500072 0.419959 2.20549 2.44286 0.404908 0.448487 0.022049 0.12216 120.549 144.286\n", " 6 0.100014 0.355644 2.0732 2.25803 0.380621 0.414554 0.103675 0.225835 107.32 125.803\n", " 7 0.150001 0.311542 1.73543 2.08388 0.318609 0.382581 0.0867483 0.312584 73.5428 108.388\n", " 8 0.200008 0.277602 1.61001 1.9654 0.295585 0.36083 0.0805122 0.393096 61.0015 96.5398\n", " 9 0.300002 0.226271 1.32748 1.75277 0.243713 0.321794 0.132739 0.525835 32.7476 75.2772\n", " 10 0.399996 0.188803 1.13036 1.59718 0.207524 0.293228 0.113029 0.638864 13.0359 59.7177\n", " 11 0.50001 0.158747 0.979815 1.47369 0.179886 0.270556 0.0979955 0.73686 -2.01848 47.3689\n", " 12 0.600004 0.132938 0.816308 1.36413 0.149867 0.250443 0.0816258 0.818486 -18.3692 36.4133\n", " 13 0.699998 0.10887 0.655942 1.26297 0.120425 0.23187 0.0655902 0.884076 -34.4058 26.2969\n", " 14 0.799992 0.0836197 0.507826 1.16858 0.0932325 0.214541 0.0507795 0.934855 -49.2174 16.8581\n", " 15 0.899986 0.0527528 0.430984 1.08663 0.0791249 0.199496 0.0430958 0.977951 -56.9016 8.66295\n", " 16 1 0.00131627 0.220458 1 0.0404742 0.183591 0.022049 1 -77.9542 0" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "Scoring History: \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", "\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", "
timestampdurationnumber_of_treestraining_rmsetraining_loglosstraining_auctraining_lifttraining_classification_errorvalidation_rmsevalidation_loglossvalidation_aucvalidation_liftvalidation_classification_error
2017-10-14 12:37:05 4 min 42.997 sec0.00.50.69314720.51.00.81608040.50.69314720.51.00.8164087
2017-10-14 12:37:05 4 min 43.132 sec1.00.44261860.58343500.64798981.81738370.30566190.44269770.58357930.64537621.78983590.3089363
2017-10-14 12:37:05 4 min 43.194 sec2.00.41176680.52603040.66823732.00293780.35219610.41184580.52617710.66512871.97157950.3528714
2017-10-14 12:37:06 4 min 43.255 sec3.00.39612880.49596430.67456952.17793850.34434420.39620650.49612230.67079462.09326850.3441008
2017-10-14 12:37:06 4 min 43.323 sec4.00.38661990.47638230.68344992.52747770.31755360.38667350.47651190.68051122.52640310.3426287
------------------------------------------
2017-10-14 12:37:09 4 min 46.356 sec18.00.37213450.43876260.70090242.92025470.33134380.37281290.44078390.69393712.86393600.3353914
2017-10-14 12:37:09 4 min 46.645 sec19.00.37200200.43844440.70153402.89981830.34048660.37272970.44056740.69443532.87320250.3441825
2017-10-14 12:37:09 4 min 46.933 sec20.00.37184890.43809000.70215262.81176500.32396270.37262880.44032230.69485602.91678870.3314252
2017-10-14 12:37:13 4 min 50.940 sec196.00.36673020.42643560.72628433.52301530.30880570.37128170.43697530.70323452.76790530.3397870
2017-10-14 12:37:15 4 min 52.513 sec250.00.36590090.42458800.72995803.54776710.29963250.37132220.43700080.70324142.84571790.3369656
" ], "text/plain": [ " timestamp duration number_of_trees training_rmse training_logloss training_auc training_lift training_classification_error validation_rmse validation_logloss validation_auc validation_lift validation_classification_error\n", "--- ------------------- ---------------- ----------------- ------------------- ------------------- ------------------ ------------------ ------------------------------- ------------------- -------------------- ------------------ ------------------ ---------------------------------\n", " 2017-10-14 12:37:05 4 min 42.997 sec 0.0 0.5 0.6931471805589422 0.5 1.0 0.816080432537513 0.5 0.6931471805591649 0.5 1.0 0.816408725696645\n", " 2017-10-14 12:37:05 4 min 43.132 sec 1.0 0.44261860970260863 0.5834349706124139 0.6479898219928322 1.8173836984392455 0.3056618674442622 0.4426976691199359 0.5835793420627537 0.6453762000125266 1.7898359480853072 0.3089362746100219\n", " 2017-10-14 12:37:05 4 min 43.194 sec 2.0 0.411766825778027 0.5260304436464857 0.6682372994206375 2.0029377609070997 0.3521960998724257 0.4118457510016101 0.5261770622602281 0.6651287245222607 1.97157954839959 0.3528714247745998\n", " 2017-10-14 12:37:06 4 min 43.255 sec 3.0 0.3961287738937405 0.4959642917391692 0.6745694791270342 2.1779384856534234 0.34434420752080674 0.39620653713774856 0.49612226036104085 0.6707945650278248 2.093268519804798 0.34410075031177806\n", " 2017-10-14 12:37:06 4 min 43.323 sec 4.0 0.3866199070203514 0.4763823446068694 0.6834499003363573 2.52747771280271 0.3175536115667335 0.38667347762105037 0.4765118683703171 0.6805111786072406 2.526403129076618 0.34262874900333246\n", "--- --- --- --- --- --- --- --- --- --- --- --- --- ---\n", " 2017-10-14 12:37:09 4 min 46.356 sec 18.0 0.37213447528939014 0.4387625831157209 0.7009024298270556 2.9202547371155116 0.33134378227325195 0.3728128580036016 0.44078392331421695 0.6939370703723837 2.863935951143635 0.3353914092368082\n", " 2017-10-14 12:37:09 4 min 46.645 sec 19.0 0.37200199050864735 0.4384444216971638 0.701533974570949 2.8998183319570603 0.34048660470202297 0.37272971066215693 0.4405674396540276 0.694435276805799 2.87320250023814 0.34418252816224726\n", " 2017-10-14 12:37:09 4 min 46.933 sec 20.0 0.3718489401358285 0.4380899869194073 0.7021526089751158 2.811764980968051 0.32396269971447667 0.37262879513569297 0.44032232091599416 0.6948559731202325 2.91678871686913 0.331425183489052\n", " 2017-10-14 12:37:13 4 min 50.940 sec 196.0 0.36673015609468546 0.42643558407189514 0.7262843283105693 3.5230152534493495 0.30880566186744424 0.37128171215614425 0.4369752968195832 0.7032344739799967 2.767905322485342 0.33978696869952774\n", " 2017-10-14 12:37:15 4 min 52.513 sec 250.0 0.3659008817349886 0.4245879659541619 0.7299580222631977 3.5477671170567224 0.2996324646133285 0.3713221885151888 0.43700082401861967 0.7032413981615196 2.845717921912641 0.3369656328583403" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "See the whole table with table.as_data_frame()\n", "Variable Importances: \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", "
variablerelative_importancescaled_importancepercentage
dti112.01.00.1505376
int_rate108.00.96428570.1451613
annual_inc105.00.93750.1411290
revol_util96.00.85714290.1290323
loan_amnt67.00.59821430.0900538
total_acc59.00.52678570.0793011
longest_credit_length53.00.47321430.0712366
addr_state49.00.43750.0658602
purpose41.00.36607140.0551075
emp_length25.00.22321430.0336022
home_ownership8.00.07142860.0107527
delinq_2yrs8.00.07142860.0107527
term8.00.07142860.0107527
verification_status5.00.04464290.0067204
" ], "text/plain": [ "variable relative_importance scaled_importance percentage\n", "--------------------- --------------------- ------------------- ------------\n", "dti 112 1 0.150538\n", "int_rate 108 0.964286 0.145161\n", "annual_inc 105 0.9375 0.141129\n", "revol_util 96 0.857143 0.129032\n", "loan_amnt 67 0.598214 0.0900538\n", "total_acc 59 0.526786 0.0793011\n", "longest_credit_length 53 0.473214 0.0712366\n", "addr_state 49 0.4375 0.0658602\n", "purpose 41 0.366071 0.0551075\n", "emp_length 25 0.223214 0.0336022\n", "home_ownership 8 0.0714286 0.0107527\n", "delinq_2yrs 8 0.0714286 0.0107527\n", "term 8 0.0714286 0.0107527\n", "verification_status 5 0.0446429 0.00672043" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/plain": [] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# show grid search results\n", "gsearch.show()\n", "\n", "# select best model\n", "gbm_model = gsearch.get_grid()[0]\n", "\n", "# print model information\n", "gbm_model" ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.7299580222631977\n", "0.7032413981615196\n", "0.7022823257893827\n" ] } ], "source": [ "# measure gbm AUC\n", "print(gbm_model.auc(train=True))\n", "print(gbm_model.auc(valid=True))\n", "print(gbm_model.model_performance(test_data=test).auc())" ] }, { "cell_type": "code", "execution_count": 16, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABPcAAANLCAYAAAAttXhCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRqIsAAAAAElFTkSuQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# analyze variable importance\n", "gbm_model.varimp_plot()" ] }, { "cell_type": "code", "execution_count": 17, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "PartialDependencePlot progress: |█████████████████████████████████████████| 100%\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsQAAAPxCAYAAAD9lbspAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzs3XmcFNW5//HvAyqCCzIQlxjRMe7RuGtwTW4UlcTBLeKN\n+algEo1ilCiYm7iQROMFjSZKbtxwvw4QNWgWozFG42g0RtyiIu5EvSozCC6DKHB+f5ye0DPMDF2n\nuvtUdX/er9e8Gnqqq576Tg88U3PqHHPOCQAAAKhXfWIXAAAAAMREQwwAAIC6RkMMAACAukZDDAAA\ngLpGQwwAAIC6RkMMAACAukZDDAAAgLpGQwwAAIC6RkMMAACAukZDDKBqzOw+M/tL4GuXmdk55a4p\n78zsVTO7JnYd1VJ4D90bu45YzOw6M3ulxG0nmtmyStcE1AIaYqAOmNmxhYay42ORmT1vZpeZ2bpl\nPtbWZnaumQ3t5tNOUsX+gzazjbuc58dmNs/MHjSz881so0odOyIXu4ByKDT2xV+7t83sr2Z2SJdN\ng87XzPoX3pf7lLj9vl3qKf64OaSGMnEqymAl51XR7zeglqwSuwAAVeMknS3pVUmrS9pL0nckHWRm\n2zrnPirTcbaRdK6kv0ia2+Vz+5fpGCtzs6Q/yP/QP0jSrpJOlXSqmR3vnJtepTpQOifpcUkXSTJJ\nn5Z0gqTbzOxE59yVKfc/QP596ST9NcHrfi7pH12eezVlLeXU23n9RNIFVa8IyCEaYqC+/NE5N6vw\n52vMbL6kcZJGSkrVJJpZP0kfyzcz3V7Fc84tSXOMBGY55zpdxTOzH0r6k6TrzOxZ59zTVaoFpXvD\nOdfc8Rczu1HSi/Lv0bQNsQW+rsU5d1vKY3fLzPo75xal3U1Pn3DOLZP/ngSwEgyZAOrbvfL/oTZK\nkpkNMrOLzOwpM3vfzBaa2R/M7PPFLyr6dfIoMzvPzF6X9KGk70qaUdjsvsI2Szt+ndt1/KeZrWpm\nPzazf5jZAjP7oPBr8i+W+0Sdc/+SdJykfpImdDmfgWb2czOba2YfmdkLZjbBzKxom47hGN8zs9MK\nv+JvL5zT57oez8y2NLNbzKytMETlUTM7uMs2HUNZ9jCzi83snUIGt5nZ4G72eZaZ/cvMPjSzP5vZ\nNt2da8D5fMvMXixs+3cz26WH85lRqLHdzGab2Xldtvm0mV1jZm8V9vVPMxvd09dkZZxzb0t6ToX3\nZ0/M7FNmNrVw3EVm9oSZHVN8rpLekf9BbWLR0IeyjEk3s5MK5/qRmb1hZlPMbGCXbe4rfF/tVHiP\nfyjp/JXs95DCfhcVXntIl8/3el7GGGKgZFwhBurbZoXHtsLjppKaJP1a0iuS1pP/tfV9ZraNc+6t\nLq8/W9JiSRfKN5p3SbpU0imSzpM0u7Ddc4XHrleO15Y0RlKz/BXAtSQdL+mPZrabc+6ptCdYzDn3\nsJm9pKKhG2bWX/5XzRtIulzSvyTtIf+r5vUlfa/Lbo6VtKakKfJDT06V9Gcz2845N6+wz89JapH0\nemE/H0o6UtJMMzvMOXd7l31eJmm+pImSNpG/IjpF0n8W1fkTST+U9DtJd0raSdLdklYt3lHA+Rxd\nOJ/L5b8+Z0q61cw2dc4tLezz85IekP9aXyHpNUmflfRVSWcVtllX0iOSlsq/B1olHSRpqpmt5Zy7\nVAmZ2SqSNtLy92d326wu6X759+5l8sMZvib/m4CBzrnLJM2TdGLhHG8rfEhSKe+vtbr54WS+c84V\njj9R0jnyX4v/kbSlpJMk7WJme3ZkKJ/tEPmhPNMk3SDp7V7Oa7ikWyT9U9L3JQ2WdK38e6rDys6r\n03hjAL1wzvHBBx81/iHfxC2V9CX5/1g3lDRK/j/UDyRtUNhu1W5eO1TSIkk/LHpuX/mbdV6QtFqX\n7Q8vHGufbvb1F0n3Fv3dJK3SZZu1Jf2fpKu6PL9M0jkrOc+NC9t9r5dtflOob83C38+S9J6kTbts\n91P5Xzdv2GXfH0hav2i7XQvPX1T03D3y42G7nluLpNldvi7L5IeyFG/3s8Kx1yr8fYikjyTd3mW7\n8wqvv6bouaTn846ktYu2O7iQz4ii5+6XtKDjtT3kerV8s7ZOl+dvlm/2+63ka/eKfKM/uPDxefkf\nlJZKuqSX99CphW2OKnqur6QHJS2UtEbhucGlvIe6eY8vLTwuK/r70C5flz90ee1Jhe2O7VL3Uknf\nLPH4jxfyXLPouS8Xani56Lkez0t+bPHSUo7HBx/1/sGQCaB+mKQ/yzfB/5JvVN6TdIhz7v8kyTn3\nyb83NutjZg2S2iU9L39FsqvrnHPBYxSdt6RwPDOzQZJWk7+JqbvjlcMHhce1Co9HyF/9XGhmgzs+\n5LNaRVLXu/d/44qulDvnHpW/MjqicB6D5H/w+LWkgV32ebekzc1sg6L9Oa04PvYB+aZu48Lf95O/\nEnxZl+1+3s35JT2fac6597oc2+SvuMrMhkjaW9JU59wb3Ryvw2GSfiupbzfnPFClfT0PkH9/zpP0\nhPwPVzfIXyHtyUGS3nLOTet4wvmrspfKX/net4Tj9uZH8vl3fOwvqePr3/F16fp1uErS+5K+0uX5\nxZKuW9kBzWx9SdvLf391vF/lnPuzpGcTnwGAlWLIBFA/nPyVqxckLZH0tnPu+eINCmNMT5OffaJR\nvinreG1rN/t8NW1RZnas/K/xt1LnX/+/nHbfPViz8Ph+4XFzSdvJN2FdOUldp6V7sZvt5sj/ml7y\nw1BM/g7/87rZtmOf/1f03L+6bPNu4XFQ4bGjMe50bOdcq5m9q86Snk+nYzvnFhSGGncce9