{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Combine Claims Edges with Quilifier Edges\n",
    "\n",
    "This notebook illustrates how to combine claims edges with qualifier edges to prepare Wikidata-like data for export to SPARQL.\n",
    "\n",
    "Parameters are set up in the first cell so that we can run this notebook in batch mode. Example invocation command:\n",
    "\n",
    "```\n",
    "papermill partition-wikidata.ipynb partition-wikidata.out.ipynb \\\n",
    "-p claims_input_path /data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/claims.tsv.gz \\\n",
    "-p qualifiers_input_path /data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/qualifiers.tsv.gz \\\n",
    "-p output_path /data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/claims-with-qualifiers.tsv.gz \\\n",
    "```\n",
    "\n",
    "The output file contains the claims and qualifiers edges interleaved.\n",
    "Each claim edge will by followed by matching qualifiers edges.\n",
    "Qualifier edges that do not match a claim edge will be omitted from the output.\n",
    "\n",
    "Here are some contraints on the contents of the input file:\n",
    "- Each input file starts with a KGTK header record.\n",
    "  - The `id`, `node1`, `label`, and `node2` columns are required.\n",
    "  - Any additional columns in either input file will be be passed on to the output file.\n",
    "- The `id` column in the claims file must contain a nonempty value.\n",
    "- The `node1` column in the qualifiers file must contain a nonempty value that matches an `id` clolumn valie in the claims file.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Parameters for invoking the notebook\n",
    "\n",
    "| Parameter | Description | Default |\n",
    "| --------- | ----------- | ------- |\n",
    "| `claims_input_path` | A file containing the Wikidata-like claim edges to combine. | '/data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/claims.tsv.gz' |\n",
    "| `qualifiers_input_path` | A file containing the Wikidata-like qualifier edges to combine. | '/data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/qualifiers.tsv.gz' |\n",
    "| `output_path` |         A file to receive the combined output. | '/data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/claims-with-qualifiers.tsv.gz' |\n",
    "| `temp_folder_path` |    A folder that may be used for temporary files. | '/data3/rogers/tmp' |\n",
    "| `gzip_command` |        The compression command for sorting. | 'pigz' |\n",
    "| `kgtk_command` |        The kgtk commmand and any common options. | 'time kgtk --debug --timing' |\n",
    "| `kgtk_extension` |      The file extension for generated KGTK files. Appending `.gz` implies gzip compression. | 'tsv.gz' |\n",
    "| `presorted_claims` |    When True, the claims input file is already sorted in the `id` column. | 'False' |\n",
    "| `presorted_qualifiers` | When True, the qualifiers input file is already sorted in the `node1` column. | 'False' |\n",
    "| `sort_extras` |         Extra parameters for the sort program.  The default specifies a path for temporary files. Other useful parameters include '--buffer-size' and '--parallel'. | '--parallel 24 --buffer-size 30% --temporary-directory ' + temp_folder_path |\n",
    "| `use_mgzip` |           When True, use the mgzip program where appropriate for faster compression. | 'True' |\n",
    "| `verbose` |             When True, produce additional feedback messages. | 'True' |\n",
    "| `cleanup` |             When True, remove temporary files at the end of processing. | 'True' |\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "tags": [
     "parameters"
    ]
   },
   "outputs": [],
   "source": [
    "# Parameters\n",
    "claims_input_path = '/data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/claims.tsv.gz'\n",
    "qualifiers_input_path = '/data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/qualifiers.tsv.gz'\n",
    "output_path = '/data3/rogers/kgtk/gd/kgtk/cache/datasets/wikidata-20200803/data/claims-with-qualifiers.tsv.gz'\n",
    "temp_folder_path =     '/data3/rogers/tmp'\n",
    "gzip_command =         'pigz'\n",
    "kgtk_command =         'time kgtk --debug --timing'\n",
    "kgtk_extension =       'tsv.gz'\n",
    "presorted_claims =     'False'\n",
    "presorted_qualifiers = 'False'\n",
    "sort_extras =          '--parallel 24 --buffer-size 30% --temporary-directory ' + temp_folder_path\n",
    "use_mgzip =            'True'\n",
    "verbose =              'True'\n",
    "cleanup =              'True'\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sort the Claims Input Data Unless Presorted\n",
    "Sort the claims input data file by id.\n",
    "This may take a while."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if presorted_claims.lower() == \"true\": \n",
    "    print('Using a presorted claims input file.')\n",
    "    sorted_claims_path = claims_input_path \n",
    "else: \n",
    "    print('Sorting the claims input file.')\n",
    "    sorted_claims_path = temp_folder_path + 'claims.sorted-by-id.' + kgtk_extension \n",
    "    !{kgtk_command} sort2 --verbose={verbose} --gzip-command={gzip_command} --use-mgzip={use_mgzip} \\\n",
    " --input-file {claims_input_path} \\\n",
    " --output-file {sorted_claims_path} \\\n",
    " --columns     id\\\n",
    " --extra       \"{sort_extras}\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sort the Qualifiers Input Data Unless Presorted\n",
    "Sort the qualifiers input data file by node1.\n",
    "This may take a while."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "if presorted_qualifiers.lower() == \"true\": \n",
    "    print('Using a presorted qualifiers input file.')\n",
    "    sorted_qualifiers_path = qualifiers_input_path \n",
    "else: \n",
    "    print('Sorting the qualifiers input file.')\n",
    "    sorted_qualifiers_path = temp_folder_path + 'qualifiers.sorted-by-node1.' + kgtk_extension \n",
    "    !{kgtk_command} sort2 --verbose={verbose} --gzip-command={gzip_command} --use-mgzip={use_mgzip} \\\n",
    " --input-file {qualifiers_input_path} \\\n",
    " --output-file {sorted_qualifiers_path} \\\n",
    " --columns     node1\\\n",
    " --extra       \"{sort_extras}\""
   ]
  },
  {  
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Combine the Claims and Qualifiers Edges\n",
    "Combine the claims and qualifiers edges.  Qualifier edges will immediately follow the claim edge they match.\n",
    "Qualifier edges that do not match a claim edge will be omitted.\n",
    "This may take a while."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "!{kgtk_command} ifexists --verbose={verbose} --use-mgzip={use_mgzip} --presorted --left-join --join-output \\\n",
    " --input-file {sorted_claims_path} \\\n",
    " --input-key id \\\n",
    " --filter-file {sorted_qualifiers_path} \\\n",
    " --filter-key node1 \\\n",
    " --output-file {output_path}"
   ]
  },
  {  
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Clean Up If Needed\n",
    "Remove any temporary files.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "if presorted_claims.lower() != \"true\": \n",
    "    print('Removing the sorted claims temporary file: %s' % sorted_qualifiers_path)\n",
    "    os.remove(sorted_claims_path) \n",
    "if presorted_qualifiers.lower() != \"true\": \n",
    "    print('Removing the sorted qualifiers temporary file: %s' % sorted_qualifiers_path)\n",
    "    os.remove(sorted_qualifiers_path)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}