PC4zPd\n7E+SH8MraR1J35YfYlPKcbvzsPywEMn/IPZcl2a9OxvLv6e7ek7+67BxN59L4p/OuZ7mPe7Y95zi\nJ51zn5jZy90c+w1X2o2l3X69C56XtGMJ+wCQAA0xUF8edctnmejODyX9WP7X32fJ/6p7maRfqPub\ncFPdIW9m35AfF3mbpMnyv75fKukHWt6Ildu2kt4puvLWR372iUnq/o79Od0815uOnC6SH1Pdna6N\nztJutrEe6inl+EnOp7tjdxw/yTEl6SZJ1/ewTSnjdVudc0ELt+RE2hklAFQIDTGAYofLj8/8dvGT\nZraOur/i2J0kN/EcLukl59wRXY734wT7KJmZDZO/GeyGoqdfkh+nWWojtnk3z22h5VfLO65sf9LL\nlcVSFOf4WtGxO47TMZxhkDpLej4r03E+2/ayzTz5K+59U55ziNfkr4h3tXXR56XK3FzWse8t1fnr\nsqr8b1j+lHK/3b3Xtuzyd26aA8qAMcQAii1VlyuDZvY1+ZvwSvVhYR/rlHi8Tsxsd0nDEhyvJIUp\nqq6TH8d5UdGnZkgaVrirv+trBppZ3y5PH2Jmny7aZjdJu8vPHiDnZ5q4T9IJhbGgXfc5JKD8e+SH\nuZzS5flx3Wyb9Hx65ZxrlZ+1Yoz1sNKf8/Pd3irpcOt+CrqQcy7VHyStb2ajio7XVz6r9+VvCJT8\nEAyptPdlqe6R9In8dIPFvil/c+jvQnZaGKP+hKRjzaxjrLvMbH/5hW+KVeK8gLrDFWKgfpTyK/Df\nSTrbzK6R9JD8lbej5a86luoJ+Ub3zMKV5cWS/lxorLo73mFmNlPS7+WHSZwgP151zW62L9XOZna0\n/A/968jPBHG4/PCPbzjn/lm07YXyU839zsyuk/SYpDXkZzk4TH4atPlF278oqcXMfqXl067NK+yn\nw8nyN6c9bWZXyV9lXU++0d9QnceA9vR1+ffzhbHCF0n6vpn9Tr4J3FHSgVrxyn3S8ynFdwvnM8vM\nrpSfEaJRfiaKjnP5vqQvSnqkcM7PSmqQtLOk/5CfkaESrpR/z1xnfv7kV+XHcw+TdKpz7kNJcs59\nZGbPShplZi/IZ/BP51yPY6NXpvB1uUDSOWb2R0l3yI+F/46kv0v63/DT0n/Jf388WPh+HCxprPw0\nbP/+3qjEeQH1iIYYqB+l/Gr1p/JLwX5dft7cx+RnT/jvbl7f02p0b5vZCfL/oV8tf2Pel7R8WVlX\ntO11ZtYx1/Fw+Sbq6MKxu86GUOqcqk7SUYWPJfIzabwg6WJJVzjnXu+0sXOLzC8c8gP5Rur/FV4z\nR35+2YVd9n+DfGN9mvyNYo9IOsX5RSQ69vlcoTk7V35qtcHy46Mflx+j3bXens6juM4fmtki+Xln\nvyh/A9pw+R8kijNNcj49ZdrpeefcU2b2BfkbBU+U/0HgNRWtbuice6dwtfwcSYfKN4Vt8j/cdFoI\npZfzLfXX/8W1fWRm+8q/R4+RvzL7vKTjnHM3dnnd8fIzdVwsP5vJj9TLzYKl1OOc+5GZvSPfrF4s\n35BeLj9NYdffgJQ8vME5d1fhtzPnyX9fviS/sMwhWvF7o7fzYkgFUAJzju8VAFiZwpCLVySd4Zy7\nOHY9AIDyYQwxAAAA6hoNMQAAAOoaDTEAlC7JOFcAQE4whhgAAAB1jSvEAAAAqGtMu1bEzAZLOkB+\nHsuP4lYDAACAXqwuP7f6Xc65tjQ7oiHu7AClm0gdAAAA1XW0pJvT7ICGuLNXJemmm27S1ltvHbmU\n/Bg3bpwuueSS2GXkDrmFIbcw5BaG3JIjszDkltxzzz2nb3zjG1Khf0uDhrizjyRp66231k477RS7\nltwYOHAgeQUgtzDkFobcwpBbcmQWhtxSST3MlZvqAAAAUNdoiJHaxx9/HLuEXCK3MOQWhtzCkFty\nZBaG3OKiIUZqTzzxROwSconcwpBbGHILQ27JkVkYcouLhhipffe7341dQi6RWxhyC0NuYcgtOTIL\nQ25xsVJdETPbSdJjjz32GAPbAQAAMmzWrFnaeeedJWln59ysNPviCjEAAADqGg0xAAAA6hoNMVKb\nOnVq7BJyidzCkFsYcgtDbsmRWRhyi4uGGKnNmpVq2E7dIrcw5BaG3MKQW3JkFobc4uKmuiLcVAcA\nAJAP3FQHAAAAlAkNMQAAAOoaDTEAAADqGg0xUmtqaopdQi6RWxhyC0NuYcgtOTILQ25x0RAjtbFj\nx8YuIZfILQy5hSG3MOSWHJmFIbe4mGWiCLNMAAAA5AOzTAAAAABlQkMMAACAukZDjNRmzpwZu4Rc\nIrcw5BaG3MKQW3JkFobc4qIhRmrNzc2xS8glcgtDbmHILQy5JUdmYcgtLm6qK8JNdQAAAPnATXUA\nAABAmdAQAwAAoK7REAMAAKCu0RAjtdGjR8cuIZfILQy5hSG3MOSWHJmFIbe4aIiR2vDhw2OXkEvk\nFobcwpBbGHJLjszCkFtczDJRhFkmAAAA8oFZJgAAAIAyoSEGAABAXaMhRmotLS2xS8glcgtDbmHI\nLQy5JUdmYcgtLhpipDZ58uTYJeQSuYUhtzDkFobckiOzMOQWFzfVFeGmujDt7e0aMGBA7DJyh9zC\nkFsYcgtDbsmRWRhyS46b6pApfAOHIbcw5BaG3MKQW3JkFobc4qIhBgAAQF2jIQYAAEBdoyFGauPH\nj49dQi6RWxhyC0NuYcgtOTILQ25x0RAjtaFDh8YuIZfILQy5hSG3MOSWHJmFIbe4mGWiCLNMAAAA\n5AOzTAAAAABlQkMMAACAukZDjNRmz54du4RcIrcw5BaG3MKQW3JkFobc4qIhRmoTJkyIXUIukVsY\ncgtDbmHILTkyC0NucXFTXRFuqgszd+5c7o4NQG5hyC0MuYUht+TILAy5JcdNdcgUvoHDkFsYcgtD\nbmHILTkyC0NucdEQAwAAoK7REAMAAKCu0RAjtUmTJsUuIZfILQy5hSG3MOSWHJmFIbe4aIiRWnt7\ne+wSconcwpBbGHILQ27JkVkYcouLWSaKMMsEAABAPjDLBAAAAFAmNMQAAACoazTESK21tTV2CblE\nbmHILQy5hSG35MgsDLnFRUOM1MaMGRO7hFwitzDkFobcwpBbcmQWhtzioiFGahMnToxdQi6RWxhy\nC0NuYcgtOTILQ25xMctEEWaZAAAAyAdmmQAAAADKhIYYAAAAdY2GGKlNnTo1dgm5RG5hyC0MuYUh\nt+TILAy5xUVDjNRmzUo1bKdukVsYcgtDbmHILTkyC0NucXFTXRFuqgMAAMgHbqoDAAAAyoSGGAAA\nAHWNhhgAAAB1jYYYqTU1NcUuIZfILQy5hSG3MOSWHJmFIbe4aIiR2tixY2OXkEvkFobcwpBbGHJL\njszCkFtczDJRhFkmAAAA8oFZJgAAAIAyoSEGAABAXaMhRmozZ86MXUIukVsYcgtDbmHILTkyC0Nu\ncdEQI7Xm5ubYJeQSuYUhtzDkFobckiOzMOQWFzfVFeGmOgAAgHzgpjoAAACgTGiIAQAAUNdoiAEA\nAFDXaIiR2ujRo2OXkEvkFobcwpBbGHJLjszCkFtcNMRIbfjw4bFLyCVyC0NuYcgtDLklR2ZhyC0u\nZpkowiwTAAAA+cAsEwAAAECZ0BADAACgrtEQI7WWlpbYJeQSuYUhtzDkFobckiOzMOQWFw0xUps8\neXLsEnKJ3MKQWxhyC0NuyZFZGHKLi5vqinBTXZj29nYNGDAgdhm5Q25hyC0MuYUht+TILAy5JcdN\ndcgUvoHDkFsYcgtDbmHILTkyC0NucdEQAwAAoK7REAMAAKCu0RAjtfHjx8cuIZfILQy5hSG3MOSW\nHJmFIbe4aIiR2tChQ2OXkEvkFobcwpBbGHJLjszCkFtczDJRhFkmAAAA8oFZJgAAABDNyy9L114r\n1cp1VRpiAAAAJPLww9KYMVJ7e+xKyoOGGKnNnj07dgm5RG5hyC0MuYUht+TILEzechs82D/Onx+3\njnKhIUZqEyZMiF1CLpFbGHILQ25hyC05MguTt9waGvxjW1vcOsqFhhipTZkyJXYJuURuYcgtDLmF\nIbfkyCxM3nLraIi5QgwUMFVMGHILQ25hyC0MuSVHZmHylNuSJdLf/+7/TENcZmZ2spm9YmaLzOxh\nM9u1xNftaWafmNkK022Y2dfM7LnCPp80s4PKXzkAAED9mDdP+vrX/Z9piMvIzEZJ+pmkcyXtKOlJ\nSXeZ2ZCVvG6gpOsl3dPN5/aQdLOkqyTtIOl2STPNbJvyVg8AAFA/iscN0xCX1zhJVzjnbnDOzZZ0\noqR2SWNW8rrLJf2vpIe7+dx3Jd3pnLvYOfe8c+4cSbMkjS1j3ZA0adKk2CXkErmFIbcw5BaG3JIj\nszB5yq211T/26yctXhy3lnJZJXYBZraqpJ0l/bTjOeecM7N7JA3r5XWjJTVKOlrS2d1sMkz+qnOx\nuySNTFszOmuvlUkIq4zcwpBbGHILQ27JkVmYPOXWcYX4zTeX31yXd9GXbjazDSS9IWmYc+6Roucn\nSdrHObdCU2xmm0v6q6S9nHMvmdm5kkY653Yq2maxpGOcc9OLnvuOpHOccxv0UAtLNwMAAPTiiiuk\nk06SPvlE6hNxrEFdL91sZn3kh0mc65x7qePpiCUBAADUjbY2f2U4ZjNcblk4lVZJSyWt1+X59SS9\n1c32a0naRdKUwuwSn8gPmdjBzD42sy8WtnsrwT47GTFihJqamjp9DBs2TDNnzuy03d13362mpqYV\nXn/yySdr6tSpnZ6bNWuWmpqa1Nox8Kbg3HPPXWHc0Ny5c9XU1LTCqjWXXXaZxo8f3+m59vZ2NTU1\nqaWlpdPzzc3NGj169Aq1jRo1ivPgPDgPzoPz4Dw4D84j+DxaW6UhQ6p7Hs3NzWpsbNQOO+zw795s\n3LhxK+wvVPQhE5JkZg9LesQ5d2rh7yZprqRLnXMXdtnWJG3dZRcnS/qSpMMlveqcW2Rm0yT1d86N\nLHrtg5KedM6d1EMdDJkI0NraqiFDep0QBN0gtzDkFobcwpBbcmQWJk+5HXus9PLL0gMPxK2jFodM\nXCzpW2Z2jJltJT97xABJ10mSmV1gZtdL/oY759yzxR+S3pH0kXPuOefcosI+fyHpQDP7npltaWYT\n5W/ey9dSMDkwZszKJgNBd8gtDLmFIbcw5JYcmYXJU27XXCPdeWfsKsor+iwTkuScm1GYc/jH8sMa\nnpB0gHNuXmGT9SVtlHCffzOzr0s6v/DxgvyNd8+Wr3JI0sSJE2OXkEvkFobcwpBbGHJLjszC5Cm3\nvn2lNdeMXUV5ZWLIRFYwZAIAACAfanHIBAAAAHLkxRelL31JmjMndiXp0RADAAAgyH33+QU68o6G\nGKl1ncoFpSG3MOQWhtzCkFtyZBYmj7l1rFI3f37cOsqBhhipzZqVathO3SK3MOQWhtzCkFtyZBYm\nj7kNHCiZ1UZDzE11RbipDgAAoHSDB0tnnilNmFD9Y3NTHQAAAKruoYeko4+W2tv93xsa/FLOeUdD\nDAAAgJI8+6zU3Cz16+f/3tBQG0MmaIgBAABQktZWadAgvziHREMM/FtTU1PsEnKJ3MKQWxhyC0Nu\nyZFZmLzk1toqDRmy/O+HHSYNHx6vnnLJxNLNyLexY8fGLiGXyC0MuYUhtzDklhyZhclLbm1t/ka6\nDt/6VrxayolZJoowywQAAEDPDj7YT7V2xx2xK2GWCQAAAETQ9QpxraAhBgAAQEm6jiGuFTTESG3m\nzJmxS8glcgtDbmHILQy5JUdmYfKS2377SbvuGruK8qMhRmrNzc2xS8glcgtDbmHILQy5JUdmYfKS\n2//8j3TkkbGrKD9uqivCTXUAAAD5wE11AAAAiM456Y03pPfei11JOjTEAAAACLJsmfSZz0gzZsSu\nJB0aYgAAAATp21daZ538L99MQ4zURo8eHbuEXCK3MOQWhtzCkFtyZBYmz7k1NNAQAxpeC4uYR0Bu\nYcgtDLmFIbfkyCxMnnOrhYaYWSaKMMsEAABA99rbpdVWk1ZZpfPzBxwgrbWWdMst1a2HWSYAAABQ\nVaecIu2994rPDx6c/yvENMQAAABYqdZW3/x21dAgtbVVv55yoiFGai0tLbFLyCVyC0NuYcgtDLkl\nR2Zh8pBbW5s0ZMiKz9fCGGIaYqQ2efLk2CXkErmFIbcw5BaG3JIjszB5yK2nK8Tf/rb0xz9Wv55y\n4qa6ItxUF6a9vV0DBgyIXUbukFsYcgtDbmHILTkyC5OH3IYMkU4/Xfqv/4pdicdNdciUrH8DZxW5\nhSG3MOQWhtySI7MwWc9t6VLp3Xe7v0JcC2iIAQAA0KsFC/wyzd2NIa4FNMQAAADoVWurf+QKMdCD\n8ePHxy4hl8gtDLmFIbcw5JYcmYXJem4bbijddZf0+c/HrqQyVln5JkDvhg4dGruEXCK3MOQWhtzC\nkFtyZBYm67mtuaaU49WlV4pZJoowywQAAEA+MMsEAAAAMuOKK6Q774xdRTgaYgAAAKRy1VXS7bfH\nriIcDTFSmz17duwSconcwpBbGHILQ27JkVmYvOfW0OCXds4rGmKkNmHChNgl5BK5hSG3MOQWhtyS\nI7Mwec+toUGaPz92FeFoiJHalClTYpeQS+QWhtzCkFsYckuOzMLkPTcaYtS9rE8Vk1XkFobcwpBb\nGHJLjszCZD23666THn2058/TEAMAAKCmTZgg3X13z5+nIQYAAEDNWrbMN7tDhvS8zeDB0gcfSB9/\nXL26yomGGKlNmjQpdgm5RG5hyC0MuYUht85mzZLGju19GzILk+XcFi6Uli71TW9PNtzQL+v84YfV\nq6ucaIiRWnt7e+wSconcwpBbGHILQ26dzZ0r/fKXUmtrz9uQWZgs59YxnVpvDfF++0lPPikNGlSd\nmsqNpZuLsHQzAAA9e/ttaYMNpGuukY47LnY1qJaHH5aGDZOeekrabrvY1SzH0s0AAKDsli71Hz1Z\nbz3pC1/I94pkSK6UK8R5R0MMAECdW7zYL7271VZSc3Pv244c6WcbWLSoOrUhvo4hMjTEQC9aextM\nhh6RWxhyC0NuYWo9t/ffly66SGpslE44Qdp++5X/SnzkSKm9Xbrnnu4/X+uZVUqWc1u2TNpkE6lf\nv9iVVA4NMVIbM2ZM7BJyidzCkFsYcgtTq7nNmyedfbY0dKj0gx9IBx0kPfecdMstvinuzVZbSVts\n0fOwiVrNrNKynNvo0dIrr8SuorJWiV0A8m/ixImxS8glcgtDbmHILUwt5nbffdKIEZKZvyr8ve9J\nn/lMsn2MHCldf70fb9y3b+fP1WJm1UBucdEQIzVm5AhDbmHILQy5hanF3HbZxV8V/s53wseEjhol\nLVnih06stVbnz9ViZtVAbnHREAMAUEfWXFM666x0+9h5Z/8BFPve96Q335SmTYtdSXKMIQYAoIaw\nvABiaW+XXnghdhVhaIiR2tSpU2OXkEvkFobcwpBbmDzltnSpvylu112lBx6IV0eeMsuSWsht8GBp\n/vzYVYShIUZqs2alWhymbpFbGHILQ25h8pDbxx/7leO22Ub62tekgQOl/v3j1ZOHzLKoFnJraFi+\niEfesHRzEZZuBgDkxQcf+MU0fvYz6Y03pEMPlb7/fWm33WJXhnp17bXSmDH+h7RVV6388Vi6GQCA\nOvbHP0obbyxNmCDtv7/07LPSbbfRDKP85s6VdtpJevzxlW/bMWvJu+9WtqZKYJYJAABy5nOfk/7f\n//N39Q8dGrsa1LK33vLNcJ8SLqE2NPjH+fOlddetbF3lxhViAAByZqONpJ//PBvN8L33+uWfUZs6\nVpQuZc7q4oY4b2iIkVpTU1PsEnKJ3MKQWxhyC0NuKzdrll8G+sMP/d/JLExWc+u4Sa6Uhvgzn5Eu\nvDD5yodZQEOM1MaOHRu7hFwitzDkFobcwpDbyo0cKX30kfSnP/m/k1mYrObW2iqtsUZpM5esvbZ0\nxhnZ+M1FUswyUYRZJgAASG6bbaTdd/ezDKC2nHWWdOON0muvxa5kRcwyAQAAMmPkSOl3v/OLg6C2\ntLZKQ4bErqLyaIgBAMiwV1+VZszwc7tm1ciRvnF66KHYlaDc2tpKGz+cdzTESG3mzJmxS8glcgtD\nbmHILUwWcrv+eumb35SWLYtdSc92201af33p9tuzkVkeZTW3Qw7xU/zVOhpipNbc3By7hFwitzDk\nFobcwsTOzTlp2jS/Ct3qq0ctpVd9+kgHH+wb4tiZ5VVWczv66PpoiLmprgg31QEAsuSpp6Ttt5d+\n/3tpxIg1BiJPAAAgAElEQVTY1fTub3+THnxQOu00aRWW/UIVlPOmOt6yAABk1LRpfrGD/faLXcnK\nDRvmP1Df3nhDeuUVaa+9YleSDEMmAADIoI7hEocfLq22WuxqgNI0N0tf/WrsKpKjIQYAIIMefdRf\naTvqqNiVAKVraJAWLpSWLIldSTI0xEht9OjRsUvIJXILQ25hyC1MzNymTZPWW0/ad99oJQThvRam\nVnJraPCPCxbErSMpGmKkNnz48Ngl5BK5hSG3MOQWJmZu220nnXmm1LdvtBKC8F4LUyu5dTTE8+fH\nrSMpZpkowiwTAAAA3jvvSPPm+aW5zUp7zTPPSNtu6xdpqfRNlizdDAAAgIpqbpZ23bX0ZljK7xVi\nGmIAAFBWN98s/fd/x64CaYUs2zxokH+kIUbdaWlpiV1CLpFbGHILQ25hyC25lpYWPfusdOGF+Ztp\nIKYsvtdaW6UhQ5K9ZvXVpQEDaIhRhyZPnhy7hFwitzDkFobcwpBbcpMnT9bIkb4hymCPl1lZfK+F\nXCGWpFdflU46qezlVBQ31RXhprow7e3tGjBgQOwycofcwpBbGHILQ27Jtbe3a/XVB2ijjaQjj5Qu\nuSR2RfmQxffal78sfepTfgrALOKmOmRK1r6B84LcwpBbGHILQ27JDRgwQH36SE1N0u23+xX3sHJZ\nfK+1tSUfMpFXNMQAAGTE+edn92pcUiNH+pX2/vnP2JUgVGtr2JCJPKIhBgAgAxYt8jMzvPBC7ErK\n40tfktZay18lRv44xxViIJHx48fHLiGXyC0MuYUhtzDVzO0Pf5A++EAaNapqh6yIjsz69ZMOPJCG\nuFRZ/B59/nnpG9+IXUV1rBK7AOTf0KFDY5eQS+QWhtzCkFuYauY2bZq0007SFltU7ZAVUZzZKadI\nr7/urzYmWdyhHmXte9RMylhJFcUsE0WYZQIAEMP770vrriv9+MdSBi8UApnELBMAANSQO+6QPvrI\nT1MG5N3TT0vHHSctWBC7ktLREAMAENn06dKwYdLGG8euBEhv/nzp+uuld96JXUnpaIiR2uzZs2OX\nkEvkFobcwpBbmGrk9u670h//KB11VMUPVRW818LUUm4NDf4xT8s30xAjtQkTJsQuIZfILQy5hSG3\nMNXIbfXVpeuuy//sEh14r4Wppdzy2BBzU10RbqoLM3fu3MzdHZsH5BaG3MKQWxhyS47MwtRSbosW\nSQMGSDfeWNlp27ipDplSK9/A1UZuYcgtDLmFIbfkyCxM1nK76irp6qvDXtu/v//NR56uENMQAwCA\nips82a/Eh3yYMUO6++7w1w8e7Fe6ywsaYgAAUHGvvSb96ld+kQ5kX1ubb2pDNTRwhRh1ZtKkSbFL\nyCVyC0NuYcgtDLkl11NmI0dKc+dKTz5Z5YJyImvvtdZWaciQ8NcPH56vVRdZuhmptbe3xy4hl8gt\nDLmFIbcw5JZcT5l98YvS2mtLt98u7bBDdWvKg6y919JeIb7oovLVUg3MMlGEWSYAANXinGQWu4rq\nOuooac4caVaq+QBQae3t0hprVH6WiLSYZQIAgBx75RVps82kp56KXUl1HXKI9PjjfugEsqvjZrg0\nV4jzhoYYAIAqmz5deust6bOfjV1JdR10kLTqqtIdd8SuBL3paIjTjCHOGxpipNba2hq7hFwitzDk\nFobcwlQqt2nTpIMP9r+WrjW9ZTZwoB9LfPvt1asnL7L0Pdqnj7TPPtL668eupHpoiJHamDFjYpeQ\nS+QWhtzCkFuYSuQ2e7afaeGoo8q+60xYWWbnnCOddVaVismRLH2Pfv7z0v33SxttFLuS6mGWCaQ2\nceLE2CXkErmFIbcw5BamErlNn+5nWzjwwLLvOhNWltlee1WnjrzhezQuZpkowiwTAIBKck7aZhtp\nt92k66+PXQ1QWc75jz4VGo9Qk7NMmNnJZvaKmS0ys4fNbNdett3TzFrMrNXM2s3sOTM7rcs2x5rZ\nMjNbWnhcZmbZmuQPAFBXnnrKD5mo1eESQIc33vA3UKZZ/rmaMjFkwsxGSfqZpG9L+rukcZLuMrMt\nnHPdjTL/UNJlkp4q/HkvSVea2QfOuauLtlsoaQtJHTM9cjkcABDNzJl+Sdv99otdCVBZgwZJS5cu\nn7Ei67JyhXicpCucczc452ZLOlFSu6RuR5g7555wzk13zj3nnJvrnLtZ0l2S9l5xUzfPOfdO4WNe\nRc+iTk2dOjV2CblEbmHILQy5hSl3bmedJT30kL9yVqt4r4Wptdz695f69ZPmz49dSWmiN8Rmtqqk\nnSX9ueM55wc23yNpWIn72LGw7X1dPrWmmb1qZnPNbKaZbVOeqlFsFksOBSG3MOQWhtzClDu3vn2l\nLbcs6y4zh/damFrLzcz/NiQvDXH0m+rMbANJb0ga5px7pOj5SZL2cc712BSb2b8kfUpSX0kTnXPn\nF33uC5I2kx9WMVDSeEn7SNrGOfdmD/vjpjoAAIAy2HZb6ctfln7xi8rsvyZvqgu0l/zV5RMljSuM\nRZYkOeceds7d5Jx7yjn3gKTDJM2TdEKcUgEAQLGTT5YuvDB2Fehq3XWlK69Mv588XSHOQkPcKmmp\npPW6PL+epLd6e6Fz7jXn3DPOuamSLpE0sZdtl0h6XP6qca9GjBihpqamTh/Dhg3TzJkzO2139913\nq6mpaYXXn3zyySuMBZo1a5aamppWWInm3HPP1aRJkzo9N3fuXDU1NWn27Nmdnr/ssss0fvz4Ts+1\nt7erqalJLS0tnZ5vbm7W6NGjV6ht1KhRnAfnwXlwHpwH55GJ81iwQLroovyfh1QbXw9J+uEPz9W8\neZO0+urpz6O4IU57Hs3NzWpsbNQOO+zw795s3LhxK+wvVPQhE5JkZg9LesQ5d2rh7yZprqRLnXMl\n/exoZudIOs45t2kPn+8j6RlJv3fOndHDNgyZAACgSmbMkEaNkl5+WWpsjF0NJD9d2mc+I/3+99KI\nEen2dfzx0jPPSA8/XJ7auqrFIRMXS/qWmR1jZltJulzSAEnXSZKZXWBm/57C3MxOMrOvmtlmhY/j\nJZ0u6caibc42s/3NrLFw093/ShoqqXhaNpRBdz+NYuXILQy5hSG3MOSWXJLMDjzQz7hxxx0VLCgn\nsvJe65gmbfDg9Ps66STpggvS76caMjEPsXNuhpkNkfRj+aEST0g6oGiatPUlFa+o3UfSBZI2kbRE\n0kuSxjvnike8DJJ0ZeG170p6TP7Gvc7X+5Ha2LFjY5eQS+QWhtzCkFuYcuQ2b55fqrlfvzIUlANJ\nMlt7bek//kO6/Xbp1FMrWFQOZOV7tGMExZAh6fflL97mQyaGTGQFQyYAAOU2Zoz09NPSo4/GriSb\nfvUr6ZRTpHfe8TdhIa5f/1o68kjp3XelddaJXU3vanHIBAAANWfxYum22/zQAHSvqcmvaPaHP8Su\nBJK/Qty3rzRwYOxKqouGGACACrn7bmnhQn/jGLq34YbSLrswjjgrWlv9+GGz2JVUFw0xUus6XQpK\nQ25hyC0MuYVJm9u0adLnPucXKKgXIZldeql03nkVKCZHsvI92tQkXXZZ7Cqqj4YYqTU3N8cuIZfI\nLQy5hSG3MGlya2/3N4sddVQZC8qBkMyGDZO22KICxeRIVr5Ht9/ejyGuN9xUV4Sb6gAA5dJxc9Kc\nOdLmm8euBqg93FQHAEDGTZvmp52iGUY9mzFDeuqp2FWsHA0xAABltmyZ9Prr9TdcAujqu9/1Q4ey\nLhMLcwAAUEv69JEeecRPJwbUs8GDpfnzY1exclwhRmqjR4+OXUIukVsYcgtDbmHS5ta3b5kKyRHe\na2FqNbeGhuXLQWcZDTFSGz58eOwSconcwpBbGHILQ27JkVmYWs2toSEfV4iZZaIIs0wAABCPc9KI\nEf7jlFNiV1N/2tqk++6T9t9fWnvt8uxz9Gjp+eelhx4qz/6KMcsEAACoOWZ+/PWtt8aupD49/bR0\nxBHS22+Xb5+MIQYAAEho5EippSUf405rTUfmQ4aUb595GTJBQ4zUWlpaYpeQS+QWhtzCkFsYcksu\nbWYHH+xn5/jDH8pUUE5k4b3W2uqv0A8cWL59Dhnih8JkfYQuDTFSmzx5cuwSconcwpBbGHILkzS3\n3/zGNxX1LO17bYMNpN13z8fcteWUhe/RtjY/xKFPGbvDb39bmjfPD4fJMhpipDZt2rTYJeQSuYUh\ntzDkFiZJbq+/Lh12WP1d2eyqHO+1kSOlP/5R+uijMhSUE1n4Hm1t9Q1xPaIhRmoDBgyIXUIukVsY\ncgtDbmGS5DZjhtSvn2/m6lk53msjR0offijde28ZCsqJLHyPtrWVd/xwntAQAwBQBtOmSQcdVN7x\nl/Vq662lzTarv2ETsdXzFWKWbgYAIKWXXpIefVQ6/fTYldQGM2n6dGmTTWJXUl8+/ND/IFKPuEKM\n1MaPHx+7hFwitzDkFobcwpSa2/Tp0oAB0le/WuGCcqBc77WddvJTdtWLLHyP3nefdMUVsauIg4YY\nqQ0dOjR2CblEbmHILQy5hSk1t+nTpaYmaY01KlxQDvBeC5OV3Pr2jV1BHCzdXISlmwEAST37rPS5\nz/kp1w45JHY1QP1g6WYAADLio4/8rAgHHhi7EiCbzjhD+slPYlfRO26qAwAghZ12kmbOjF0FkF1z\n5sSuYOW4QozUZs+eHbuEXCK3MOQWhtzCkFtyZBamlnNraJDmz49dRe9oiJHahAkTYpeQS+QWhtzC\nkFsYckuuUpktW1aR3WZGLb/XaIhRF6ZMmRK7hFwitzDkFobcwpBbcuXObPFiaZttpBtuKOtuM6eW\n32uDB9MQow5kZaqYvCG3MOQWhtzCkFty5c6sXz9p0KDaX7Uu9nvtppukY46pzL4bGvyy0Fme2IyG\nGAAAZNrIkdLdd0uLFsWupHY98YT0yCOV2XdDg7RkifTBB5XZfznQEAMAgEwbOVJqb5fuuSd2JbWr\ntVUaMqQy++5YcTDLwyZoiJHapEmTYpeQS+QWhtzCkFuYnnL7+9+lH/3Iz0GMzirxXttyS/9Ry8Mm\nYn+PtrX5sb6V8NnPSqec4oe/ZBUNMVJrb2+PXUIukVsYcgtDbmF6yu3aa6Vrrsn2f/CxVOq9NnKk\n9NvfSkuXVmT30cX+Hq3kFeJNN5UuvVRaf/3K7L8cWLq5CEs3AwBW5pNPpE9/Who9Wpo8OXY19eOh\nh6Q995QefFDaY4/Y1dSeLbbwP3RceGHsSkrH0s0AAERy773+atpRR8WupL7svru07rrS738fu5La\nVMkrxHnA0s0AACQwfbq0+ebSjjvGrqS+9O0r/fWv/tfvKK8lS6QFC+q7IeYKMVJrbW2NXUIukVsY\ncgtDbmG65rZ4sXTbbdKoUZJZpKIyrpLvtS23lFZdtWK7jyrm9+gnn0jf/ra03XbRSoiOhhipjRkz\nJnYJuURuYcgtDLmF6ZrbXXdJCxcyXKI3vNfCxMytf3/p8sul3XaLVkJ0NMRIbeLEibFLyCVyC0Nu\nYcgtTNfcpk2Ttt1W+tzn4tSTB7Hea2+/HeWwZcP3aFzMMlGEWSYAAL255x6/WtrBB8euBMXee8/P\nofvZz0oHHSQdeKC0zz7+yiey4b33/MdnPlO+fTLLBAAAEey3H81wFq22mr/ZcZ99pFtu8Q3x4MHS\niBF+/ts5cySu/8V17rnS8OGxq+gZDTEAAMi11VeXDjtMuvJKae5c6Z//lH7yE3+z2Pjx0lZbZXvZ\n4HrQ0JDtrwENMVKbOnVq7BJyidzCkFsYcgtDbsnFzszMj/E+/XTpT3/ySxL/5S+VW5a4XGLnVmkd\nDXFWr9TTECO1WbNSDdupW+QWhtzCkFsYcksua5mtuaa0774r3+6HP5RuvdXPIhJD1nIrt4YGf8X+\nww9jV9I9bqorwk11AABJevVVadkyFoGoF++951fCmz3bLwCyxx7Lb87bfnupT41fPnznHWmNNfxH\npdx1l8/ztdekoUPLs09uqgMAoALef1/6wQ/8mNNzz41dDapl7bWl556TXnlF+uUv/dXMn/5U2mkn\n6dOflo47Tpo3L3aVlXPoodLJJ1f2GB1DVtraKnucUDTEAIC6t3SpdPXVfknmn/9cOvNMv1AB6ssm\nm0gnnCDNnLl87PGxx0rPPy8NHBi7usppba38GOuGBv+Y1RvrVoldAAAAMf3lL9K4cdKTT0pHHy1d\ncIG00Uaxq0Jsq60mffGL/qPWtbVJQ4ZU9hhZb4i5QozUmpqaYpeQS+QWhtzCkNuKli2TjjhC+o//\n8As4PPywdNNNnZthckuOzMLEym3pUt+kVrohXntt6e9/l/bfv7LHCcUVYqQ2duzY2CXkErmFIbcw\n5LaiPn38MsyHHy4ddZSfrqsrckuOzMLEym3BAj8VWqWHTPTpI+26a2WPkQazTBRhlgkAAFBPnn/e\n30R6//1+pb88YZYJAACASBYulB59NHYV5dEx60PWFy6pNBpiAEDNevfd2BWgFv3mN9Juu0mLFsWu\nJL3WVv9Y6THEWUdDjNRmzpwZu4RcIrcw5Bam3nJra5NOOcXfIPfKK+H7qbfcyqEeMmts9I+vvlq+\nfcbKbe+9pQceoCGmIUZqzc3NsUvIJXILQ25h6iW3jz/28whvtpl0/fXSOedIG2wQvr96ya2c6iGz\njoY4zQ9bXcXKbdAgaa+9/Ap99Yyb6opwUx0A5JNz0u9+J51+uvTSS9I3vyn9+MfSeuvFrgy1aOlS\nP1XfJZdUfoU39Iyb6gAAKHjuOT+3aVOTtPHG0hNPSFdcQTOMyunbVxo6tLxXiOvBE0/4JdGzeC2W\nhhgAkGsLFkivvy799rfS3XdL220XuyLUg8ZGGuKknn3W/+amvT12JStiYQ4AQK4NGyY98wxjIFFd\njY3SP/4Ru4p8KV6+eY014tbSFVeIkdro0aNjl5BL5BaG3MLUem6VaoZrPbdKqJfMyn2FuB5yK26I\ns4YrxEht+PDhsUvIJXILQ25hKpnbLbf4G9pWWcU3psWPq6wibbihdOqpve9j+nTpgw9WfG3Hn7fe\n2s8cUW2835Krl8xOOsl/lEs95JblhphZJoowywQAdPboo9Ibb0iHHNLzNlddJd1wg7Rkif9YurTz\n4xZb+PG9vdliC+mFF3r+/PnnSz/4Qdg5AOjesmXST38qHXaYtM02lT/e/Pl+RbxbbpEOPzz9/so5\nywRXiAEA3VqwQDrySGnddf0MDn16GGT3rW/5jzTmzPF3nndtpjv+PGBAuv0DWNHChdLZZ0tbbVWd\nhnjgQMksm1eIaYgBACtwTjr+eN8U33tvz81wOZktHyoBoPLa2vxjtVap69vXLwTScdws4aY6pNbS\n0hK7hFwitzDkFiZpblOmSLfdJl177fJVueoR77fkyCxMjNxaW/3j4MHVO+bOO0trrlm945WKhhip\nTZ48OXYJuURuYcgtTJLcHntMOuMM6bvf7X3scD3g/ZYcmYWJkVtHQ1ytK8SSnyt87NjqHa9U3FRX\nhJvqwrS3t2sAA/wSI7cw5Bam1NwWLpR22snfDd7SIvXrV4XiMoz3W3JkFiZGbtdfLx13nLR4sbTa\nalU9dFlwUx0yhX/4wpBbGHILU0puzvmb41pbpT/9iWZY4v0WgszCxMittVVaa618NsPlxpAJAIAk\nf1NbU5N03XXSppvGrgbIvmee8dOHzZsXu5IwbW3VHT+cZTTEAIB/+8Y3pEMPjV0FkA9LlvibT198\nMXYlYQYMkLbbLnYV2UBDjNTGjx8fu4RcIrcw5BaG3MKQW3L1lFnHDCzlWMI5Rm5nnSXdcUfVD5tJ\nNMRIbejQobFLyCVyC0NuYcgtDLklV0+Zrb22H3JQjoa4nnLLImaZKMIsEwAAIIldd5W23166+urY\nldSfcs4ywRViAACAQI2N5blCXC+eflraZBN/Q2KW0BADQB16//383hkPZAkNcTL9+kmvvbZ8UZCs\noCFGarNnz45dQi6RWxhyC1Ocm3PSiSdKe+/t75JHz3i/JVdvmTU2SnPnpv9eqpfcGhr8Y1tb3Dq6\noiFGahMmTIhdQi6RWxhyC1Oc29Sp0s03SxMnSquwPFOveL8lV2+Z7bKLNGaMtGhRuv3US26DBvnH\n+fPj1tEVN9UV4aa6MHPnzuXu2ADkFobcwnTk9tRT0u67S8ccI11xReyqso/3W3JkFqaechs0SPqv\n/5LS/gzATXXIlHr5Bi43cgtDbmGGDh2qDz6QjjxS2mIL6ec/j11RPvB+S47MwlQ7tz//Wdpyyzj3\nEjQ0ZO8KMb8sA4A64Jz0ne9Ir78uPfaY1L9/7IoAxPTmm9KcOdJaa1X/2DTEAIAorr1Wuukm/7Hl\nlrGrARBba6u0xhrS6qtX/9hZbIgZMoHUJk2aFLuEXCK3MOQW5tJLJ+n446Wjj45dSb7wfkuOzMJU\nO7e2NmnIkKoe8t9OPNHfx5AlXCFGau3t7bFLyCVyC0NuYZqa2nXWWbGryB/eb8mRWZhq59ba6ped\njuHQQ+MctzfMMlGEWSYAAEA9OOIIv0DPXXfFriQcs0wAAABkxNKlfrW6Dz+MXUnp2triXSHOIhpi\nAACAFN58U9p0U+n++2NXUrrW1nhjiLOIhhiptWZtQfKcILcw5BaG3MKQW3L1mNmnPy2tuqq/Shyq\n2rmddlo2x/LGQkOM1MaMGRO7hFwitzDktnLd3RpCbmHILbl6zKxvX2njjdM1xNXO7fjjpS99qaqH\nzDQaYqQ2ceLE2CXkErmFIbfe3Xijn1rtk086P09uYcgtuXrNrLExXUNcr7llBQ0xUmNGjjDkFobc\nejZ7tl+NbtVVpVW6TKpJbmHILbl6zSxtQ1xPuS1dKrW0SP/3f7ErWY6GGABqwKJF0pFHShttJP3y\nl5JZ7IqA+pK2Ia4nzkl77y3deWfsSpZjYQ4AqAGnniq98IL0979La64Zuxqg/jQ2SgsW+I911old\nTbatsoo0cGC2lm/mCjFSmzp1auwSconcwpDbim6+WbrqKmnKFGm77brfhtzCkFty9ZpZY6N/DL1K\nXG+5NTTQEKPGzJqVanGYukVuYcitszlzpBNO8DfS9XaTOrmFIbfk6jWzz39eev55adttw15fb7ll\nrSFm6eYiLN0MIG/231+aO1f6xz+ktdaKXQ2APHjhBam9Xdp++3g1HHCAtPba0q9/Hb6Pci7dzBhi\nAMix666T3nuPZhhA6S65RPrb36THH49XQ0OD9M478Y7fFUMmACDHNtxQ2nrr2FUAyJMsLNuctSET\nNMQAAAB1pK1NGjw4bg00xKg5TU1NsUvIJXILQ25hyC0MuSVHZmGqmVsWrhBPmCD9859xayjGGGKk\nNnbs2Ngl5BK5hSG3MOQWhtySI7Mw1cwtC1eIs3bfA7NMFGGWCQAAUMuck/r3ly68UDrllNjVpFPO\nWSYyM2TCzE42s1fMbJGZPWxmu/ay7Z5m1mJmrWbWbmbPmdlp3Wz3tcLnFpnZk2Z2UGXPAgAq46WX\npJkzY1cBoDdLl0onnyzdd1/sSnrW3i4tXhz/CnHWZKIhNrNRkn4m6VxJO0p6UtJdZtbTCJcPJV0m\naW9JW0n6iaTzzOybRfvcQ9LNkq6StIOk2yXNNLNtKnUeAFAJixdLo0ZJZ5zh/wwgm/r2laZPlx58\nMHYlPWtt9Y+xxxBnTSYaYknjJF3hnLvBOTdb0omS2iV1u+6Sc+4J59x059xzzrm5zrmbJd0l3yB3\n+K6kO51zFzvnnnfOnSNpliQGN5XZTC5bBSG3MPWY2/jx0tNPSzNmSP36he2jHnMrB3JLrt4za2wM\nW765WrlttJFvivfZpyqHy43oDbGZrSppZ0l/7njO+YHN90gaVuI+dixse1/R08MK+yh2V6n7ROma\nm5tjl5BL5Bam3nL79a+lyy6TLr5YSnNrQ73lVi7klly9ZxbaEFcrtz59/HCJ1VevyuFyI/pNdWa2\ngaQ3JA1zzj1S9PwkSfs453psYM3sX5I+JamvpInOufOLPrdY0jHOuelFz31H0jnOuQ162B831QHI\njMcfl/bcUxo5Urr5ZsksdkUAVubMM/0Psi+/HLuS2leTN9UF2kv+6vKJksYVxiIDQO69/bZvhLfZ\nRpo6lWYYyIvGRmnuXGnJktiVZN9ZZ/kf9rMgCw1xq6Slktbr8vx6kt7q7YXOudecc88456ZKukTS\nxKJPvxWyT0kaMWKEmpqaOn0MGzZshfE9d999d7cTaZ988smaOnVqp+dmzZqlpqYmtXaMZi8499xz\nNWnSpE7PzZ07V01NTZo9e3an5y+77DKNHz++03Pt7e1qampSS0tLp+ebm5s1evToFWobNWoU58F5\ncB4ZP4/TTx+vI46QPv64Y2aJfJ5HrXw9OA/OI8l5NDb62SZefz3f51GsUudx553S2WeXdh7Nzc1q\nbGzUDjvs8O/ebNy4cSuca6joQyYkycwelvSIc+7Uwt9N0lxJlzrnLixxH+dIOs45t2nh79Mk9XfO\njSza5kFJTzrnTuphHwyZAJAJN90kffaz0jDuegByZc4cacstpXvvlb70pdjVZNv++0uDBvkbhkPU\n4pCJiyV9y8yOMbOtJF0uaYCk6yTJzC4ws+s7Njazk8zsq2a2WeHjeEmnS7qxaJ+/kHSgmX3PzLY0\ns4nywyumVOeU6kd3P7li5cgtTL3k9o1vlLcZrpfcyo3ckqv3zDbeWDr4YL/4RRL1mFtDgzR/fuwq\nvEws3eycm1GYc/jH8sManpB0gHNuXmGT9SVtVPSSPpIukLSJpCWSXpI03jl3ZdE+/2ZmX5d0fuHj\nBUkjnXPPVvh06s7w4cNjl5BL5BaG3MKQWxhyS67eM+vXT7rjjuSvq8fcGhqkF16IXYWXiSETWcGQ\nCQAAUMt+8ANphx2kI4+MXYm/qe6mm6RXXw17fS0OmQAAAECF3XSTX+gnCxoapLa22FV4NMQAAAB1\norU1O8s2NzRIH3zgZ9SJjYYYqXWdWgWlIbcwtZbblVf6OYcrrdZyqxZyS47MwlQjt/Z2adEiv1Jd\nFk+GN/QAACAASURBVGy+uXTIIdLixbEroSFGGUyePDl2CblEbmFqKbf//V/phBOk226r/LFqKbdq\nIrfkyCxMNXLrGJ6QlSvEe+4p/eY30lprxa6Em+o64aa6MO3t7RowYEDsMnKH3MLUSm6PPirtvbd0\n1FHStddWfiW6Wsmt2sgtOTILU43cnnhC2nFH/+/PLrtU9FBVwU11yBT+4QtDbmFqIbc33/S/Jtxx\nR+nyy6uzLHMt5BYDuSVHZl57u7RgQenbVyO3jkXpsjJkIktoiAGgij76SDr0UN8E33abtPrqsSsC\nUAk77CCdd17sKjrL2pCJLMnEwhwAUA+ck779bempp6QHHpA22CB2RQAqZZNNpFdeiV1FZw0N0le+\nIq25ZuxKsocrxEht/PjxsUvIJXILk+fcrr5auvFG6Zprqj9+L8+5xURuyZGZ19iYrCGuRm777y/9\n7nfVGaaVN1whRmpDhw6NXUIukVuYPOc2apTUv7/0n/9Z/WPnObeYyC05MvMaG6UZM0rfntziYpaJ\nIswyAQAAymH6dD+LzLvvSuusE7uabFuyRFol4BIts0wAAABkWGOjf8zaOOKsOfpoacSI2FXQEAMA\nAJQdDXFp1lpr+ewXMdEQI7XZs2fHLiGXyC0MuYUhtzDklhyZeUOGSGusUXpDXK+5NTRI8+fHroKG\nGGUwYcKE2CXkErmFIbcw5BaG3JIjM89Muu8+6dhjS9u+XnOjIUbNmDJlSuwSconcwuQhtwcflH7x\nCz/vcFbkIbcsIrfkyGy5XXYpfRGMSufmnL95LWsaGqT33pM++SRuHTTESI2pYsKQW5is5zZ3rnTY\nYdKtt0pLl8auZrms55ZV5JYcmYWpdG5vvimtuqp0110VPUxiDQ3+Mcky15VAQwwAZfLhh9LIkX6u\n4VtvDZtGCAAqobXVPw4aFLeOrjoa4tjDJvjnGgDKwDlp9GhpzhzpoYekT30qdkUAsFzHTA6DB8et\no6uOemLPNMEVYqQ2adKk2CXkErmFyWpu558v/frX0g03SNtvH7uaFWU1t6wjt+TILEylc+u4Qlzq\nmOZq2Xhj6ZZbpC22iFsHV4iRWnt7e+wSconcwmQxt9/8Rjr7bGniROnww2NX070s5pYH5JYcmYWp\ndG5tbX4Y19prV/Qwia25Zjb+3WTp5iIs3QwgqYULpU02kb78ZWnGDKkPv3cDkEE/+Yn0y19Kb70V\nu5LyKefSzVwhBoAUBg6UfvtbaccdaYYBrOjaa/2cxMcdF7eOtrbsDZfIEv75BoCU9trLr0gFAF3d\nead0442xq/BjiLN2Q12W0BAjtdaOkfpIhNzCkFsYcgtDbsmRWWeNjaUt31zp3MaN88Mm0D0aYqQ2\nZsyY2CXkErmFIbcw5BaG3JIjs84aG/2CPStbJa7Sue28s7TPPhU9RK7RECO1iRMnxi4hl8gtDLmF\nIbcw5JYcmXXW2OhXrXz99d63I7e4aIiRGjNyhCG3MOQWhtzCkFtyZNZZY6N/XNmwiXrO7emnpebm\nuDXQEANACd5/XzriCL8SHQCUauON/SwTpYwjrld33imddFLcGmiIAWAlli2TjjlGuvvulY8DBIBi\n/fpJG25IQ9ybhgZpwQI/tCQWGmKkNnXq1Ngl5BK5hYmR27nnSrffLt18s7TNNlU/fFnwfgtDbsmR\n2Yq+8hVp3XV736aec+uYDu7dd+PVQEOM1GbNSrU4TN0itzDVzm36dOm886QLLpC++tWqHrqseL+F\nIbfkyGxFl18unXJK79vUc24NDf5x/vx4NbB0cxGWbgYg+V9tXn+9dP/90oMPSl/7mnTTTX4cIADk\nzZw50t/+Jh19tLRKBtcofvpp6fOf9zV+4Qulv66cSzdzhRgAunjjDenSS/2yzBddJE2dSjMMIL/+\n8hdpzJjsLi+fhSvEGfw5AQAq5733/M0bQ4f2vM0ee/hlTrP6nwcAJNGxbHNW/03LQkOc0WgAoDwW\nLJB++1vpjDOkXXeVBg2STjut99f06ZPd/zgAIKm2tuU3rmVR//6+Kf7oo3g18E8+UmtqaopdQi6R\nW5hScnv0UWncOGmnnfw/sk1N/ua4Lbf0N7dMnlyFQjOG91sYckuOzMJUMrfWVmnIkIrtviza2qRv\nfjPe8RkygdTGjh0bu4RcIrcwpeT2j39IM2dK++7r7+zed1+/WlQ9jwPm/RaG3JIjszCVzC3rV4iz\ngFkmijDLBJB9b74p9e0rrbdez9ssXeq3AYCscE5atEgaMKD6x959d2m77aSrr67+sSuJWSaAOnPh\nhX6d92XLYldSfXPnSjfe6H+VtvnmfsWnX/2q99fQDAPImlNPlfbcM86xO26qQ89oiIGMc056+GHp\n61/38zTeemvtN8YffCD98pd+VbiNN/bLJj/yiHTAAdKMGfHXvAeApD796XjLN/fp0/tv1UBDjDKY\nOXNm7BJyqTi3+fP9VdDumPkm+KGHpA02kI44Qtp5Zz9zQq2OeDrzTH81Zdttpdtuk+bN8xO3T5ki\nrbrqzJUugYoV8X0ahtySI7PuNTZKCxf2vDxxJXN74QXpe9+r2O5rAg0xUmtubo5dQi515DZ/vrTf\nfv4fq7a2nrcfNkz605/86mlrr+1nTvjCF6S77qq9xvj735deftlfDT700M53R/N+C0NuYcgtOTLr\nXmOjf+zpKjG5xcVNdUW4qQ7V9u670v77S6++Kt17rx8SUQrn/PZnny2tvrr/MwAgu+bNk9ZdV7rl\nFunww2NXUxu4qQ6oAQsWSMOH+2b4z38uvRmW/DCKL39ZevBBP6QgT956S7r99thVAEB1DRkirbFG\nvHHEWfe3v0n77OP/b4yBhhiIoKMZfvll3wxvv33YfsykddYpb22VMmuWdOyx/ia544/30w8BQL0w\n88MmaIi7t3ix9MADfkaMGGiIgSpbuNDPlvDii+ma4VLF+mlbkpYs8TcE7rOPvxHw/vul88/3N3j0\n7x+vLgCIgYa4Zw0N/nH+/DjHpyFGaqNHj45dQq60tPgrw3vtNVo77FDZY7W3S1tvLR12mJ+loVqc\nk372M2mzzfysGJIfN/fii9IZZ0iDBoXvm/dbGHILQ27JkVnPfvQjadKk7j9X77nRECP3hg8fHruE\nXPnKV6SXXpL+8z8rn9tqq0n//d/Sk0/6K9FHHSXNnl3xw8rMX/3ed1/pscekv/7V30SyShkWi+f9\nFobcwpBbcmTWsx139CvGdafec+tYOKS32ZYqiVkmijDLBGrVJ59I110n/eQn0htvSEcfLZ1zjr+C\nWynLlvnJ4AEA8Uyc6IfqXXJJ7EpWrn9/afJk6ZRTStueWSYAJLLqqtK3vuXH7l56qXTPPdJWW/mm\nuFJohgEgvsce88P08qChgSETAKqgXz/p5JP9kI2LLur5V3e9mTvXryR36aXlrw8AUF6trcuHI2Qd\nDTFyraWlJXYJuRQzt/79pdNOk772tdK2d87PeXzkkdKmm0pXXOFv2IuB91sYcgtDbsmRWZhK5dba\n2nm1zyw79lhpjz3iHJuGGKlNnjw5dgmZ88EH0ne+0/tPunnI7eOPpZtuknbbTdprL39z3i9+Ib3+\nul9e+f+zd+fxUVXnH8c/h0UxrLKDGoyguJAoICpudUXc4lYFrBtorQLWpYJVRLAgCv7cEa0WXGoF\nVBQVFXGtYlU0KaJVVBQkUVmCyBbW5Pz+eDISQhJmztw7996Z5/16zWtgMrlz8mWAk5NznicIUcgt\njDQ3N5pb4jQzN37ltmJFdFaIr78e+vQJ5rX1UF0leqjOTWlpKVlZWUEPIzTWroVTToG5c+Gdd6T+\nbnXCntuWLdClC3z9tTQRufpq6N07+L3BYc8trDQ3N5pb4jQzN37ktmWLnCH5xz+kIVK68fJQnQdF\nkFSm03/4tlq7VsqqzZ0Lr79e82QYwp/bokUyEX7+edh//6BHs1XYcwsrzc2N5pY4zax2c+bIYskN\nN2z7uB+5rVwp91FZIQ6SbplQyiPr1sFpp0mL4pkzoWfPoEeUnE6d5OBcmCbDSikVdYWFMGyYrN76\nLdYGOSp7iIOkE2KlPBCbDBcUyGQ4qEMBSimlwi0nB8rKoKjI/9dq1AiuvVZeU9VOJ8QqaUOGDAl6\nCIEqLYXTT4dPPoHXXoMjjojv8zI9N1eamxvNzY3mljjNrHaxyenChds+7kdue+wBd98Nu+3m+aXT\njk6IVdKys7ODHkKg1q2TSfHMmVKJIV6Znpsrzc2N5uZGc0ucZla7Dh2kvX3VCbHmFiytMlGJVplQ\nrqyVf+CUUkqpHdljD6m5O3p00CMJl7Iy+PFH2fMczxlDbd2sVMjoZFgppVS8cnK2XyFW8NNPsoL+\n/vupf22dECullFJKpZBOiKvXvLncB9G+WSfEKmnz588PegiRpLm50dzcaG5uNLfEaWY71r07tG+/\n7WOam2yT2GknnRCriBo6dGjQQ0gJr7fbZ0puXtPc3GhubjS3xGlmO/bnP8Nzz237mOYm2w+bN5d2\n06mmE2KVtPHjxwc9BN9t2ABnngnTpnl3zUzIzQ+amxvNzY3mljjNzI0fuS1YAKtWeX5ZXzVvrivE\nKqLSvVTMxo1wzjkwaxY0berdddM9N79obm40NzeaW+I0Mzd+5HbYYfDQQ55f1lctWuiEWKnQiU2G\n334bXnoJTjgh6BEppZRSO1ZWJhPLFi2CHklidMuEUiGzcSP8/vfw5pvw4otw4olBj0gppZSKz6+/\nytmXli2DHklidMuEiqyxY8cGPQTPbdwI554Lb7whk+Fevbx/jXTMLRU0NzeamxvNLXGamRuvcysp\nkfuorRAPHw5PPpn6162X+pdU6aa0tDToIXju8stlz/D06XDSSf68RjrmlgqamxvNzY3mljjNzI3X\nucW2HURthTgnJ5jX1dbNlWjrZhUzZAgcdxycfHLQI1FKKZWuysuhjk8/q3/pJTjjDFiyBNq08ec1\ngqatm5VytHSp1H789dfan3fnnToZVkop5Z9XXpFGFCtX+nP92ApxrPubqp1umVBpy1ppjfn++/De\ne3L/7bfysddeg969gx2fUkqpzNWqlZxXWbgQdt3V++uXlEip0Pr1vb92OtIVYpW0ktjO/ZCwFi68\nEHbfHTp2hP794ZNPpErE5MlQXByOyXDYcosKzc2N5uZGc0ucZhaf2F7ZhQvl3uvcLr0UPvjA00um\nNZ0Qq6QNGDAg6CFswxho0AAuuABefll+bDRvHjz4IPTtC7vtFvQIRdhyiwrNzY3m5kZzS5xmFp+W\nLaFhw60TYq9za94cDjjA00umNd0yoZI2cuTIlLzOmjXw4YdQWAg33CAT35o8+mhKhpSUVOWWbjQ3\nN5qbG80tcZpZfIyRVeLYhFhzC5ZWmahEq0yEy/Llsu83dps7VzrvtGwJX30VvVIySimlVGX5+bBl\nC7z6atAjCZdx4+CII+RWGy+rTOgKsQqdxYul9u/8+fL7Dh3gqKOkNvBRR8G++9a+OqyUUkpFQU6O\n1LxX27r7bjlwuKMJsZd0QqxCp317OP546VZz1FGwxx5Bj0gppZTyXk4OLFokh8Ejt9BTVgabN8uh\nHY8F0b5ZD9WppE2cONHT69WrB+PHw/nnp/dk2OvcMoXm5kZzc6O5JU4zi9/pp8Mzz0iDjsjlds89\ncOSRvlxaJ8QqkgoLk9q2k7E0NzeamxvNzY3mljjNLH4dO8qkuG7dCOY2bx7873+yvO2xICbEeqiu\nEj1Up5RSSqmoW7IE7rsPBg708Setxx4L774rrfaaNfP00pdcIo20dlRHORStm40xRxljnjLGfGiM\n2a3isQuNMf6snyullFJKqR1auBDuuAN+/dXHFykulvuffvL80pHZMmGMOQd4HVgPdAV2rvhQU+Am\nb4amlFJKKaUStWKF3PtWntRaXyfELVps/RpSxXWF+GbgCmvtH4HNlR7/ANC9Bo42bZLTppmmvDzo\nESillFLpI9YFukULn15gxQrYsEF+7cOEeM895ZbKXb2uE+LOwHvVPL4K8HYjSQa5/34pwfLNN0GP\nJDH5+fnOn7typbSW/Pe/PRxQRCSTWybT3Nxobm40t8RpZm68zG3FCmjcGHbaybNLbquoaOuvfZgQ\n/+EPMGdOakvRuU6IlwCdqnn8SOB79+FktkGD5KTpVVel9ruiZA0ePNj5c8eNk0YcnTt7OKCISCa3\nTKa5udHc3GhuidPM3HiZW0mJz91cYxPiFi3g5599fKHUcZ0QPwrcZ4w5FLBAe2PMH4D/Ax7yanCZ\nZpddYPp06Vrz3HNBjyZ+vXr1cvq8n36SU7DXXQdt23o8qAhwzS3TaW5uNDc3mlviNLPELF8uPyHu\n2tW73Fas8HG7BMj+4fr14aCDfFkhDoLrhPgO4GngLaARsn3iH8DfrbUPeDS2jHTaaXDGGXDttbBm\nTdCj8dett0JWFlx/fdAjUUoppYKxciVcfTV8/rl310zJCvFuu8ktkyfEVtwGNAe6AIcBray1w70c\nXKa67z4pN3LrrUGPxD9ffw0TJ8JNN0HTpkGPRimllApGhw6yV3bhQu+uufvusnjrm6IiKXDcvn1m\nT4hjrLWbrLVfAvOBE4wx+3kzrMzWoQMMHw733uvtd4x+mT59esKfc/PN8vdo4EAfBhQRLrkpzc2V\n5uZGc0ucZpaYnXeW/w9nzvQut/vvh9tv9+xy2ysulll3bEIcpYNPNXCtQ/yMMWZwxa93AT4BngHm\nVdQoVrXYsAG2bKn9OX/5C3TqJIfOwm7y5MkJPf+TT2SP9N/+Bg0a+DSoCEg0NyU0NzeamxvNLXGa\nWeL22gvmzIlQbpVXiDdtkn0fEVfP8fOOBm6r+PVZyMS6GXAxUqN4WvJDS0/l5XDxxbB5Mzz/fM3P\n22knmDFDvgELu6lTpyb0/NWrZa/0hRf6NKCISDQ3JTQ3N5qbG80tcZpZ4nJyYMuWiOQWa8oRWyEG\nWSVu3jzYcSXJdctEUyDWVK83MM1aWwq8AuztxcDS1YgR8MwzcMEFO35up07puYJ6/PHw8stSYk4p\npZTKdDk53u4h9tXy5bIqvMce0K6dPObDPuJ+/eCaazy/bI1cV4iLgJ7GmF+QCXHfisd3BTZ4MbB0\n9OSTMHq09Bc/++ygR6OUUkqpMMjJgSVLYP16KcEaarEaxD5PiNesgdJSzy9bI9cV4nuBfwHFwE/A\nuxWPHw1E4BhY6r3/Plx2GQwYAEOHBj0apZRSSoXF3nvDPvtsbbkcasXFcr/77nIisEULXybELVpI\nPeVUcS27NgHoCQwAjrTWlld86HtkD7GqZMECOOssOPJIeOih1LYiTIX+/fsHPYRI0tzcaG5uNDc3\nmlviNLPEHX44HH54f/bYI+iRxKGoSA46tWolv2/f3pdudc2bSwnaVHHdMoG19lPg0yqPvZL0iNLM\nypVw6qnync5zz/nYVzxA2pXIjebmRnNzo7m50dwSp5m5iUxuRUWyOlynYk3Vp1rEqZ4Qu5Zdq2uM\nudQY87Qx5k1jzNuVb14PMspKSqBJE3jlFW8OYJaVwdq1yV/HS/369Qt6CJGkubnR3Nxobm40t8Rp\nZm68ym38eOjc2ZNLVS9WYSKmXTtfJ8SpKnHsuof4vopbXeAL4LMqN1Vh771hzhypGOGF006DwYO9\nuVaqzJ+fFjW7lVJKqdD7+WfYuNHHF4jVII7xcYV482ZYt87zS1fLdULcFzjPWtvHWnuNtfbayjeX\nCxpjBhljFhpj1htjPjLG9KjluWcZY2YZY5YZY1YZY/5jjOlV5TkXG2PKjTFlFfflxpgUnlesPBbv\nrnX22fDEE3JILwpWrIBDD4U77wx6JEoppVT6KymRbZq+qbpCHNtD7PHKV+yn6qnaNuE6Id4ELPBq\nEMaYPsBdwAigK7LK/LoxpmUNn3I0MAs4GegGvAO8bIw5sMrzVgFtK906eDXmoFx6qUwwBw6U75zC\nYPbs2TV+7PbbpRnJJZekbjxRUVtuqmaamxvNzY3mljjNzI1Xua1YAS1rmj0lq7xcJsRVV4g3b/a8\nJMQBB8Ddd0OjRp5etkauE+K7gKuN8Wzt81rg79baJ62184ErgFKkisV2Klai/89aW2Ct/c5aOwz4\nFjh9+6fa5dbaZRW35R6NNzB16sCECfDll9KrPAzG1dBfevFi2ct0/fXQunWKBxUBNeWmaqe5udHc\n3GhuidPM3HiVm68rxMuWyeS38oTYp1rEu+8O116bugZ4rhPiI4E/AN8ZY142xjxf+ZbIhYwx9YHu\nwFuxx6y1FngTKe0WzzUM0Jit3fNiGhljFhljFhtjphtj9k9kbGHVrZusEI8cubUcYJCmTJlS7eMj\nR8qBwuuuS+14oqKm3FTtNDc3mpsbzS1xmpkbr3LzdYW4cg3imMrtmyPMdUL8K/AC8G+gBNmaUPmW\niJbI4bylVR5fimxziMcQoCHwTKXHvkZWmPORyXsd4D/GmPYJji8ua9fCxx/7ceXqjRoFDRuGY7KZ\nlZW13WNffil7nYcPh8aNAxhUBFSXm9oxzc2N5uZGc0ucZuZmy5YssrPhhReSu46vK8SVu9TFtK2Y\nqkV8QuxUh9haG5qq28aY84HhQL619rceL9baj4CPKj3vQ+Ar4E/IXmXPlJXB+efDhx9KL/JU7Hdp\n1gzuugsuuADefBNOOMH/10zEsGGQnQ2XXx70SJRSSqnwa9xYDpB99537Naz1eYW4qEi601V+gViT\nDh+ac6SS6woxAMaYVsaYIyturRwvUwKUAW2qPN4GWLKD1+8LPAKca619p7bnWmu3AP8FdlgA7ZRT\nTiE/P3+bW8+ePZk+ffo2z5s1axb5+fkMHSp1hp98UibDgwYNYuLEids8t7CwkPz8fEqq9GUcMWIE\nY8eO3eaxxYsXk5+fz/z587d5/IEHHmDIkCG//f788+H++0u555787TbjT548udpuQX369Knx66jK\n9ev46COYPh2uvnox5567468DoLS0lPz8cH0dMfH+eejXoV+Hfh36dejXoV+H69exfn0pderk85//\nuH8dM2fOYv/98zn2WJ++js8+k+0SFUfIfvs6KpVe8+vPY/LkyeTk5HDQQQf9Nje79lqnwmbVs9Ym\nfEO2J0wCtgDlFbfNwEQgy+F6HwH3Vfq9AYqAIbV8Tj9gHXBanK9RB1kh/r9antMNsAUFBTZeDz9s\nLVj7wANxf0rauf7667f5/a+/WnvPPdaWlQU0oIiompuKj+bmRnNzo7klTjNzc/3119vTT7f25JOD\nHkkt+va19phjtn+8d29rzzwz5cMpKCiwgAW6WYf5bOWb6wrx3cDvkKoOzSpuZ1Q8dpfj9f5ojLnI\nGLMv8DCQBTwOYIy53RjzROzJFdskngD+AnxijGlTcWtS6TnDjTEnGmNyjDFdgX8B2cA/HMZXrTfe\ngEGDpFFG1JpleCk7O3ub3zdtCtdcs7Wro6pe1dxUfDQ3N5qbG80tcZqZm+zsbHJyZOtlaFWtQRzj\nU3OOVDLWoZCyMaYE+L219t0qjx8LPGOtTXj7hDFmIDAU2SoxF7jKWvtpxcceAzpYa4+r+P07SC3i\nqp6w1g6oeM7dwFnIwbyVQAEwzFo7r5YxdAMKCgoK6NatW63j/fJL6NkTjjgCXnoJ6jntxlZKKaWU\nEvfeCzfeCKWl3jb18syee8p+zTFjtn18+HB4/PGth+488v33ckjwkEOq/3hhYSHdu3cH6G6tLUzm\ntVyncVlsXxUCYFnFxxJmrZ0ATKjhY/2r/P7Y6p5X5TnXAb7UYFi1SlooZ2fDlCk6GVZKKaVU8nJy\nYMMGWLJka3nf0Cgrgx9/3LbCREz79jLo8nJPfzx8//1SOOCLLzy7ZI1cR/0hcKsxpkHsAWPMLkj1\nhg+9GFiYNWkiWyRmzJBfK6WUUkolKydH7kO5bWLZMtiypeYtE1u2yHKuh5o3D3/r5quBI4BiY8xb\nxpi3kENwh1d8LK0ZI/V/O0S+EbQ3qp6eVfHR3Nxobm40NzeaW+I0Mzfz58+nUyd49NGtE+NQqa4G\ncYxP3epiE2KH3b0Jc5oQW2u/APYGbkT2+84F/grsba39n3fDU4n68Uc4/nj46qvUvebQoUNT92Jp\nRHNzo7m50dzcaG6J08zcDB06lKwsuOyyEG6XgNonxD51q2vRAjZuhPXrPb1stZx3v1prS4FHPRyL\n8kCLFvDDD1L94q23UrMpv1ev8cydCwcd5P9rpZPx48cHPYRI0tzcaG5uNLfEaWZuvMjt44+hQQM4\n8EAPBlRVcbFcvHnz7T/Wpo1MOHxYIQZpNuJ3A0Tnnc/GmM7GmPGxLRMVv97Xy8GpxDVoAOPHwzvv\nyIE/vy1bBjfemM2//uX/a6UbLU3kRnNzo7m50dwSp5m58SK3m26C22/3YDDVKSqS1eHqVtrq14fW\nrT3vVhebEKdiH7HThNgYcw7wBdAd+Kzi1g34vOJjKkC9e8M558g+51Wr/H2t226TA6V//au/r6OU\nUkqp2pWU+Ni2uaYaxDE+1CIO/YQYGAfcbq3taa29ruJ2ODCm4mNp4ddfZVK5bl3QI0ncPffAmjUw\nYoR/r7FwITz0ENxwg2zVUEoppVRwVqzw8f/j2ApxTdq1y8gJcTvgyWoef6riY5G3eTOce67Umfb4\nJwApscceMhl+4AGYO9ef17jlFvmLV1Y2dsdPVtup2itexUdzc6O5udHcKtm4Ef75zx3ux9PM3CSb\nm7U+rxDvaELswwpxs2awfDmceaanl62W64T4XeCoah4/EnjfeTQhYS1cdRW8+y5MmwadOgU9IjfX\nXAP77gsDB0qtbC/Nmwf/+pdMirdsKfX24hmitFRzc6G5udHc3GhuyGGRv/1Nao1edBFcfHGt9WY1\nMzfJ5rZunXzP4ssKcVmZTHZ3tGXC4xVEY2SCX7eup5et/rUcWzdfAfwNeAb4qOLhw4BzkeYcv32L\nYK19KflhpkasdfN11xVw993dmDgRBgwIelTJee89eP55uOMOOXDnldNOg6+/lhbW9et7d12lkCKy\nFAAAIABJREFUlFIh8dlncN998PTTMiO5+GL4wx+ktuett8p+OeWL556DffaBvLz4P+eHH6Sz8uuv\nQ69eHg/oxx9lMjxjBpx6avXP+fvfpcTVxo2pmcESjtbNsRbLAytu1X0MwAKpScVDd98NQ4dGfzIM\ncPTRcvPSd9/BzJmyQqyTYaWUSiNlZfDKK3DvvVKuaPfdZXX4ssu2bujs00cOkFx/fcomPpnmmmvg\nkksSmxDHFu19WSGurQZxTLt28v5ZvhzatvVhEP5ybcxRJ85bJP+mHHusj2VL0kDHjjB/vuyxVkop\nlQbWrJFDJ507wxlnwIYNMHUqfP+9rBBVrj07aJAsR772WnDjTXM5OYm3b44dPPNlD3FxsdzvaMsE\neL6POFWc6xBXZYxp5tW1gjZqlJQSUzXr1GlrRiUe9y7PFJqbG83NjebmJu1zW7gQ/vIXmehcdx0c\ncgh89BH85z9w3nnV/xjwkEPg4IPhwQervWTaZ+aTyrm5TIhPPBFKS2tfxHVWVCSdMXbdtebnZOKE\n2BhzgzGmT6XfPwv8Yoz50RjjR3+UlNpll6BHEC0D0mFvSQA0NzeamxvNzU1a5mYtvP++FKzv1EnK\nKQ0aJDOwp5+GQw/d8TUGDpS9cwsWbPehtMwsBSrnttdeiU+IQeYvvizoxWoQ19b+tnVrefEolubC\nfYX4CqAIwBhzInAC0Bt4DbjTm6GpqBg5cmTQQ4gkzc2N5uZGc3OTVrnFyqYdfLAcLvnqK9kLXFQE\nY8bU/uPwqvr2lW0UDz+83YfSKrMUqpxbTg4sWQLr1wc3nm3sqOQaQL160sI5k1aIgbZUTIiB04Bn\nrLWzkKYcPbwYmIqObt26BT2ESNLc3GhubjQ3N2mR27Jlshdwzz2lbFqbNlKK4H//g8svlx+FJ2qX\nXeTk+aRJ8nP6StIiswBUzi0nR+4XLQpmLNuJZ0IMvjTneO896N/f00tWy3VCvBKIJdMbeLPi14YI\nVpXINN98A3fdFfQolFJK+WrePLj0UsjOlpPiZ50lq8Kvvip1uWr78Xc8rrhCWrruoFGHSlxsQvz9\n98GO4zc7atsc40NzjqIi2dXj92q564T4eeBpY8wbQAtkqwRAV2D7DUUqVD74QKrlvPtu0CNRSinl\nqfJyePllqRV84IEwa5bUDC4uhgkTpFuTVzp2hN695XCdQ08DVbP27aVaxOrVQY8E2LJFJrnxrBD7\nMCFOVftm1wnxtcB44EvgRGvt2orH27FtHWIVQhdfDEccIWcoNm3a8fNHj4Zhw2r++MSJE70bXAbR\n3Nxobm40NzeRya1y2bT8fNnGMGWKLDHecMO2ZdO8NGgQFBbCnDm/PRSZzEKmcm5160o53379AhxQ\nzM8/yzda8a4Qe3yoLvbWXbHC08tux7UO8WZr7f9Za6+21v630uP3WGv/4d3wlB/q1JGFgq+/ltrr\ntVmyRH7SVlvr58LCpJrDZCzNzY3m5kZzcxP63CqXTbv2Wjkw9+GHcuvTx//uSb17y97kCVvXwkKf\nWUiFNrdYDeJ4V4iXLpVVZY+kaoXYqXUzgDHmQuBPwF5AT2vtD8aYa4CF1toXPRxjysRaNxcUFGTE\noYDrrpNOi/Pn1/w+HzQIJk+WRYZmaVNpWimlIsxamD1bVjSmT5d/nC+/XP7BTqRShFfGjYNbbpGJ\nky9dIVQ8zj0X/vQnOOEEjy/8zDPyzdXKlTueCMyYAaefLq2eY3WJk/TLL9J9b9o0OPvsbT/mZetm\n1zrEVwJ3I3uHm7H1IN2vwDXJDEilzsiR0LSptIiszoIF8Mgj8Ne/6mRYKaUCV14OTz1Vfdm0228P\nZjIMUm0CQLdKBKa0FJ57ThZnPVdcDA0byoRhR3xoztG0qZz/DOse4quAP1prbwPKKj3+KZCb9KhU\nSjRpAnffDc8/X30HzuHDpTrPVVelfmxKKaWqeOghuPBCaYAwc2ZyZdO81LKlrCA+/DCUle34+cpz\nsf21LVr4cPFYybV4qpL4MCGuW1cW5UK5hxjIAf5bzeMbgYbuw1Gp1qePtHssKNj28cJCOY8xcqR2\n7lNKqcBt3Ah33AEXXCArGCedlHzZNC8NGiRFc6tbXVG+i3V99mXHSrw1iAFatZIZrMcH6844Q6oH\n+sl1QrwQOKiax3sDX7kPR6WaMVKS8uabt338xhulOs8ll+z4Gvn5+b6MLd1pbm40Nzeam5vQ5PbE\nE7Ivs7aSP0Hq0QO6d4cHHwxPZhGTTG6+rhDHW4MYZDLctq3npdcee8z/ihuuE+K7gQeNMX2QZhyH\nGGOGAbcj3epUhNSrt+3v166VAti33bb9x6ozePBgfwaW5jQ3N5qbG83NTShy27xZ9gife663dYS9\nZIysEs+cyeCqJ59UXJJ5r4VmhRh86VaXCnFMd7Znrf2HMWY9MBrIAp4GfgKuttZqy5qIa9QI/v3v\n+J/fq1cv/waTxjQ3N5qbG83NTShye/pp2Y7wYsgLOPXtC3/5C72++CLokURSde+1P/wBOnSAMWNq\n/9ySEqmw16iRx4PavFm2PyRyYNOH5hypkPAKsRHZwDRr7d5AI6CttXZ3a60eMU0TxoRre5pSSmWk\nsjKZDZ1xBuTlBT2a2u2yi1ScmDRJyh6opK1ZA3Pn7vh5K1bI6rDn/2///LOU+UtkhdiH5hyp4LJl\nwiDtmfcAsNaWWmuXeToqpZRSSkkN2G++2f6gR1hdeSX8+itMnRr0SNJCTo70XtmRzp1lgd5zRUVy\nn+iEOBNWiK215cC3gB9bt1UETZ8+PeghRJLm5kZzc6O5uQk0t/JyOczRu7fUHo6Cjh2Z3rUrPPig\nrCyquFX3XsvJkd0yO4qyb18po+q5WJe6RLZMtGsHy5bJdosIcT1U91fgTmNMFy8Ho6Jp8uTJQQ8h\nkjQ3N5qbG83NTaC5TZ8utYaHDw9uDA4mZ2VJLc85c4IeSqRU917LyYENG2DJkgAGBLJC3LhxfE05\nYtq3lxm8L11C/OPUutkYsxI5TFcP2ASsr/xxa21zT0aXYpnWulkppVRIWStlzHbdFd56K+jRJKas\nDDp1km56TzwR9Ggibd48OPBA+OADOPzwAAZw9dXw5pvyjVm85s6Frl3h44/hkEM8G0pZmeyRrlNp\nKdfL1s1OVSbQ9sxKKaWUf159Ff77X3j77aBHkri6deGKK2DECLjrLp9qgWWGnBy5X7gwoAlxIjWI\nY2Ld6jw8WDdnDhx2mHyD0MWnvQmuZdfi+pbPGPNX4GFr7a8ur6OUUkplHGth1Cg44gg45pigR+Pm\n0ktlQjxpEgwdGvRoIqtxY2m2Ec/BOl8UFSVe3aRlS2li4OHBumbN5K+Fn+2bXfcQx+smIJLbJ5RS\nSqlAvPWW/Lj55pujW/+yZUvo0wceekh+1q2c3XEHHH98QC9eVJT4CnGdOp53q2teMZP85RfPLrkd\nvyfEEf2brBLRv3//oIcQSZqbG83NjebmJpDcRo2SqhInnZT61/bAb5kNHCglEl57LdDxREVN77XL\nLoOePVM8GIBNm+RgXCIl12I8Lr3WrJncR3lCrDJAKDo5RZDm5kZzc6O5uUl5bu+9J7fhwyO7Ovxb\nZoccIgcDJ0wIdkAR4fpe27RJGnh4XuXup58Sb8oR4/GEuF49KXShE2IVav369Qt6CJGkubnR3Nxo\nbm5Sntvo0bJn8/TTU/u6HvotM2Ng0CCYORO++y7YQUWA63tt9mxo0sSHiF1qEMf40K2ueXOdECul\nlFLp7+OP4Y03or13uKo+feTn3Q89FPRI0lZJidy3auXxhV261MX40K1OJ8RKKaVUJhg1CvbbD845\nJ+iReCcrCwYMkGoT69fv+PkqYStWyJaCJk08vnBRkexTaNw48c9t1w6WL5f9HB5p0SLaVSbep0rT\nDpV+Zs+eHfQQIklzc6O5udHc3KQst//+F155BW66advOAxG0XWZXXgkrV8KUKcEMKCJc32slJTJZ\n9PyHCi41iGNitYg9bLE3fDgMGeLZ5bbj/LfOGFPHGLOPMeZIY8zRlW+x51hrT7HWeruJRIXOuHHj\ngh5CJGlubjQ3N5qbm5TlNno0dOwIffum5vV8tF1mHTtC797w4IM+nPxKH67vtRUrfOp9UlTktl0C\ntk6IPdw2ceSRcOihnl1uO06NOYwxhwFPAx3YvrSaBeomOS4VIVP0u34nmpsbzc2N5uYmJbl98QU8\n/zxMnCg/+464ajMbNEgOCn7yiaftfNNJbe+1zz6DVaukG3ZVsRVizxUVSQtmFz50q/Ob6wrxw8Cn\nQBek8caulW7aiCPDZGVlBT2ESNLc3GhubjQ3NynJbcwYyM6GCy7w/7VSoNrMTj4ZOnSQVWJVrdre\na3ffDX/9a/Uf822FuLjYfYW4RQuoX9/zg3V+cp0Q7w3cZK39ylr7q7V2VeWblwNUSiml0tY338DU\nqTLb2WmnoEfjn7p1ZS/x1KlbyyKouOXk1Ny+2ZcV4o0b3ZtygGxobtcuIybEHwOdvByIUkoplXHG\njIE2bSATOgleeqncT5oU7DgiKCdHzqdVV6jjoYfgmms8fsHYRNb1UB34UnrNT64T4geAu4wxlxhj\nuhtj8irfvBygCr8hfh77TGOamxvNzY3m5sbX3BYuhKeegqFDoUED/14nxWrMrGVLOO88mcGVlaV2\nUBFQ23stJ0fuFy3a/mMHHwz77+/xYJKpQRzjQ3MOP7lOiKcB+wGTgE+AucB/K92rDJKdnR30ECJJ\nc3OjubnR3Nz4mtsdd0i3gcsv9+81AlBrZoMGyaxu5syUjScqasstNiH+/vsUDSY2Ic6gFWJjHUqg\nGGM61PZxa+0PziMKkDGmG1BQUFBAt27dgh6OUkqpdFVUJOXIRo2CG24IejSpYy306AGtW8OrrwY9\nmsgoK4NddpHDdYMHp+AFx46Vb9hWrnS/xpgxMmCP9oxv2iRb0A8/XP7qABQWFtK9e3eA7tbawmSu\n77RCbK39obZbMgNSSiml0t6dd0oHsIEDgx5JahkjX/PMmfDdd0GPJjLq1oU996z5YJ3nkqlBHNO+\nvZTA2LjRkyEZAxddBO+958nltpNUOxxjzP7GmN7GmPzKN68Gp5RSSqWdJUvg0UflJJRLW9yo69sX\nmjWDhx8OeiSRstde8MsvKXqxoqLktkuA57WI69eXvy5+ZeA0ITbG7GWM+Qz4AngFmF5xe6HipjLI\n/Pnzgx5CJGlubjQ3N5qbG19y+7//kxJrV13l/bVDYIeZZWXBgAFSbaK6sgkZake5vfIKPPZYigaT\nTA3iGB+ac7RoIYvOfnBdIb4PWAi0BkqBA4CjkWYdx3gyMhUZQ4cODXoIkaS5udHc3GhubjzPraRE\nqixcdZWskqahuDK74gpZ6tMOir/ZUW51U9kD2KstE+DpwbrmzUO2Qgz0BG6x1pYA5UC5tXY2cCNw\nv1eDU9Ewfvz4oIcQSZqbG83NjebmxvPc7rlHNkN6Xjg2POLKrFMn6N0bJkzwf0AR4fJemz0bnnjC\n44Fs2ADLlye/ZWLXXWHnndN+QlwXWFPx6xKg4tsAfgA6JzsoFS1azsmN5uZGc3OjubnxNLeVK+GB\nB6Rjmy+9dsMh7swGDYJPP4U5c/wdUES4vNemTZNiEJ768Ue5T3aF2IdudWGcEH8BHFjx64+BocaY\nI4BbgFRVyVNKKaWi44EHYPNm+Mtfgh5JOJx8MnToAA8+GPRIImvFCh++t/KiKUeMx7WIwzghHl3p\nc28BcoD3gVOAP3swLqWUUip9rF4N994Lf/wjtG0b9GjCoW5dWS2fOtWzWrWZpqTEhwlxcbHcJ7tl\nAjzvVteqFZSXe3a5bbjWIX7dWvt8xa8XWGv3BVoCra21b3s5QBV+Y8eODXoIkaS5udHc3GhubjzL\nbcIEWLdO2jSnuYQyGzBAmnVMmuTfgCLC5b22YoVUXvBUUZEsxWZlJX8tj1eI//Y3mDvXs8ttI9k6\nxJ2MMScZY3ax1qaqOp4KmdLS0qCHEEmamxvNzY3m5saT3Natg7vugv79vVl1C7mEMmvVCvr0kZrE\nZWX+DSoCXN5rvqwQe1GDOMbjPcR+cm3d3AJ4BjgWsMDe1trvjTGTgJXW2khukNLWzUoppTx3zz0w\nZAh8+y3k5AQ9mvD5+GM47DCYMQNOPTXo0YTao49KTC++KL9v1gxuvhmuv97DF8nPl30JM2Ykf60n\nn4SLL4bSUuk97bHAWzcD9wCbgWykDnHMVKB3MgNSSiml0saGDdKm+cILdTJck0MOgW7d9HBdHEpL\nYdYs2WWyeTOsWuXTlgkvDtSBL805/OI6Ie4F3GCtLa7y+LdAh+SGpJRSSqWJSZNg6VK48cagRxJe\nxkgJtpkz4bvvgh5NqOXkyPdYS5bAmjVSpMPzM5rFxd5tmciACXFDtl0ZjmkObHQfjoqiEj0d7ERz\nc6O5udHc3CSV26ZNUiS2Tx/YZx/vBhVyTpn17Ss//3/4Ye8HFBHx5Bb7IcPChXLubdEiqV7nmfXr\nZWOy1yvEEdhH7Dohfh+4qNLvrTGmDjAUeCfpUalIGTBgQNBDiCTNzY3m5kZzc5NUbv/8p/z4edgw\n7wYUAU6ZZWXJocNJk2RSloHiya3yhNgXsZJrXk2ImzaFBg3SekI8FLjcGPMasBMwDmnWcTRwg0dj\nUxExcuTIoIcQSZqbG83Njebmxjm3LVtgzBg45xw44ABPxxR2zpldeaV0XZg61dPxREU8uTVqJFUl\nfJ8Qe7VlwhjPS6/5xbUO8RdIi+bZwIvIFornga7WWt0AlGG0Iocbzc2N5uZGc3PjnNvkyfD99xm3\nOgxJZNapE/TunbGH6+LNLSfHxwlxrEudl+UBPW7OcdFF8NBDnl3uN/WS+NwNwBvAZ2ydWPcwxmCt\nfSnpkSmllFJRVFYmq8OnnQZduwY9mmgZOFDKfs2ZI9Un1HZ8nxC3aOFtiTSPV4i/+EJWyr3mNCE2\nxvQG/okcojNVPmyBukmOSymllIqmadNg/nx4/PGgRxI9p5wipRMmTNAJcQ3OP9/HTtfFxd7tH45p\n1w4+/9yzyzVvLjtrvOa6h/gBpDFHe2ttnSo3nQxnmIkTJwY9hEjS3Nxobm40NzcJ51ZeDqNHw4kn\nwqGH+jOokEvqvVa3LlxxBUyZ4uOsL5zize2MM+DSS30ahJc1iGM8XiEO24S4DXC3tXapl4NR0VRY\nmFRzmIylubnR3Nxobm4Szu3ll2U1bPhwfwYUAUm/1y69VDpPTJrkzYAiIhR/R72sQRzTvr10EPGo\nfXyLFrBihSeX2oZr6+ZJwAfW2rRactDWzUoppZxZCz16QMOG8O9/Bz2aaLvoIpg9W9pd19UfPFen\nb1/Iy4ObbvLwoi1aSB9oLxvJvP02HH88LFgAHTsmfblhw+Dpp2UftZetm10P1Q0GnjXGHAV8jrRx\n/o219v5kBqWUUkpFzsyZUFAAb7wR9Eiib+BAqeM8cyacemrQowmlggLIzvbwgqWlshfBjy0TINsm\nPJgQ+7VlwnVC3A9p37wBOAY5SBdjAZ0QK6WUyhzWwqhRcNhhshqmknPoodCtmxyu0wlxtUpKpCax\nZ7yuQRzTrp3ce7SPuHlzWL0aNm/e8XMT4Tohvg0YAdxhrS33cDxKKaVU9LzzDnz4IcyYIc0IVHKM\ngUGD4LLLpJ7zXnsFPaJQ2bIFfv1Vdjh4JlaD2OsV4iZNpBOhRxPiAw+Ea66RDLzkeqhuJ2CqToYV\nQH5+ftBDiCTNzY3m5kZzcxN3bqNGyYrmKaf4O6AI8Oy91rcvNGvmTxeGEEokt9iWAU9XiGMT4t12\n8/CieN6trls3uOceb0slg/uE+Amgj5cDUdE1ePDgoIcQSZqbG83NjebmJq7cZs+Gd9+Fm2/W1WE8\nfK9lZUH//lJtYv16b64ZYonkNn++3Hu6QlxcDK1aQYMGHl60gsfd6vzgWmXifuAipEvdPLY/VHed\nJ6NLMa0yoZRSKmG9e8OPP8Jnn0Ed13UmVa0FC2DvveGxx+CSS4IejSgvlxIHOTmB/Xn/8Y/wj39A\nYaGHzRD/9Cf49FM5ree1fv1g6VKpOOEhL6tMuP5J5gL/BcqBLkDXSreDkhmQUkopFRlz5sDrr0st\nKJ0Me69TJzjpJHjwwWDHUVICkydLObh27WRc06YFNpwbb5Szm/vu6+FF/ahBHNOunafNOfzgdKjO\nWnus1wNRSimlIue222CffeDcc4MeSfoaNAjy8+GTT6TOcyqUlclq6WuvSem3OXOkksiBB8o2jief\nlMcC+nPfay94802PL1pUBEcf7fFFK3jcrc4P+u2sStr06dODHkIkaW5uNDc3mpubWnP77DN46SXp\njKDNI37j+XvtlFOgQwf/V4mXLpWJ7vnnQ5s2UkLvvvvktSdOlG0xc+fCHXfAwQdLR0IPBf531I+2\nzTHt28OaNbB2rT/X94BOiFXSJk+eHPQQIklzc6O5udHc3NSa2+jRso/0/PNTN6AI8Py9VrcuXHEF\nTJkiWxe8smWLHIi8+Wbo3h3atoWLL5bueFdeCR98AMuXw9SpsiocazAB0iJu3jzvxkLAf0fXrpU6\nbn5tmYhlF+KDdU6H6tKVHqpTSikVly+/hC5d4OGH4fLLgx5N+lu+XCZro0fDkCHu1/nxR9nz/dpr\n0lFw1Sop1XDSSXDyydCrF7RuvePrTJ0qZeGWL/e49llA5s+H/faTluN+bJv45hvo3Fmqsfzud0lf\n7pdfYMMGWLIk+NbNSimlVOYaM0bqtV58cdAjyQytWsF550lN4uuui3+LyqZN8J//bN0LPG+elMY7\n9FC5Tu/esjqc6JaXvDy5//xzODYNjlXFahD7eagOPNtH/Kc/Sbe622/35HKAToiVUkqpxCxYIBUH\n7rsPdt456NFkjkGD4KmnZGJbWzvnH36Q57z2Grz1lmwHaNNGJr833ggnnph8Ad+995Y/+3SZEMfa\nNnvdlCOmcWNo1MjT9s2LFnlyqd/ohFgppZRKxO23y4rlpZcGPZLMcuih0qZswoRtJ8QbNsD7729d\nBf7qK1nxPfxwmQCffLJUh/CyLF69erD//p7vIw5MUZF80+DnN3geNudo3nxrtz6v6KE6lbT+/fsH\nPYRI0tzcaG5uNDc32+W2aJFUIhgyxPvesWnCt/eaMTBwoEx8334bxo+H006T1d5eveCZZ2QS/Nxz\ncvjuvfekAkjXrv7UiM7L87TSRKB/R/2sQRzjYem15s1hxQpPLvUbXSFWSevVq1fQQ4gkzc2N5uZG\nc3OzXW7jxkHTplL1QFXL1/dav37yzcjxx0P9+nDkkTBypGyH6NIlta2zc3Ph2WelZrEHZfcC/Tvq\nZ8m1GA+bczRvLucht2zx5HKAToiVB/r16xf0ECJJc3OjubnR3Nxsk9uPP0o92hEjoGHD4AYVcr6+\n17KypPbzihVw3HGyNzUoeXlQWgrffy97ipMU6N/RoiLJ00/t20uzEw/EtoCvWePJ5QCdECullFLx\nufNOmZANHhz0SDLbkUcGPQKRmyv3n3/uyYQ4UBHcMgGySuwV3UOslFJK1WbdOulO9vDD8Oc/Q5Mm\nQY9IhUGbNnK4MuoH69askZml31sm2reXv0seLOvGJsSrVyd9qd/ohFglbfbs2UEPIZI0NzeamxvN\nzcHGjcy+9lro2BFuuQX++Ee44YagRxV6GfNeM0ZWiT06WBdYbn7XII6JdavzYJW4UycoKPB2YV4n\nxCpp48aNC3oIkaS5udHc3GhuCSgrg8cfh86dGXfvvVK265tv4IEHZMuEqlVGvdc8bOEcWG6xGsSp\nOFQHnkyIGzSQCnxeFnrRCbFK2pQpU4IeQiRpbm40NzeaWxyshWnTZNWvf3/o0YMpBQXw2GOw555B\njy4yMuq9lpcH330nWwGSFFhuRUWy2h1bwfWLx93qvKYTYpW0LF0xcaK5udHc3GhutbAWXn8devSA\n3/8esrPlNPyzz5LVrVvQo4ucjHqv5ebK++d//0v6UoHlVlws+6F32snf12nUSPbf64RYKaWUCpkP\nPoBjjpE6tjvvDO++K93OuncPemQqCvbfX5p+RPlgXSpqEMd42K3OazohVkoplXnmzpUuZ0ceKSfs\nZ8yA2bPhd78LemQqSrKy5ISXhx3rUq6oyP8DdTEell7zmk6IVdKGDBkS9BAiSXNzo7m50dwqfPON\ndDvr2lV+PWUKFBbCqadW2+VMc0tcxmXm0cG6wHIrLk7dCrGH3eq8phNilbTs7OyghxBJmpsbzc1N\nxudWVCRl0/bfX1aCH31U9n326SM/8q5BxufmIOMyi5VeszapywSWW6q3TIR0Qmxskn+A6cQY0w0o\nKCgooJsepFBKqehbvhxuvx0mTJA2v8OGwRVXSN0mpbwwfTqcdZa09va7UoPXVq2CZs1g8mTo29f/\n17v3XrjpJqnKUc1PZBLx/vswbVoh993XHaC7tbYwmetp62allFLpZ9UquOsuuOceWQEeNgyuuUYm\nxUp5KS9P7j//PHoT4lTVII5p3x7Wr5cWc02bJnWpDz+ESZM8Ghch2jJhjBlkjFlojFlvjPnIGNOj\nlueeZYyZZYxZZoxZZYz5jzGmVzXPO9cY81XFNT8zxpzs71ehlFIqUKWlMG4c5OTAnXfClVfC99/D\n8OE6GVb+2HNPaNgwmpUmYl3qUjkhBk+2TTRv7kkX6N+EYkJsjOkD3AWMALoCnwGvG2Na1vApRwOz\ngJOBbsA7wMvGmAMrXfNw4GngUeAg4EVgujFmf7++jkw1f/78oIcQSZqbG83NTdrntmkTPPSQnPgf\nNkx+/PvddzI5btHC+bJpn5sPMi6zOnU8aeEcSG7FxbJ1IdY0w28eNudo3jzpS2wjFBNi4Frg79ba\nJ62184ErgFJgQHVPttZea639P2ttgbX2O2vtMOBb4PRKT/sz8Jq19m5r7dfW2luAQmCwv19K5hk6\ndGjQQ4gkzc2N5uYmbXMrK4OnnoL99oNBg+D442H+fNkz7MGPr9M2Nx9lZGa5uUmvEAdENZlLAAAg\nAElEQVSSW1GRTFLr10/N6+mEuGbGmPpAd+Ct2GNWTvq9CfSM8xoGaAz8UunhnhXXqOz1eK+p4jd+\n/PighxBJmpsbzc1N2uVmrRxmOvBAuPDCraWv/vlP6NjRs5dJu9xSICMzy8uDL7+EzZudLxFIbqms\nQQxSt7lZM0+ac6TdhBhoCdQFllZ5fCnQNs5rDAEaAs9UeqxtktdUccq4Ejse0dzcaG5u0iq3t96C\nww6Tk/1t28JHH8ELL0CXLp6/VFrlliIZmVlurkyGv/nG+RKB5JbKGsQxHpVe83pCHPkqE8aY84Hh\nQL61tiTo8SilVNrZvBkWLZK9knXrbr1V/X11H0uytNI2PvpI9ge//TYccgi8+aZskVAqaLm5cj9v\nHhxwQLBjSURRkbQtTyWPmnMkcTSgWmFYIS4ByoA2VR5vAyyp7RONMX2BR4BzrbXvVPnwEpdrApxy\nyink5+dvc+vZsyfTp0/f5nmzZs0iPz9/u88fNGgQEydO3OaxwsJC8vPzKSnZds4+YsQIxo4du81j\nixcvJj8/f7sN9g888MB2nWxKS0vJz89n9uzZ2zw+efJk+vfvv93Y+vTpo1+Hfh36dejXEf/XsWAB\ndOsG++zDiE6dGJuTA9nZsNtu0K4di1u3Jr9FC+Y3ayZVHLKyYOedeaB+fYbUqSMT4/r1oUEDSrOy\nyK9Xj9mNGsn/Zq1bQ7t2TG7enP6NGsFee8mhuM6dYf/96dO0KdNzcqB7d9ka0bMns777jvwePWRy\nXGkynDF/Hvp1hPPraN6cEY0bM/aRR6LzdRQUkP/tt5Tsuus2j/v+51FlhTjer2Py5Mnk5ORw0EEH\nkZ+fT58++TRseO12X6urUDTmMMZ8BHxsrb264vcGWAzcb629s4bP6Qf8A+hjrZ1RzcenALtYa8+o\n9NgHwGfW2oE1XFMbczgYO3YsN9xwQ9DDiBzNzY3m5sYpt5kzpc1xq1ZSUH+XXeQQW9VbeXn1j3v5\nsfJyOPpoqR5Rt64/IVVD32+Jy9jMTjlFvgGcsd2UJC4pz+3XX2HXXWHqVDjvvNS97l//Cs88I+UQ\nk1RYWEj37unVmONu4HFjTAEwB6k6kQU8DmCMuR1ob629uOL351d87M/AJ8aY2Erwemvt6opf3we8\na4y5DngF6Icc3vtjKr6gTFJaWhr0ECJJc3OjublJKDdr4Y47ZHvCqafKQbVmzfwbXIjp+y1xGZtZ\nXp50fHOU8txSXYM4pn17OVRnrbdbqpIUihViAGPMQGAosq1hLnCVtfbTio89BnSw1h5X8ft3kFrE\nVT1hrf2tVJsx5hzgNqADUpZtiLX29VrGoCvESqnMtnYtXHIJTJsmzSxGjpRVL6VU7f71L7jgAli5\nMhrfQL72mqxqL16c2knxc8/BuefCL7/ICnUS0nGFGGvtBGBCDR/rX+X3x8Z5zWnAtORHp5RSGeDb\nb6Vyw+LFUrXhzDODHpFS0VG5hfNRRwU7lngUFck3u6lqyhFTuRZxkhNiL+m3/UoppWS1qEcPqSjx\n8cc6GVYqUZ07Q716SXesS5lYU456KV4b9bB9s5d0QqySVvXEsIqP5uZGc3NTY27Wwpgxslf4qKNg\nzhzp+qYAfb+5yNjMdtpJ/u44dqxLeW5B1CAGT7vVeUknxCppAwZU22Fb7YDm5kZzc1NtbmvWwO9/\nL4fnbrkFXnwRmjZN/eBCTN9vicvozHJznVeIU55bUVEwE+IGDaSrhgfd6rwUmj3EKrpGjhwZ9BAi\nSXNzo7m52S63b7+VbRFFRdIC+Ywzqv28TKfvt8RldGZ5efDyy04VFFKeW1HR1n3PqeZRtzov6Qqx\nSppW5HCjubnR3Nxsk9srr8h+4bIy2SKhk+Ea6fstcRmdWV6e/OTlhx8S/tSU5mZtcFsmwLNudV7S\nCbFSSmWK8nIYPRpOPx1+9zs5PLfvvkGPSqn0UbmFc5itXAmlpcFNiHWFWCmlVCBi+4WHD4cRI6Ss\nmu4XVspbu+0mpcTCPiEuLpb73XcP5vV1QqzSUdX+6Co+mpsbzc3BN98wsXNneOstOTg3YoQ224iT\nvt8Sl9GZGeN8sC6luQXVpS6mcre6kNB/EVXSCguTag6TsTQ3N5pbgmbMgB49KFy/XvYL5+cHPaJI\n0fdb4jI+s7w8pxXilOZWVAR160Lbtql7zcrat4dNm6RbXUiEpnVzGGjrZqVU2ojtFx4xQg7NPfkk\nNGkS9KiUSn+PPAJXXgnr1kmJsTC6+Wb5N2Hx4mBe/8MP4fDD5RuH2L5rB162btYVYqWUSjerV8PZ\nZ8tk+NZb4fnndTKsVKrk5ck3pF9+GfRIahZUDeKYEHar0zrESimVTr7+WuoL//ST1EM97bSgR6RU\nZunSRe4//xzC+tPmoqLgDtTB1q0aIWrOoSvESimVLl56CQ45RA72fPKJToaVCkKjRrDXXuGuNBFk\nDWKAnXeGli1DtUKsE2KVtHw9pONEc3OjuVWjvFy2RpxxBhx3HHz0EeyzzzZP0dzcaG6J08yQbRMJ\nVppIWW7WBr9lAkLXnEMnxCppgwcPDnoIkaS5udHcqli9Gs46SybEo0bBtGnV7hfW3NxobonTzJCD\nYgmuEKcst19+gQ0bgt0yAaGrRax7iFXSevXqFfQQIklzc6O5VTJ/vuwX/vln2S986qk1PlVzc6O5\nJU4zQ1aIly6FZcugdeu4PiVluQVdgzimfftQHTzUFWKllIqi2H7hunVlv3Atk2GlVIrFSok5NOjw\nXWxCHIYVYj1Up5RSykl5OYwcKfuFTzih2v3CSqmAdeokNYjDeLCuuBjq1YM2bYIdR2xCXF4e7Dgq\n6IRYJW369OlBDyGSNDc3GZ3bqlWyReJvf5OmG889B40bx/WpGZ1bEjS3xGlmyE9uDjggoRXilOVW\nVAS77SZjDFK7drB5M6xYEew4KuiEWCVt8uTJQQ8hkjQ3Nxmb21dfyRaJ996TdszDhkGd+P8Jz9jc\nkqS5JU4zq5BgC+eU5RZ0DeKYkDXn0NbNlWjrZqVUKE2fDhddJIdgpk+HvfcOekRKqR2591648UZY\nuzb41djKjj1WGmME/Y1LURFkZ8Orr8LJJztdQls3K6VUJigvh1tukbJqvXrJfmGdDCsVDbm5Ut5s\nwYKgR7KtMNQghtB1q9MJsVJKhdGqVXJwbvRouO02ePbZuPcLK6VCIC9P7sN0sM5aOVQXhi0T9etL\nSbqQbJnQCbFSSoXN/Plw6KHw/vuyX/imm6Qds1IqOlq1kkoOYSq9VlICGzeGY4UYQtWtTifEKmn9\n+/cPegiRpLm5SfvcqtYXPuUUTy6b9rn5RHNLnGZWSQIH61KSW1hqEMeEqFudTohV0rQrkRvNzU3a\n5lZeLu2XK9cX9nC/cNrm5jPNLXGaWSW5uXGvEKckt+JiuQ/LCnGIJsRaZaISrTKhlArE6tVw4YXS\nfvlvf5MtEgmUVFNKhdQTT8All8jf8TCcAXjwQbj2WjnsF4Z/Y265BR57bOvKdYK8rDJRL5lPVkop\nlaSvv5ZmGz/9JNslTjst6BEppbwSO1j3v//BYYcFOxbY2pQjDJNh2LZbXcBjCkkiSimVgWbMkP3C\nAHPm6GRYqXSz335yHiAslSaKi8OzXQLkUF1ZGSxfHvRIdEKskjd79uyghxBJmpubtMitvFzKqeXn\nS5H8jz+Gzp19fcm0yC0AmlviNLNKGjSAffaJa0KcktzCUoM4JkTd6nRCrJI2bty4oIcQSZqbm8jn\ntmYN/P73MHw4jBgBzz8PTZr4/rKRzy0gmlviNLMq4jxYl5LcwlKDOCY2IQ5Bcw7dQ6ySNmXKlKCH\nEEmam5tI5/btt7JfuKgIXnxRVohTJNK5BUhzS5xmVkVeHsyaJU0xaqkn7ntu5eXh2zLRpo1koivE\nKh1kZWUFPYRI0tzcRDa3V1+FHj1gyxbZL5zCyTBEOLeAaW6J08yqyM2FX3+FH3+s9Wm+57Z8OWza\nFK4V4nr1QtOtTifESinlJ2thzBg5MHfUUTIZ3nffoEellEqVsLRwDlsN4piQ1CLWCbFSSvll7Vo4\n91wYNkz2DL/4IjRtGvSolFKp1KGD1CAOuoVzrNavToirpRNilbQhQ4YEPYRI0tzcRCa3BQuk7ujr\nr8MLL0gXugDrbEYmt5DR3BKnmVVhjGyb2MEKse+5FRXBTjtBy5b+vk6iYrWIA6YTYpW07OzsoIcQ\nSZqbm0jkNnOm7BfetElKqp15ZtAjikZuIaS5JU4zq0Ze3g5XiH3PLVZhIixNOWJCskKsrZsr0dbN\nSqmkWAtjx0rr5ZNPhn/9C5o1C3pUSqmgTZgAV18N69bJKm0Qzj9fJp7vvhvM69fk73+HgQNlAaFu\n3YQ+1cvWzSH7NkEppSJq3Tro2xduvFH2DL/8sk6GlVIiL08qzMyfH9wYwlaDOKZ9eykJt2xZoMPQ\nCbFSSiXr+++hZ0945RWYNg1GjQrfjyWVUsHp0kXugzxYF7YudTEh6Van/2KrpM0P8jveCNPc3IQu\nt1mz4OCDobQUPvoIzj476BFVK3S5RYTmljjNrBrNmkF2dq0H63zNrbxc6iCHdYUYAj9YpxNilbSh\nQ4cGPYRI0tzchCY3a+HOO2Wv8KGHwiefbF0FCqHQ5BYxmlviNLMa7KCFs6+5LVsGmzeHc4W4dWv5\niZquEKuoGz9+fNBDiCTNzU0oclu3Tg6oDB0KN9wAM2bArrsGPapahSK3CNLcEqeZ1SAvr9YVYl9z\nC2sNYpCDdG3aBD4hrhfoq6u0oCV23GhubgLPbeFCOOss+PZbeOYZabwRAYHnFlGaW+I0sxrk5cm2\nhV9+gebNt/uwr7nFJsRh3DIBoSi9pivESikVrzfflP3Ca9bIfuGITIaVUiGQmyv3QRysKy6GBg3C\n15QjRifESikVAdbCXXfBSSfJhPiTT7b+56aUUvHYZx+pQbyDjnW+KCqS1WFjUv/a8QhBtzqdEKuk\njR07NughRJLm5ibluZWWwgUXwPXXy+3VV6v9cWfY6fvNjeaWOM2sBvXrw3771bhC7GtuYa1BHBOC\nFWLdQ6ySVlpaGvQQIklzc+N7bhs3SvH8L76Q/7hefhkWLYIpU6BPH39f20f6fnOjuSVOM6tFLQfr\nfM2tqAj22su/6yerXTtYulSal9QLZmqqrZsr0dbNSmWQ8nJpqBGb+Mbuv/kGysrkOXvsAQcdJI02\nDjww2PEqpaLvzjvh1lth9erUNu/p0AH+8AcYMyZ1r5mIV16B006Tlezddov707xs3awrxEqp9GYt\nLFmy/cT3yy9lOwTIFojcXDjuOLj6aqkn3KULNG0a7NiVUuklL0/KNi5cCB07puY1y8pkO0IYS67F\nVG7OkcCE2Es6IVZKpY/Vq2XCG5v0xibAK1bIx3fZBQ44QCa7/frJfW4utG0b3sMmSqn0kZcn959/\nnroJcWwrQhQmxAHuI9YJsUpaSUkJLcNayiXENDc3JSUltGzceNt9vrH7xYvlSXXryonuLl22rvjm\n5kJOjnwsA+n7zY3mljjNrBZt20KLFrKP+Mwzt/mQb7mFvQYxQKtW8m9zgBNirTKhkjZgwICghxBJ\nmlscysthwQKYPl328fbpw4CcHGjYUPb2XnABTJ4s2yL69YN//hPmzoW1a2VLxDPPwPDh0kijU6eM\nnQyDvt9caW6J08xqYYysEldTacK33IqL5T7MK8R16sg3C7pCrKJs5MiRQQ8hkmrNbeFCmDkTTjgB\n9t47ZWMKhfJy+OADePppePbZrdsdKvb5jjzlFNnrm5sr2x90n29c9O+pG80tcZrZDuTmyr/vVfiW\nW1GRNOUIe7nIgEuv6YRYJU0rcripMbcZM2Tlc9Uq+X3XrnDeedIVLVV7zlLNWvjsM5kET5ki/4Bn\nZ8Oll8o3BV26/LbPV99tbvTvqRvNLXGa2Q7k5cEDD8ih3qys3x72LbfiYlkdDvs5iYCbc+iWCaXC\noqwMRoyA00+Ho4+W75Sfe05WiEeNkh/5H3wwjBsnK8jp4LvvYPRoWent2hUee0y+/tmz5WscOxZO\nPFFqVIb9H3OllIpHbq4sAnz5ZWper6go3NslYgJeIdYJsVJh8MsvUoNx1Ci47TbZM9uuHZxzDkyd\nCsuWyX7YnBwYOVIKrPfoITUtFy0KevSJWbIE7r8fDjtMJvl33AHdu0sHuJ9+ggcfhCOOSG2NTqWU\nSpUDDpBv8FPVwjnWtjns2rXTCbGKtokTJwY9hEj6LbfCQpkQfvKJ7Cu76abtJ4MNG8qWiWeflcnx\nlCmypeCWW2SSfOihcNddW6sshM2qVfD449Crl9SYvP56aNNGvo5ly+Qw3MknS2vTHdD3mxvNzY3m\nljjNbAcaNpTtb1UO1vmWW2zLRNi1by//H2zeHMjL64RYJa2wMKnmMBmrsLBQtggcfji0bAkFBTJh\n3JFGjaSF8LRp8o/H00/LPyTDhkk3op494Z57tpbaCcqGDTLGc86Rye+AAfIP3cMPyyrxiy/K11Fp\nD1089P3mRnNzo7klTjOLQzUtnH3JLQpNOWJitYiXLAnk5bV1cyXaulmlzMaN8Oc/wyOPwGWXyQGL\nBg2Su+bq1fDyy7K1YuZM2LRJJtvnnQe//31quv9s2QLvvCOT9OeflzF17w7nny+T34A6ECmlVKjc\neiuMHy+LGn6ej4itDs+YAaee6t/reOGzz6Sc5scfwyGHxPUpXrZu1hVipVJt8WI46ih44gl49FG5\nJTsZBmjSRHrVv/ii/CP75JNSZmfIENk/dtRRMvH2eo+WtfDRRzLB3313WeX+4AO49lppnvHpp3Dd\ndToZVkqpmNxcKCmRLnJ+ikIN4piAu9XphFipVHrrLVkxXbpUKilcdpk/r9O0KVx4oawYL1sm+3eb\nNoW//EUmrb/7naxOJFPi5ssv4eab5WBcz55SEeP882Uv9Ndfy+G/zp29+oqUUip9xFo4+32wLgpd\n6mJatIB69XRCrFRas1aqKfTqJeXFCgqkhFoqNGsGF18sPzJbuhQmTZJDHddeK6u2xxwDEybEt1Kx\neLGUfTvoIDkp/eCD0iTjrbfkH96775avS0ukKaVUzfbaS85PVNOxzlPFxfI6u+7q7+t4oU6dQCtN\n6IRYJS0/Pz/oIYTbqlVw9tlw441ye+01aNkymNx23RUuuURKnC1dCv/4h2zX+POf5cdVxx0nh96W\nLdv6OSUl8tjRR8uhvREjZOV3+nQ5/PDoo/J5KWqLrO83N5qbG80tcZpZHOrUkYZDlVaIfcktVoM4\nKosUAdYi1k51KmmDBw8Oegjh9cUXMhleulT29lb6By/w3Jo3l8oPAwbIpHf6dDmQN3gwDBoExx4L\nO+8Ms2bJCveJJ8q+5DPPhMaNAxt24LlFlObmRnNLnGYWp9xcKbtZwZfcolKDOCbAbnVaZaISrTKh\nPDVlirQe7thRyo/tvXfQI4rP8uXwwgtS83jjRqkOce650Lp10CNTSqn0cf/9MHQorF0re2f90LMn\n7LuvlPiMgsGD4f33peJEHLysMqErxEp5bfNm+Ufu3nvlkNkjj8ie3aho1Qouv1xuSiml/JGXJ4sO\n334L++3nz2sUFclP96JC9xArlSZ+/ln2044fLyXOnnoqWpNhpZRSqZGbK/d+VZrYskX+T4ralomS\nEqmjn2I6IVZJmz59etBDCIfZs6FbN/juO/j3v+VHP7UcZNDc3GhubjQ3N5pb4jSzOLVoIRPAigmx\n57n9/DOUl0ejBnFMgN3qdEKskjZ58uSghxAsa+G+++QQ2j77yCGJww/f4adlfG6ONDc3mpsbzS1x\nmlkCcnN/K73meW5RqkEcE2BzDj1UV4keqlMJW7cO/vhHmDxZurHdcQfUrx/0qJRSSkXB0KFS3WfR\nIu+v/cwzcih65UqpRx8FJSVyjmXaNKnQtAPaulmpMPjmGzjsMHjpJZg6Fe66SyfDSiml4pebCz/8\nIPXqvVZUBI0aSZfSqGjRQv4fDWCFWCfESrl48UXo0UMqSsyZA+edF/SIlFJKRU2shfMXX3h/7VgN\n4qg05QAZa0DNOXRCrFQiysrgppukOcUJJ8hkeP/9gx6VUkqpKNp3X6lB7EcL5+LiaB2oiwmoOYdO\niFXS+vfvH/QQUmP5cujdG8aOhXHj4LnnoEkT58tlTG4e09zcaG5uNLfEaWYJ2Hln6NwZ5s3zPrdY\n2+ao0RViFVW9evUKegj+mzMHuneX7jlvvAFDhiT9Y6iMyM0Hmpsbzc2N5pY4zSxBeXnw+efe5xa1\nts0xATXn0AmxSlq/fv2CHoJ/rJVOc0cdJd+1FhZK4w0PpHVuPtLc3GhubjS3xGlmCcrNhXnz6Ne3\nr3fX3LxZavnqCnHcdEKsVE3Wr4fLLoM//QkuvVSabUTxu22llFLhlZcHq1fD4sXeXfOnn2RBJ4r/\nZ7VvD7/8Ahs2pPRldUKsVHUWLYIjj4Snn4bHH4cJE2Svl1JKKeWlWAtnLw/WFRfLfVRXiCHl3ep0\nQqySNnv27KCH4I01a+CFF2Q1+KCDpJj5hx/CxRf78nJpk1uKaW5uNDc3mlviNLME7bEHNG3K7Jde\n8u6asS51UZ4Qp3jbhE6IVdLGjRsX9BDcff89PPAAnHQStGwpnXE+/FC2SRQUyMTYJ5HOLUCamxvN\nzY3mljjNLEHGQG4u4155xbtrFhVB48ZJVUIKTLt2cp/iCXG9lL6aSktTpkwJegjx27JFJrwzZsjt\nyy+lK84xx8Cdd8Kpp0LHjikZSqRyCxHNzY3m5kZzS5xm5iAvjyklJd5dL6o1iAF23VW2KOqEWEVN\nVlZW0EOo3S+/wOuvywT4tddkK0Tr1jL5HTUKTjxRvpNOsdDnFlKamxvNzY3mljjNzEFeHll//zts\n3OjNeZWo1iCGwLrV6YRYpR9rYf78ravAH3wgHea6doXBg+G00+Dgg6GO7hhSSikVArm58v/UV195\ns1WvqAgOPDD56wQlgG51OiFW6WHjRnjvva2T4O+/h112kfbKEybAKadEs/yMUkqp9Neli9zPm+fN\nhLi4WBZ/oiqAFWJdIlNJGzJkSDAvvHQpPPYYnHOOHIjr1UuqRJx0ErzyCqxYAS+9BJdfHsrJcGC5\nRZzm5kZzc6O5JU4zc9CkCUOaNPGm9NqmTfL/Ywj/34tbAN3qdIVYJS07Ozs1L2QtzJ27dRV4zhzZ\na3TYYXDjjfLdcG5u0i2VUyVluaUZzc2N5uZGc0ucZuYmOztbVoiTFWvKEdU9xBDICrGx1qb0BcPM\nGNMNKCgoKKBbt25BD0cBlJbCW29tnQT/9JOUkTnpJJkAn3wytGoV9CiVUkqp5Nx8M0yalPxE8P33\n4eijpYrSfvt5M7ZU++c/4aKLZA6wyy41Pq2wsJDu3bsDdLfWFibzkrpCrMLnp5/gxRdlAvz229K+\nce+9oW9fmQQfeaSUSlNKKaXSRV6eHCQrKZFtgK5iTTmivGUi1pzj559hr71S8pI6IVbh8tVXsgWi\ntFS+wx0zRsqj7bNP0CNTSiml/JOXJ/effw7HHvv/7d13mFTV/cfx96FKEaVJUxQVsewiiqBYfhoL\nCtHVRA2WJAp2QY0mWEgUUgwuiEZBNFHsijGiBDuiGEOUsLJKQFgLCIiIgFRZ+p7fH2dXdmHbPVPu\nvTOf1/PMszI7c+c7H+7id++c4n+cJUtgjz1CWU40acpvzpGmhliT6iRhRUVFyTnQmjVw9tnut9pv\nvnFDJW68MWOb4aTllmWUmx/l5ke5BafM/BRt2+bWIE50HHGc1yAuE8L2zWqIJWE333xz4gfZvh0u\nvBBWrHDDJRL5uCgmkpJbFlJufpSbH+UWnDLzc/OQIXDYYclpiOM8XALcFe5GjdQQS7yMGTMm8YMM\nGQKTJ8Pf/w4HHpj48WIgKbllIeXmR7n5UW7BKTM/Y8aMcSslJbr0Wpy3bS5TtltdGjfnUEMsCUt4\niZ1nn4URI2DkSLeWcJbQ0kR+lJsf5eZHuQWnzPx07NjRjSOeM8d9auorE64QQ9qXXlNDLOGaORMu\nuwx+8Qs3XlhERCRb5ebCxo1ut1UfmzfD8uXxv0IMad+cQw2xhOfbb+Gcc9w/AH/9a2w21BAREUmJ\nspUmfMcRf/21+5oJDbGuEEvc5OfnB3/Sli1uy+Vt29x2y9UsvJ2pvHIT5eZJuflRbsEpMz/5+fnQ\npg3stZf/OOJMWIO4TJobYq1DLAkrLi4O9gRrYdAgKCiAd9+FDh1SUlfUBc5NAOXmS7n5UW7BKTM/\nP+SWm+t/hXjJEvc1U64Qr1sHGzZAkyYpfzlt3VyOtm5Ok7FjYeBAt0Vl//5hVyMiIhIdN90EL78M\nn38e/Ll33eUmqa9alfy60m3qVDj5ZJdDFatPJXPrZg2ZkPR691244Qa4/no1wyIiIjvLzYX58+H7\n74M/N1NWmICKu9WlQWQaYmPMQGPMl8aYjcaY6caYHtU8tq0x5hljzKfGmO3GmHsqecwlxpiS0u+X\nlN70OU6YFi6E8893WzLffXfY1YiIiERP165uaOEnnwR/biasQVwmzbvVRaIhNsb0A0YBQ4EjgFnA\nm8aYqrYrawgsB/4IfFzNodcCbcvd9k1WzUm3fLlbfqxrV78fghCtXLmy5gdt2OBWlNh9d3j+eahf\nP/WFRVytcpNdKDc/ys2PcgtOmfn5IbdDD4U6dfwm1mXSFeLdd3djh7OpIQZuBP5qrX3SWlsEXA0U\nAwMqe7C1dpG19kZr7dPAumqOa621K6y1y0tvK5JfeoK2bYP774eDDnKrLWzZAsceC1OmhF1ZrQ0Y\nUOlf0w7WuuERX3zhtmVu2TI9hUVcjblJpZSbH+XmR7kFp8z8/JBbo0bQubPfxLpMukKc5t3qQm+I\njTH1ge7A22X3WTfTbwrQK8HDNzXGLDTGLDbGTDTGHJrg8ZLrX/+CI4+EX/0K+vWDzz5zKy8cdxz0\n6QOPPBJ2hbUybNiw6h8wfDj84x/wxBNubJQAtchNKqXc/Cg3P8otOGXmp0JuPgeNHWoAACAASURB\nVFs4b9oEK1ZkTkMMaV16LfSGGGgF1AW+3en+b3HDHHx9irvCnAdcjHuv7xtj2idwzOT4+mu46CI4\n6ST3cUBBgduYolUr9xHBpElw5ZVwxRVwyy1QUhJ2xdWqdkWOl1+G3/0O7rjDrTssP9BKJn6Umx/l\n5ke5BafM/FTIrWtXd4U4yEpgZUuuZcqQCUjrbnVRaIhTwlo73Vr7tLX2f9bafwM/BVYAV4VW1JYt\nkJ8PXbrA22/DY4/Bf/4DbsmQHerVgzFj4N57YeRI+NnP3FaOcTNvHlx8MeTlwdChYVcjIiISD127\nuqXTggwXyKQ1iMtk2RXilcB2oM1O97cBliXrRay124CPgMoXsyunb9++5OXlVbj16tWLiRMnVnjc\n5MmTycvL2+X5AwcOZNy4cRXuKxwzhrwWLVg5ZIi78vvZZ3DppQz9/e932dVn8eLF5J19NkVnnOHG\nFb/+Opx0EqPvvJPBgwdXeGxxcTF5eXlMmzatwv3jx4+nfyXLmvXr1y+x91FYSF5e3i6TJoYOHVrx\nfaxezeK+fckDim6/3U0QKDV69Oj4vA9K/z7y8igqKqpwv96H3ofeh96H3ofeR0reR9nwwv/9r/bv\n46uvmAzk3XhjdN5HOV5/H+XGEI8fP55OnTrRrVu3H3qzGyt5r96staHfgOnAfeX+bICvgMG1eO5U\n4J5aPK4OMA+4u5rHHAnYmTNn2qRZsMDac86xFqw96SRrZ88OfowPP7S2XTtr993X2k8+SV5tSfLI\nI49UvGPbNmtPP93a5s2t/eKLcIqKgV1yk1pRbn6Umx/lFpwy81Mht+3brW3a1Nr8/Nof4M47rW3R\nIvmFhenZZ13/tG5dpd+eOXOmBSxwpE2wF43CFWKAe4ArjDG/NMYcDDwENAYeBzDGDDfGPFH+CcaY\nw40x3YCmQOvSPx9S7vu3G2NOM8Z0MsYcATwDdATSM1Nt40YYNswtn1JQAM89B++8Azk5wY/VvTv8\n97/QrBn06hW5FSgKC3faHOa22+Ctt+Dvf4cDDginqBjYJTepFeXmR7n5UW7BKTM/FXKrU8f1C0Em\n1mXSChNlyjbnSMNKE5HZutkYcy1wM26oxMfAddbaD0u/9xiwr7X25HKPL8H9VlDeImvt/qXfvwf4\nCW5i3mpgJvBba22V65gkZetma93SYjfe6CbP/eY3MGQING3qd7zy1q1zq1FMmQIPPgiXX574MZPt\nmWfg5z93459/9auwqxEREYmnq66C6dNh1qzaPf6ss1wP8sorqa0rnT77zM27mjrVLUSwk2Ru3Vwv\nkScnk7V2LDC2iu/tMujEWlvt1W1r7U3ATcmprpY+/dRtS/zmm27ZtDffdOsLJ0uzZm7Vhuuvd+OQ\nv/gC/vznCuNzQ/Xhh65Jv+QSl4OIiIj4yc11k++3bq3dZlZLlsAxx6S+rnRK4/bNEemkYm79erc8\nWm6u+21m0iR49dXkNsNl6tWDBx6Ae+6BESPcFeMorECxbJnbia5rV3joIbegtoiIiPjp2tU1w59+\nWrvHf/VV5g2Z2H13d0vDkAk1xImwFp59Fg4+2O02d/vtMHeu+9gilQ2hMW5IxksvwWuvuY8Rvt15\nGec02rzZrTFcUuJq2m238GoRERHJBGUrTdRmHHFxMXz3XWatQVwmTUuvqSH29b//uUb04ovdRxTz\n5rmGOJ3N4Nlnw3vvweLFcPTRrhlPN2vJO+ggN1zixRfdiSu1UtkSP1Iz5eZHuflRbsEpMz+75Na8\nuWtwa7OF89dfu6+ZdoUY0rY5hxrioNascWN4jzgCli+HyZNhwgTYb79w6gl7BYqxYxm0eLEbJpFp\nY5dSbNCgQWGXEEvKzY9y86PcglNmfirNrWvX2l0h/uor9zUTG2JdIY6YkhJ49FE3Lvixx9yOc7Nm\nwWmnhV0ZdOwI06bBsce6yXyPpGdlOd59F264gd433ACVLLYt1evdu3fYJcSScvOj3Pwot+CUmZ9K\nc8vNrd0V4rKGuEOH5BYVBWqII6SgwF19vewyOP10N8D9N7+BBg3CrmyHshUorrjC3W691TXxqbJw\nIZx3Hpx4Itx9d+peR0REJFt17eqa3dWrq3/ckiXQqhU0apSeutKpbLe6FC8TrIa4OitWuOby6KPd\nxLH33oOnnoruONl0rUCxYYMbv9ysGTz/vHtdERERSa6yiXVz5lT/uK++yswJdeB6rg0b3IpeKaSG\nuDLbtrnG8qCD4IUXYMwYN2nshBPCrqxmZStQvPiiW4HiRz9K7goU1sKll8L8+W4DkpYtd9mDXGpH\nuflRbn6Umx/lFpwy81Npbl26uDWIaxo2kYm71JVJ01rEaogrc/HFcN11cP75bl3ha6+N31XQc86B\nf/0LFi1yk92StQLFn//sfkl46qkffnMdP358co6dZZSbH+XmR7n5UW7BKTM/lebWoIFb2rWmiXWZ\nuAZxmbJP5VPcEEdm6+Yo+GHr5sMO48jHHoMePcIuKXGLF8OZZ7qvL7wAp57qf6xJk9xQiWHDYOjQ\npJUoIiIiVfj5z2HBAnj//aof06IFDB4Mt92WvrrSZcMGaNrUXYj7+c8rfCuZWzfrCnFlHn88M5ph\n2LECRa9eia1AMXeuOxF/8hO33rKIiIikXteubgxxVRPlN2xwk+4y9Qpxkyawxx4p361ODXFl6mRY\nLGUrUFx+ud8KFKtXuyvD++4LTzyRefmIiIhEVW6um1C2aFHl31+yxH3N1El1kJal12I2MFa81asH\nY8dC585uybj58+HJJ2teomXbNrjgAli1yi0/t/vu6alXRERE3BVicBPrOnXa9fuZvClHmTTsVqdL\nfdnEGLjpJrcCxauv1m4Filtvhbffdsur7b9/pQ/pr005vCg3P8rNj3Lzo9yCU2Z+qsytfXu3jXNV\nE+vKrhBn4qYcZdJwhVgNcTY65xy3pnJNK1A89RSMGuVup5xS5eG0K5Ef5eZHuflRbn6UW3DKzE+V\nuRnjrhJXtfTaV19B69aw226pKy5sZZtzpJBWmSjnh1UmZs7kyCOPDLuc1Fu8GH78Y/fDtPMKFAUF\nbt3lCy90W1YbE16dIiIi2ey662DKFJg3b9fvXXWV2yth5sz015Uu993nVtDYsKFCP6JVJiQ5OnaE\n//zHXSUuvwLFN9+4q8jdusGDD6oZFhERCVPXrm5fhMp2n83kNYjLtGvn3vvatSl7CTXE2a5ZM3jl\nlR0rUNxyC5x7rtuR7sUXM/sjGBERkTjo2tWtDlXZFeJM3ra5TBo251BDLDtWoBg1CkaOdB+7vPTS\njhOwBtOmTUtxgZlJuflRbn6Umx/lFpwy81Ntbocd5r5WNo44k7dtLqOGWNKmbAWKt96C116Do4+u\n9VNHjBiRwsIyl3Lzo9z8KDc/yi04Zean2tyaNoUDDti1If7+e1izJvOvELdr576mcGKd1iGWiqpZ\nTaIqzz33XAoKyXzKzY9y86Pc/Ci34JSZnxpzy83ddem1bFiDGNyeCc2b6wqxRFvjxo3DLiGWlJsf\n5eZHuflRbsEpMz815lbZ0mtlaxBnekMMKd+cQw2xiIiISNTl5sLy5RU31Cq7QlzLOT+xluLNOdQQ\ni4iIiERd2RbO5YdNLFkCbdpAw4bh1JROaogl6gYPHhx2CbGk3PwoNz/KzY9yC06Z+akxtwMOcGNp\nyzfE2bAGcZkU71anhlgS1rFjx7BLiCXl5ke5+VFufpRbcMrMT4251a3rll8rP444G9YgLlN2hThF\nOyxr6+Zysm7rZhEREYmPyy6DWbPcVs0AOTlw8slw//3h1pUOL7wA558P330HLVoA2rpZREREJPvk\n5sInn8C2be7P2XaFGFI2jlgNsYiIiEgcdO0KmzbBF1/AunXulk1jiEENsURXUVFR2CXEknLzo9z8\nKDc/yi04ZeanVrnl5rqvs2dn1xrEkPLd6tQQS8JuvvnmsEuIJeXmR7n5UW5+lFtwysxPrXJr3Rra\ntnUT68rWIM6WIRMNG0LLlrpCLNE1ZsyYsEuIJeXmR7n5UW5+lFtwysxPrXMr28J5yRIwBjp0SG1h\nUZLC3erUEEvCtMSOH+XmR7n5UW5+lFtwysxPrXMr28L5q6/c1eL69VNbWJSkcHMONcQiIiIicdG1\nK3z5Jcydmz3DJcqkcHMONcQiIiIicVE2se6tt7JnQl0ZXSGWKMvPzw+7hFhSbn6Umx/l5ke5BafM\n/NQ6t0MOcbvWrVmTfVeIy8YQp2BTOTXEkrDi4uKwS4gl5eZHuflRbn6UW3DKzE+tc9ttNzjoIPff\n2XiFeOtWt1tdkmnr5nK0dbOIiIhE3gUXwN//Ds89B/36hV1N+kyfDr16ue2ru3bV1s0iIiIiWats\nHHG2DZko260uBRPr1BCLiIiIxEnPnm4c8QEHhF1JerVt676mYGKdGmJJ2MqVK8MuIZaUmx/l5ke5\n+VFuwSkzP4FyO/VUWLBgR4OYLRo0gFat1BBLNA0YMCDsEmJJuflRbn6Umx/lFpwy8xMoN2MgWzdA\nSdHSa2qIJWHDhg0Lu4RYUm5+lJsf5eZHuQWnzPwot1pSQyxRpRU5/Cg3P8rNj3Lzo9yCU2Z+lFst\npWi3OjXEIiIiIhIPukIsIiIiIlmtXTt3hbikJKmHVUMsCRs3blzYJcSScvOj3PwoNz/KLThl5ke5\n1VL79rBtGyR5NRM1xJKwwsKENofJWsrNj3Lzo9z8KLfglJkf5VZLKdqcQ1s3l6Otm0VEREQibMkS\n2GcfeO01Ctu00dbNIiIiIpJl2rRx6zAneWKdGmIRERERiYf69aF1azXEIiIiIpLFUrD0mhpiSVhe\nXl7YJcSScvOj3PwoNz/KLThl5ke5BZCCzTnUEEvCBg0aFHYJsaTc/Cg3P8rNj3ILTpn5UW4BpOAK\nsVaZKEerTIiIiIhE3B13wKOPUjhpklaZEBEREZEs1L49LFsG27cn7ZBqiEVEREQkPtq3d83w6tVJ\nO6QaYknYxIkTwy4hlpSbH+XmR7n5UW7BKTM/yi2Ast3qkrh9sxpiSdj48ePDLiGWlJsf5eZHuflR\nbsEpMz/KLYCyhnjFiqQdUpPqytGkOhEREZGI27YNGjSgcMgQut95J2hSnYiIiIhklXr13BbOSbxC\nrIZYREREROKlfXuNIRYRERGRLNa+va4QS7T0798/7BJiSbn5UW5+lJsf5RacMvOj3ALSFWKJmt69\ne4ddQiwpNz/KzY9y86PcglNmfpRbQO3aaZWJVNEqEyIiIiIx8Le/UXj11XR3faxWmRARERGRLNO+\nPSTxoq4aYhEREREJ5L334Pzzk9qTBlO2OUeSqCGWhE2bNi3sEmJJuflRbn6Umx/lFpwy8xO33L7/\nHl54ARYtCqkANcQSNSNGjAi7hFhSbn6Umx/l5ke5BafM/MQtt5wc93XOnJAKaN0ajEna4TSprhxN\nqvNTXFxM48aNwy4jdpSbH+XmR7n5UW7BKTM/ccvNWthzT7jtNrj11nBqKGzdmu5u6TVNqpPwxekH\nOEqUmx/l5ke5+VFuwSkzP3HLzRh3lXj27BCLaN06aYdSQywiIiIigeXkhDhkAuCii5J2KDXEIiIi\nIhJYTg4UFcHWrSEV0Ldv0g6lhlgSNnjw4LBLiCXl5ke5+VFufpRbcMrMT1xy27wZiovdf+fmwpYt\n8Pnn4daUDGqIJWEdO3YMu4RYUm5+lJsf5eZHuQWnzPzEJbcpU6BZM/j6a9cQX3UVNGgQdlWJ0yoT\n5WiVCREREZGqDRsGDzwAy5cnddUzL4WFhXTv3h20yoSIiIiIpMuMGdCjR/jNcLKpIRYRERGRGlnr\nGuKePcOuJPnUEEvCioqKwi4hlpSbH+XmR7n5UW7BKTM/ccht4UL47js1xCKVuvnmm8MuIZaUmx/l\n5ke5+VFuwSkzP3HIbcYM97VHj3DrSAU1xJKwMWPGhF1CLCk3P8rNj3Lzo9yCU2Z+4pBbQQHst19S\nN4iLDDXEkrC4LBUTNcrNj3Lzo9z8KLfglJmfOORWNqEuE9ULuwARERERib4XX4Tvv9/1/m++gUWL\n4Jhj0l9TsugKsYiIiIjUqFUrN2RiZ2PGwLnnpr2cpFJDLAnLz88Pu4RYUm5+lJsf5eZHuVW0YgXM\nn1/9Y5SZnzjnlpMDS5fCqlVhV+JPDbEkrLhsU3MJRLn5UW5+lJsf5VbR119D9+6wbFnVj1FmfuKc\nW26u+zpnTrh1JEJbN5ejrZtFRESq9t13cMABcNpp8I9/hF2NRMWWLdCkCdx3H1x7bfpeV1s3i4iI\nSNLVdI2sZUt46CF44QWYODE9NUn0NWgABx8c7yvEaohFRESy3PLlkJ/vmprFi6t/bL9+cOaZMHAg\nrF2bnvok+nJyYPbssKvwp4ZYErZy5cqwS4gl5eZHuflRbn4yOTdrYepUuOAC2HtvGDoUjj4atm2r\n/nnGwNixsG4d3Hrrrt/P5MxSKe655ea6K8RxHYmrhlgSNmDAgLBLiCXl5ke5+VFufjIxt+++g3vu\ncVeDTz4ZPv7YXR3++mt48knYf/+aj7HPPnDXXW74xL//XfF7mZhZOkQ5t9/+Fu68s/rH5ORAcbFb\niSSONKmuHE2q81NYWKi8PCg3P8rNj3Lzk4m5/exnbvzvuefCVVfBiSe6q75BlZTA8cfD+vUwaxbU\nKb3ElomZpUOUc+vSBXr3htGjq37M1q3ua/366akJkjupTg1xOWqIRUQk0y1YAE2bwl57JX6sTz+F\nTZvg8MMTP5ZE05o10Lw5PPEE/PKXYVdTUTIbYm3dLCIikkVqMySitrp0Sd6xJJo+/NB97dkz3DpS\nTQ2xiIhIBli/Hp59FjZvhuuvD7sayRQFBdCsGRx0UNiVpJYm1UnCxo0bF3YJsaTc/Cg3P8rNTxxy\n++gjuPpqaN/ebYpQmNAHx4mLQ2ZRFNXcCgrgqKN2jBHPVBn+9iQdCsP+1zemlJsf5eZHufmJam4b\nNsCjj7pl0o48El5+GX79a1i4EB5/PNzaoppZ1EU1txkzoEePsKtIPU2qK0eT6kREJOoWLXKT2Nat\ng9NPd1eHf/xjqKdBkJJkS5dChw4wYQL89KdhV7MrTaoTERHJUh07wu23uwalU6ewq6lccTE0bhx2\nFZKo+vXhz3+GY48Nu5LU05AJERGRGDHGDY+IajM8YQIceGB8N2iQHVq3httug7Zta/+cm2+Ga65J\nXU2pEpmG2Bgz0BjzpTFmozFmujGmyhErxpi2xphnjDGfGmO2G2PuqeJx5xtj5pUec5Yxpk/q3oGI\niIiccIJb6eLGG8OuRMKwaRO8+27YVQQXiYbYGNMPGAUMBY4AZgFvGmNaVfGUhsBy4I/Ax1Uc81jg\nWeBhoBvwT2CiMebQ5FYveXl5YZcQS8rNj3Lzo9z8KLfgLr88j3vvhWeegddfD7ua+MiUcy0nBz7/\n3DXGcRKJhhi4EfirtfZJa20RcDVQDFS6sbe1dpG19kZr7dPAuiqOeT3wurX2Hmvtp9baO4BCYFAK\n6s9qgwYpUh/KzY9y86Pc/Ci34AYNGsQvfgGnneYm/H3/fdgVxUOmnGs5ObB9u9vFME5Cb4iNMfWB\n7sDbZfdZt/TFFKBXAofuVXqM8t5M8JhSid69e4ddQiwpNz/KzY9y8xNWbtbC8OFuln/c9O7dG2Pg\nr3+FlSvht78Nu6J4yJSf0Zwc93X27HDrCCr0hhhoBdQFvt3p/m+BAMO4d9E2BccUERFJuZEjYcgQ\nmDYt7Er8deoEf/wjjB4N06eHXY2kS7NmbiWUOXPCriSYKDTEkdO3b1/y8vIq3Hr16sXEiRMrPG7y\n5MmVjvkZOHDgLjvOFBYWkpeXx8qVKyvcP3ToUPLz8yvct3jxYvLy8igqKqpw/+jRoxk8eHCF+4qL\ni8nLy2PaTv9qjh8/nv79++9SW79+/fQ+9D70PvQ+9D4i/D4GDRrHrbe6K6s/+1l838e4ceO4/nro\n3h2uugo+/DC+76O8OP99pOt9bNqUx3vvJfd9jB8/nk6dOtGtW7cferMbkzlz01ob6g2oD2wF8na6\n/3HgpVo8fypwTyX3LwKu3+m+YcBH1RzrSMDOnDnTSu299NJLYZcQS8rNj3Lzo9z8pDu3jz6ytnFj\na887z9rt29P60kmzc2azZ1s7Y0ZIxcRIlH5Gt2yx9r77rF2yxO/5t9xibceOya2pMjNnzrSABY60\nCfajoV8httZuBWYCp5TdZ4wxpX9+P4FDf1D+mKVOK71fkmj8+PFhlxBLys2PcvOj3PykM7dvvoGz\nzoJDDoEnnoA6of8f2s/OmeXkZMfWv4mK0s/oJ5/ADTfAl1/6Pf+882DYMDcWPi4isXWzMeZnuCvC\nVwMzcKtOnAccbK1dYYwZDrS31l5S7jmHAwa3rFoRcDewxVo7r/T7vYB3gduAV4ELgVtxv0XMraIO\nbd0sIiJpt3EjnHQSLFkCM2a47XIlmt5+Gy6/HB57zP2dZaKHH3aba6xdC02ahF1N1TJu62Zr7fOl\naw7/AWiDW1v4dGtt2T43bYF9dnraR7jL5OCGOlyEGyaxf+kxPzDGXATcWXr7HDi7qmZYREQkLH/6\nk5uV/+9/qxmOsocfhmuvhW3b4OmnM7chnjEDDjss2s1wskWiIQaw1o4FxlbxvV1GYVtra/wwyVo7\nAZiQeHUiIiKpc9tt0KePm4Am0TR8uFv545proF49ePFFNyTAmLArS76Cguwb5hLTEUoiIiKZo2lT\nOP74sKuQ6vTuDfffDw88AHl58PXX8VtrtzaKi92SaT17hl1JeqkhloRVtoyK1Ey5+VFufpSbH+UW\nXG0z27w5xYUkWffucN117orwCSdAq1YwN4mDMKNyrn30kdtpTleIRQLKlN110k25+VFufpSbH+UW\nXG0ye+kl6NIFVq1KQ0Ep0LChWxXkgguSd8yonGsFBbDbbjt2nMsWkVhlIiq0yoSIiEjqLV0Khx4K\n554LO+0XISG7/343ufMf/wi7kpolc5UJXSEWERGRtGrf3m1P/eijbhkziY7rr09eMzxxInwQk90f\n1BCLiIikyZAhMGJE2FVEw2WXwYknum2di4vDrga2bnXN4MyZYVeSOf74x/h8AqCGWBK2877rUjvK\nzY9y86Pc/CQzt8cfd0t31a+ftENGUm0zq1MH/vY3txnJ73+f4qJqsHq1W/buwQfhs8/CqSETf0Zz\nctyKFXGghlgSNkKXO7woNz/KzY9y85Os3P79b7jySrfD2a9+lZRDRlaQzA46CIYOhVGjoDChEaD+\n5s+HY491r//WW3DhheHUkYk/o7m5bhvokpKwK6mZJtWVo0l1foqLi2ncuHHYZcSOcvOj3PwoNz/J\nyG3BArema24uvPkmNGiQpOIiKmhmW7e6Jb7q1XMrHKRzo4tp0+Ccc6BFC3jlFdeghyUTf0bfeMNd\neV+wADp1Sv7xNalOIiXTfoDTRbn5UW5+lJufRHNbuxbOPBOaN4cJEzK/GYbgmdWv74aT3Hdfepvh\np5+GU05xH+t/8IF/M2xtcoZZZOLPaNnSbXEYNqGGWEREJAW2bYN+/dx6ta+84q5CSuW6dYPjjkvf\n6731FvziF3DRRTB5MrRs6X+sxx5zS8itXp28+jJFhw6w555qiEVERLJWQQG8955bwqpLl7CrkfJO\nOQVeeMEt+5boVftTT3U7u731VnJqC8t338H33yf3mMa4q8Rx2OJaDbEkbPDgwWGXEEvKzY9y86Pc\n/CSSW69esHCha5iySRzOtTp13KYgyRii0bEjHHYYvP56YscJO7e773ZXupPtmGOgbt3kHzfZ6oVd\ngMRfx44dwy4hlpSbH+XmR7n5STS3vfZKUiExko3nWt++8NRTbjWFOp6XGsPOraAA3Py05Bo5MvnH\nTAWtMlGOVpkQERGRoKZOhZNPdpt6xLF9KClxEz9vucVtHhMXWmVCREREMt62bYk9f/v25NRRk+OO\ng6ZNEx82EZbPP4d169zygNlKDbGIiIhEzj//6ca0rl0b/Lnbt7sNUPr3d8uipVqDBnDaafFtiGfM\ncF+POircOsKkhlgSVlRUFHYJsaTc/Cg3P8rNT21ysxbWr09DMTGRrHPtiCPcknW33RbseevXw9ln\nw5gxbmJjutY27tPHTaLctMnv+WH+jBYUQOfObom0bKWGWBJ28803h11CLCk3P8rNj3LzU5vc/vIX\nt47uunVpKCgGknWudewIw4fDgw+6HeVqY/FiOP54t1X2q6/CNdckpZRaueQS9/q77eb3/DB/RmfM\nyO7hEqCGWJJgzJgxYZcQS8rNj3Lzo9z81JTbq6/Cr38N558PzZqlqaiIS+a5ds017irvFVfUfOW1\noACOPtoNsXj/fTj99KSVUSsNGvivMAHh/Yxu2wazZrnts7OZGmJJWNhLxcSVcvOj3PwoNz/V5TZ7\nNlxwgft4/s9/TmNREZfMc61uXXj4YZg/v/qMJ0yAE0+E/faD//7XrQscN2H9jNarB0uXuivcqWRt\n4pMkU0kNsYiISEDLl8NZZ8GBB7r1ZxO5MijVO+wwtxTY8OGV73i2fr27kpyXB++8A23apL/GuGve\nPLXjh7dscdtjP/VU6l4jUfoRFhERCWDTJjjnHNi8GV5+2S23Jal1221u0teVV+66asTuu7urws8+\nC40ahVOfVK9BA9cQR3kLZzXEkrD8/PywS4gl5eZHuflRbn52zs1auPxy+OgjmDQJ9t47pMIiLBXn\nWsOG8PjjbthEZatGdOoU/6v0mf4zmpMDc+aEXUXVtHWzJKy4uDjsEmJJuflRbn6Um5+dc9u40Q2X\neOIJTUKqSqrOtUxfBSHTf0ZzcuCRR8Kuomraurkcbd0sIiI1sTZ9a9tKPBUXu9Uu2rULu5LoeP55\n6NcPVqyAVq2Sc0xt3SwiIhISNcNSkx/9CLT0d0U5Oe5rVIdNqCEWERERSaKTT4Y334SSkrAriY7O\nnaF+fTXEksFWrlwZdgmxpNz8KDc/yq32/vc/mDvX/bdyC06ZQd++bmjAWH16nQAAIABJREFUzJm1\nf04YufXu7VbnSIf69eGQQ9QQSwYbMGBA2CXEknLzo9z8KLea/fe/bi3bww+HUaPcfcotOGXmdtfb\nYw947bXaPyfdua1YAW+9ld7VOR56CH7zm/S9XhBqiCVhw4YNC7uEWFJufpSbH+VWOWth6lQ49VQ4\n5hj4/HO3gsRDD7nvK7fglJnb/a13b3j99do/J925ffih+5rO1Tt69XKb2USRGmJJmFbk8KPc/Cg3\nP8qtImvh1VfhuOPceM/vvoN//MN9nPvLX7qPd0G5+VBmTp8+MGOGuxJbG+nObcYMt1lGp05pfdnI\nUkMsIiJZ5+uv3W5z4BrjwkI47zyoWzfcuiRznHGG+8Vr8uSwK6lcQYFbS1urpjjamENERLLO3nu7\nbWS7dFFDIKnRrh0ccUQ0J5FZ664QX3112JVEh64QS8LGjRsXdgmxpNz8KDc/ym1XBx9cczOs3IJT\nZjv85z8wfHjtHpvO3BYtckM5Mn33vyDUEEvCCgsT2hwmayk3P8rNT7bltmEDrFqV+HGyLbdkUGY7\nNGpU+8emM7eCAvdV24/voK2by9HWzSIi8bZmDYwZA3/5C1xwgftvEano88/hjTfguuvCriQxydy6\nWWOIRUQk9lasgHvvhQcegM2b4bLLYPDgsKsSiabOnd0tDMuWwd13u2Z8333DqaEyaohFRCS2lixx\n/3P929/cBgPXXAM33eQmNIlI9NSp4za+OeYYNcQiIiIJ+/JLNzGucWN3Nfj66926qiISXXvtBa1b\nu9U3zjsv7Gp20KQ6SVheXl7YJcSScvOj3PxkYm6dOsEjj7gZ87//fWqa4UzMLdWUmZ9syi03N3rL\n0ekKsSRs0KBBYZcQS8rNj3Lzk4rc/vY3t8FFnTqV33r0gJNOqvr5338PEyZU/fw6deDEE6FFi6qP\n8YtfJP1tVaDzLThlVrmlS6FNm6o3f8mm3HJy3KS+KNEqE+VolQkRyXYLF8Lbb8OsWXD//dU/Ni8P\nPvoISkoqvw0cCH/6U9XPX7QI9tuv+teYPh2OPjrouxCJlo8+giOPhPffh169wq4mfA8/7DYF+f77\nYEvT7UyrTIiISFJ8+y288467vf22G5dbp477n/e6ddCsWdXPnTQpsdfu2BG2bq3YRG/fXvHP1b2+\nSFx07eo+6Xj9dTXE4IZMlJRAUZHbzS8KNIZYRCQLLVrkPrZs2xYuugg++ADOPBMmToSVK93C/alu\nRo2BevWgQQPYbTc3OW733WGPPaB5czcmuH791NYgkg5160Lv3vDaa+HW8dlncM897spsmA491H2d\nPTvcOspTQywJmzhxYtglxJJy86Pc/OycW/v2cPzx8Oyz8M03boLL/ffD2We7ZlQcnW/BKbPK9ekD\nM2e6T2Uqk47cJk+GW28N/xfNZs3cxjl77BFuHeWpIZaEjR8/PuwSYkm5+VFu1du6Ff7zH3juuYr3\n75xb/frw0ENw4YXuKrFUTudbcMqscmec4b5WNZksHbkVFMDhh0PDhil/qRqNH+9+AY8KTaorR5Pq\nRCRuSkrg4493jAH+979hwwY3WW3BAjcsQUSioUcPOOCAXX9hTZdDDoGTT3Y7OmaCZE6q0xViEZEY\n+uQTOPdct8B99+5wxx2uOb7jDncV6Isv1AyLRE3fvm7YwrZt6X/ttWvh009dUy670ioTIiIx1KAB\nLF8O113nrvgcfXQ0PgYVkar16QP5+a4xPeyw9L72zJlgLfTsmd7XjQs1xCIiEWRt9Vd4O3d2wyNE\nJD569oRVq9yKKulWUABNm0KXLul/7TjQkAlJWP/+/cMuIZaUm59Mzs1a1+Reeimcfnpyj53JuaWS\ncgtOmVWtTp2qm+FU51ZQAEcdVfVOedlODbEkrHfv3mGXEEvKzU8m5vbNN3DXXe7Kzf/9n2uKTzzR\njQlOlkzMLR2UW3DKzE+qc9t7b/jxj1P6ErGmVSbK0SoTIpIuW7fCq6/CuHFu96r69eG882DAANcM\n19HlChHJcNu3w/z5bgiYzyRgrTIhIhJzGzfCxRfDsmUwZoy7SvzUU/CjH6kZFpHs8PLL7pOxZcvC\nrkST6kREQtGsmdtGtUOHsCsREQlHTo77OmcOtGsXbi26DiEJmzZtWtglxJJy8xOX3Kx1t+qksxmO\nS25Ro9yCU2Z+sjG3/feHRo1g9uywK1FDLEkwYsSIsEuIJeXmJ+q5ffstjBwJhx4K06eHXc0OUc8t\nqpRbcMqsdqx1/16Uycbc6tRx6zHPmRN2JZpUV4Em1fkpLi6mcRiLKsaccvMTxdy2bXMT4x59FF55\nxS1r9NOfwpAhOz4SDFsUc4sD5RacMquda6+F9993W69D9ubWv7/beXPGjODP1aQ6iZRs/AFOBuXm\nJ0q5ff453HYbdOwIeXmwaBH85S+wdCk8+2x0mmGIVm5xotyCU2a1c/zxMGuW+/cCsje33FzXECdz\nmUkfaohFRDyNHQsPPeSuBhcWutvAgdCiRdiViUjUnX66W2rsjTdS+zoLFsD69al9jUTk5EBxMXz5\nZbh1qCEWEfF0++3u6s6YMXDEEWFXIyJx0rIlHH00vPZaal/nkkvg8stT+xqJKPskLeyJdWqIJWGD\nBw8Ou4RYUm5+Up3b1q3w1Vfw3//C//5X/WNbtHAzpONA55sf5RacMqu9Pn3grbfcvzupyG3bNvfJ\nVc+eST900rRrB1984YadhUnrEEvCOnbsGHYJsaTc/CQzt2eegXffdVd5y27Ll+/4/nnnwT/+kbSX\nC5XONz/KLThlVnt9+8LQoW5yXSpymzvXDUfo0SPph04aY+CAA8KuQqtMVKBVJkTip6QEvvuuYlO7\ndKnb+e0vf4EGDap+7i23wDvvQPv2u946dIC999Z4YBFJnZISd4W0f3+4667kH3/cOLjySli7Fpo2\nTf7xw5bMVSZ0hVhEYunDD+Hcc13ju3XrjvuNgb32ck3t2rXQunXVx8jPT32dIiJVqVMHzjgDCgpS\nc/wZM9ya6JnYDCebGmIRiYytW90/4G+95SapnX121Y9t1w5+/vNdr+q2aQP166evZhGRRIwenbqG\ntaAg2sMlokST6iRhRUVFYZcQS8rN7dQ0dy7cfz+cdZYbnnD88e7Pn31W+XPKcuvQAe680y1z9pOf\nuNnae++tZrgqOt/8KLfglFkwzZq5K8XJzm3jRjcxOMoT6qJEDbEk7Oabbw67hFjK9txGj3YN7GGH\nweDBbuLHkCHuCvGKFe6+ymR7br6Umx/lFpwy85Ps3ObNg+3bdYW4tjSprhxNqvOzePFizSr2UF1u\nJSXuikEmmzDBLW126qnuqnBtN2nS+eZHuflRbsEpMz+pyG3NGjcco16GDpDV1s0SKfqHz09luc2d\nC2eeCcOHV//cRYvg5Zfd2o3bt6eoQE9bt7olhBYurP5x554LI0ZA7961b4ZB55sv5eZHuQWnzPyk\nIrc994xPM/zSS3DBBeG9vhpikQhYvhyuuQa6dnUfc5Xt3FOV115zi5h37gxNmrjn9esHw4bB3/+e\n3h1/rHU133+/q6llSzjuOHj66fTVICIi8bZ+vfv/V1jbTMfk9waRzLRxo1srd/hwqFvXXTEdOBAa\nNqz+eVdf7a4kz5tX8TZ1qht/26mT278+lV5+GV54AaZMcev+NmjgGuFbb4XTTgONOhIRkdoquxA0\nd66bJJ1uukIsCcvXYq6BlZTABRfk06UL3HEHDBjghj/cdFPNzTC4tXb32ccNN7jhBnjoIfjXv9yV\n5pUr4ZVXaj7GJZfAddfB2LGukV62zF3tra1XX3UzmC+6CN58E1avdptcDBniJnHUrVv7YwWh882P\ncvOj3IJTZn7y8/PZuNFNMM5Ghxzi5s7MmRPO6+sKsSSsOFt/ehNgLUydWsxxx7nNITp3Tt6xW7Z0\nt5pef9Mmt0blQw+5/e7BjTc75BB3GzTIrQVclbFjw5n4p/PNj3Lzo9yCU2Z+1q4tZq+9YNQot7tc\ntmnUCA48MLyGWKtMlKNVJiSd1q+H3XcPuwo3CW7+/F2HX4waBSeeGHZ1IiLZ44QToFUrN8EsG517\nrtthdMqU2j1eWzeLZIAoNMPgNrI4+GB3+8lPwq5GRCR79enj5pRs2eLmZWSbnBz461/DeW2NIRYR\nERGJgD594PvvYdo0/2MsXeqG4RUUJK+udMnNhW+/dZPD000NsSRs5cqVYZcQKdbC889D375uOEJV\nlJsf5eZHuflRbsEpMz8rV66kWzdo184trelrxgw3Sbtdu+TVli49esAtt7iJ5+mmhlgSNmDAgLBL\niIwPPnBLj/Xr51ZZWLOm6scqNz/KzY9y86PcglNmfgYMGIAxcMYZ8Prr/scpKIC2baFDh+TVli77\n7gt33QVt2qT/tdUQS8KGDRsWdgmh+/JL1wQfe6xbW3jKFLdOb+vWVT9HuflRbn6Umx/lFpwy81OW\nW9++bi3eRYv8jlNQAD17uuU5pfbUEEvCsnlFjjVrYPBgNyFt2jR47DH48EM45ZSan5vNuSVCuflR\nbn6UW3DKzE9Zbqee6rZb/uCD4Mew1jXEPXokubgsoFUmRBLw2GPw4IPwu9+5TTWaNAm7IhERibM9\n93QbJdW0nnxlvvjCXajp2TP5dWU6NcQiCbj2WrjggnhOXhARkWjyaYbBTagDOOqo5NWSLTRkQhI2\nbty4sEsITcOG/s1wNueWCOXmR7n5UW7BKTM/ycitoMDt9taiRRIKyjJqiCVhhYUJbQ6TtZSbH+Xm\nR7n5UW7BKTM/ycjtkkvg3nuTUEwW0tbN5WjrZilv3Tq3/EufPm47TREREUm9zz5zm3Mcd1z1j0vm\n1s26Qiyykw0b3ES5zp3db9pFRWFXJCIikj1GjYKBA9P7mmqIRXA7yr32Glx8Mey1l5ssd/rp7rfU\nK64IuzoREZHskZMD8+bBtm3pe001xJL1pk51O/r8+Mfw8cfw29/CggXw5JOwzz5hVyciItls7dqw\nK0i/3FzYsgU+/zx9r6mGWBKWl5cXdgkJOeQQNxHho49gzhwYMgQ6dUr968Y9t7AoNz/KzY9yC06Z\n+akst3Hj3AWbTZtCKChEhx3mvs6Zk77XVEMsCRs0aFDYJSSkbVsYORK6dUvvVpdxzy0sys2PcvOj\n3IJTZn4qy61nTzev5b33QigoRK1bQ5s2MHt2+l5Tq0yUo1UmMsvatfDii1BYCKNHh12NiIhIMNa6\noXvnn599y6mddhrsvrv7/3hVtMqESBU2b4aXXoLzznO/XV52GXzyCWzcGHZlIiIiwRgDffu6Sd/V\nefVVGDMmPTWlS05Oeq8QqyGW2CspgXffdatBtG0LP/2pmxR3553w1VfwzjvQqFHYVYqIiATXp49b\n8WjBgqof89RTMH58+mpKh5wcWLPGTa5LBzXEkrCJEyeG+vobN8KZZ8Lbb7t1C+fOdcMkfv1rNxkh\nqsLOLa6Umx/l5ke5BafM/FSV2ymnQP368PrrVT+3oMCNN84kl14Ky5dDgwbpeT01xJKw8SH/Wtqk\niVsubf58+NOf3KoRcRB2bnGl3PwoNz/KLThl5qeq3Jo1g+OPr3rYxMqV7upxjx4pLC4Edeumd6K7\nJtWVo0l10bRyJbRoAXX065uIiGShkSNh+HB3xbRevYrfe+MNN6zi88/hwAPDqS8smlQnGa+42I2H\nOvNMaNcO3n8/7IpERETCcdVVsGTJrs0wwIwZ0Lw5HHBA+uvKJGqIJVKWLnWT4/baCy66CFatckvN\nxGUYhIiISLI1awaNG1f+vYICN1wincMLMlFkGmJjzEBjzJfGmI3GmOnGmGpHwxhjTjLGzDTGbDLG\nfGaMuWSn719ijCkxxmwv/VpijClO7bsQX1u2wIgR0KUL/POfcMstbkzw++/DoEHQsmXYFYqIiESL\ntZk5oS4MkWiIjTH9gFHAUOAIYBbwpjGmVRWP3w94BXgbOBy4D3jEGHPaTg9dC7Qtd9s3BeVnvf79\n+yd8jPx8t2XygAFueZnbb4f9909CcRGWjNyykXLzo9z8KLfglJkfn9w2bXKrUJx8cgoKyjKVjEYJ\nxY3AX621TwIYY64GfgwMAEZU8vhrgAXW2ptL//ypMeb40uO8Ve5x1lq7InVlC0Dv3r0TPsZ118E5\n50BubhIKiolk5JaNlJsf5eZHuQWnzPz45NaoETzzTAqKyUKhrzJhjKkPFAPnWmsnlbv/cWAPa+1P\nKnnOv4CZ1tqbyt13KXCvtbZ56Z8vAR4GluKuhBcCQ6y1c6upRatMiIiIiETERRdB+/Zw9927fi/T\nVploBdQFvt3p/m9xwxwq07aKxzczxjQs/fOnuCvMecDFuPf6vjGmfTKKFhEREZHUstatpJFqUWiI\nU8JaO91a+7S19n/W2n8DPwVWAFeFXFpW+uQTN0lORERE/FkLGzaEXUX65OTA7NnufadSFBrilcB2\noM1O97cBllXxnGVVPH6dtXZzZU+w1m4DPgJqXLa6b9++5OXlVbj16tVrl20VJ0+eTF5e3i7PHzhw\nIOPGjatwX2FhIXl5eaxcubLC/UOHDiU/P7/CfYsXLyYvL4+ioqIK948ePZrBgwdXuK+4uJi8vDym\nTZtW4f7x48dXOkC/X79+SX8f06ZNq/J99OmTxy9/WcThh7uJc1F+H5Dev49p06ZlxPuA9P59TJgw\nISPeR7r/Pu65556MeB/p/vv41a9+lRHvI51/H9OmTcuI9wHp/fuYMmVKrd7HccfBrbdG930k++8j\nJwfWrIGxY8fTqVMnunXr9kNvduONN+5yPG/W2tBvwHTgvnJ/NsBXwOAqHn8XMGun+54FXqvmNeoA\n84C7q3nMkYCdOXOmldo766yzdrlv+3Zrx42zdq+9rG3SxNq77rJ206YQiouwynKTmik3P8rNj3IL\nTpn5qW1uV19t7f77W1tSkuKCImL+fGvB2jfe2PV7M2fOtIAFjrQJ9qJRuEIMcA9whTHml8aYg4GH\ngMbA4wDGmOHGmCfKPf4hYH9jTL4xposx5lrgvNLjUPqc240xpxljOhljjgCeAToCj6TnLWWP5557\nrsKfZ8yAXr3gssvg1FPh00/dusING1ZxgCy1c25SO8rNj3Lzo9yCU2Z+aptb376wYIHbqjkb7Lef\n25Rk9uzUvk4kGmJr7fPAb4A/4IY1dAVOtzuWTGsL7FPu8Qtxy7KdCnyMW27tMmvtlHKHbQ78DZgL\nvAo0BXpZayte75eENS7dPsdauPJKOPpo2LwZ3nvPLQfToUPIBUZU46q2HZJqKTc/ys2PcgtOmfmp\nbW4nnwwNGsAf/gDr1qW4qAioUwcOOwzmzEnt60RlHWKstWOBsVV8b5dBJ9ba94Du1RzvJuCmqr4v\nyWcM7LknPPCAa4wr23NdRERE/DVp4j6FfeYZdwHquuvCrij1cnNh1qzUvoZaFkmqEZVtoyIiIiJJ\nc+ih8K9/QbZsmXDFFbAixdusRWLIhMTbzjNMpXaUmx/l5ke5+VFuwSkzP0Fy+8MfYMgQd4U4Gxxz\nDJx1VmpfQ1eIJRBr3dCI8jp27BhOMTGn3PwoNz/KzY9yC06Z+QmSW6tWcOedKSwmC4W+dXOUaOvm\nqlkLr7zifiN9+WU361NEREQkLJm2dbNE3GefuWVe8vLcfuL6HUpEREQyiYZMVGLSJPj4Yygpcbft\n293X5s3hoouqf+4DD8Dy5Tues/Pt9NPdrSrLl7uJaS1auFvz5rv+9x577DpsIRXWr4c//Qnuvdct\nnfbSS3D22el5bREREZF0UUNcid//fsd/16mz43booTU3xBMmwBdfVHxe+dsBB1TfEK9a5YYkrFrl\nbiUluz5m3jw4+OCqjzF9utsMo7KGujabY1gLzz4LgwfD6tXwu9+5/27UqPLHFxUVcXB1BUmllJsf\n5eZHuflRbsEpMz/KLVwaQ1xO2RjiGTNmctRRR4Z+JdRad5W2rDlevdp97dvXrUNYlcGD4e67K/9e\n48Zw0knw6qtVP7+wELp3h3PPhVGjYN99q68zLy+PSZMm1fh+pCLl5ke5+VFufpRbcMrMj3ILLplj\niNUQl5NJk+o2b3YNdFkTXb6hbt4cLrmk+ufPmQM5ObV7rcWLF2tWsQfl5ke5+VFufpRbcMrMj3Kr\nnrXwyCPQrRv06OHuU0OcIpnUEIuIiIhkkn33hQsvhLvucn/WKhMiIiIiklVyctwn2KmghlhERERE\nIi83Vw2xRFh+fn7YJcSScvOj3PwoNz/KLThl5ke51SwnBxYtgnXrkn9sNcSSsOLi4rBLiCXl5ke5\n+VFufpRbcMrMj3KrWdlk/08+Sf6xNamuHE2qExEREYmmTZugaVN48EG44gpNqhMRERGRLLPbbtC5\nM8yenfxjqyEWERERkVg4+eSqd85NhBpiSdjKlSvDLiGWlJsf5eZHuflRbsEpMz/KrXYeeABSMf9Q\nDbEkbMCAAWGXEEvKzY9y86Pc/Ci34JSZH+UWLjXEkrBhw4aFXUIsKTc/ys2PcvOj3IJTZn6UW7i0\nykQ5WmVCREREJB60yoSIiIiISJKoIRYRERGRrKaGWBI2bty4sEuIJeXmR7n5UW5+lFtwysyPcguX\nGmJJWGFhQsN2spZy86Pc/Cg3P8otOGXmR7kFs2ULbN2avONpUl05mlQnIiIiEm3ffgt77w0jRxZy\n442aVCciIiIiWWavvaBJE/jii+QdUw2xiIiIiMSGMZCbq4ZYRERERLJYTg7Mn5+846khloTl5eWF\nXUIsKTc/ys2PcvOj3IJTZn6UWzA5ObBwYfKOp4ZYEjZo0KCwS4gl5eZHuflRbn6UW3DKzI9yCyY3\nF7ZtS97xtMpEOVplQkRERCT6vvsOWrUqBLTKhIiIiIhkoZYtoVWr5B1PDbGIiIiIxM6oUck7lhpi\nSdjEiRPDLiGWlJsf5eZHuflRbsEpMz/KLbicnOQdSw2xJCw/Pz/sEmJJuflRbn6Umx/lFpwy86Pc\nwqWGWBLWunXrsEuIJeXmR7n5UW5+lFtwysyPcguXGmIRERERyWpqiEVEREQkq6khFhEREZGsVi/s\nAiJmN4B58+aFXUeszJgxg8LChNbDzkrKzY9y86Pc/Ci34JSZH+UWXLl+bbdEj6Wd6soxxlwEPBN2\nHSIiIiJSaxdba59N5ABqiMsxxrQETgcWApvCrUZEREREqrEbsB/wprX2u0QOpIZYRERERLKaJtWJ\niIiISFZTQywiIiIiWU0NsYiIiIhkNTXEpYwxA40xXxpjNhpjphtjeoRdU5QZY4YaY0p2us0Nu66o\nMcacYIyZZIz5ujSjvEoe8wdjzFJjTLEx5i1jzIFh1BolNeVmjHmskvPvtbDqjQJjzG3GmBnGmHXG\nmG+NMS8ZYw6q5HE638qpTW4633ZljLnaGDPLGLO29Pa+MeaMnR6jc62cmjLTeRYuNcSAMaYfMAoY\nChwBzALeNMa0CrWw6JsDtAHalt6OD7ecSGoCfAxcC+wyg9UYcwswCLgS6AlswJ17DdJZZARVm1up\n16l4/l2YntIi6wRgNHA0cCpQH5hsjGlU9gCdb5WqMbdSOt8q+gq4BTgS6A68A/zTGHMI6FyrQrWZ\nldJ5FhKtMgEYY6YD/7XW3lD6Z4M7ce+31o4ItbiIMsYMBc621h4Zdi1xYYwpAc6x1k4qd99SYKS1\n9t7SPzcDvgUusdY+H06l0VJFbo8Be1hrfxpeZdFW+gv9cuD/rLXTSu/T+VaDKnLT+VYLxpjvgN9Y\nax/TuVY7O2Wm8yxEWX+F2BhTH/eb2ttl91n3W8IUoFdYdcVE59KPtOcbY542xuwTdkFxYozphLsC\nUP7cWwf8F517tXFS6UfcRcaYscaYFmEXFDF74q6urwKdbwFUyK0cnW9VMMbUMcZcADQG3te5VrOd\nMyv3LZ1nIdHWzdAKqIv7zbW8b4Eu6S8nNqYDlwKfAu2AYcB7xpgca+2GEOuKk7a4//FWdu61TX85\nsfI6MAH4EjgAGA68ZozpZfWxV9mnXH8Bpllry8b263yrQRW5gc63ShljcoAPcJsjrAd+Yq391BjT\nC51rlaoqs9Jv6zwLkRpi8WKtfbPcH+cYY2YAi4CfAY+FU5Vki50+cv3EGDMbmA+cBEwNpahoGQsc\nChwXdiExU2luOt+qVAQcDuwBnAc8aYz5v3BLirxKM7PWFuk8C1fWD5kAVgLbcYPYy2sDLEt/OfFk\nrV0LfAZk9SzigJYBBp17CbPWfon7Wc76888YMwboC5xkrf2m3Ld0vlWjmtx2ofPNsdZus9YusNZ+\nZK39LW5C+g3oXKtSNZlV9lidZ2mU9Q2xtXYrMBM4pey+0o/NTqHiuB6phjGmKe6Httr/kcgOpf/Y\nLaPiudcMN9td514Axpi9gZZk+flX2tSdDfzIWru4/Pd0vlWtutyqeLzOt8rVARrqXAukDtCwsm/o\nPEsvDZlw7gEeN8bMBGYAN+IGuj8eZlFRZowZCbyMGybRAfg9sBUYH2ZdUWOMaYL7RcGU3rW/MeZw\nYJW19ivceMXfGWO+ABYCfwSWAP8ModzIqC630ttQ3Fi7ZaWPy8d9QvHmrkfLDsaYsbglmvKADcaY\nsqtza621m0r/W+fbTmrKrfRc1Pm2E2PMn3FjXhcDuwMXAycCvUsfonNtJ9VlpvMsAqy1urmx6tfi\nfmg34ga8HxV2TVG+4RrfJaV5LQaeBTqFXVfUbrh/7Epww3LK3x4t95hhwFKgGPcP34Fh1x32rbrc\ncJNR3sD9T2MTsAB4EGgddt0hZ1ZZXtuBX+70OJ1vAXLT+VZlbo+UZrGxNJvJwMk7PUbnWi0z03kW\n/k3rEIuIiIhIVsv6McQiIiIikt3UEIuIiIhIVlNDLCIiIiJZTQ2xiIiIiGQ1NcQiIiIiktXUEIuI\niIhIVlNDLCIiIiJZTQ2xiIiIiGQ1NcQiIiIiktXUEIuIZChjzFRjzD3VfP8SY8yqdNYkIhJFaohF\nRLKAMeZLY8z1O939HHBQGPWIiERJvbALEBGRcFhrNwObw65DRCSYFOCkAAABdklEQVRsukIsIpIB\njDGNjTFPGmPWG2O+NsbcVO57U4F9gXuNMSXGmO2l919qjFkdVs0iIlGhhlhEJDPcDZwAnAX0Bk4C\njiz93k+AJcDtQFugXen9tvQmIpLVNGRCRCTmjDFNgAHARdbad0vvuwTXBGOtXVN6Vfh7a+3y0AoV\nEYkoXSEWEYm/A4D6wIyyO6y1q4FPQ6tIRCRG1BCLiIiISFZTQywiEn/zgW3A0WV3GGOaU3FJtS1A\n3TTXJSISC2qIRURizlq7ARgHjDTG/MgYkwM8Bmwv97CFwP8ZY9obY1qGUKaISGRpUp2ISGYYDDQB\nJgHrgVFAs3LfvwN4CHc1uQG6Wiwi8gNjrVbcEREREZHspSETIiIiIpLV1BCLiIiISFZTQywiIiIi\nWU0NsYiIiIhkNTXEIiIiIpLV1BCLiIiISFZTQywiIiIiWU0NsYiIiIhkNTXEIiIiIpLV1BCLiIiI\nSFZTQywiIiIiWU0NsYiIiIhktf8HXv9LFVCj68QAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# partial dependence plots are a powerful machine learning interpretation tool\n", "# to calculate partial dependence across the domain a variable\n", "# hold column of interest at constant value\n", "# find the mean prediction of the model with this column constant\n", "# repeat for multiple values of the variable of interest\n", "# h2o has a built-in function for partial dependence as well\n", "par_dep_dti1 = gbm_model.partial_plot(data=train, cols=['dti'], server=True, plot=True)" ] }, { "cell_type": "code", "execution_count": 18, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Are you sure you want to shutdown the H2O instance running at http://127.0.0.1:54321 (Y/N)? y\n", "H2O session _sid_ae78 closed.\n" ] } ], "source": [ "# shutdown h2o ... be careful this can erase your work\n", "h2o.cluster().shutdown(prompt=True)" ] } ], "metadata": { "anaconda-cloud": {}, "kernelspec": { "display_name": "Python [conda root]", "language": "python", "name": "conda-root-py" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.5.2" } }, "nbformat": 4, "nbformat_minor": 2 }