{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train / validation / test split\n",
    "\n",
    "This is a notebook that was used to split the CC dataset into train/valid/test samples for the reference annotation NLP task."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sec_certs.dataset import CCDataset\n",
    "from sec_certs.sample import CCCertificate\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "dset = CCDataset.from_web()\n",
    "df = dset.to_pandas()\n",
    "reference_rich_certs = {x.dgst for x in dset if (x.heuristics.st_references.directly_referencing and x.state.st_txt_path) or (x.heuristics.report_references.directly_referencing and x.state.report_txt_path)}\n",
    "df = df.loc[df.index.isin(reference_rich_certs)]\n",
    "\n",
    "# The following certs go straight to the test set as they represent super rare categories that we cannot split\n",
    "certs_from_rare_categories = df.loc[df.category.isin({\"Multi-Function Devices\", \"Mobility\", \"Data Protection\"})].index.tolist()\n",
    "df = df.loc[~df.index.isin(certs_from_rare_categories)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This splits 30/20/50 (train, valid, test)\n",
    "x_train, x_test = train_test_split(df.index, test_size=0.5, shuffle=True, stratify=df.category)\n",
    "x_train, x_valid = train_test_split(x_train, test_size=0.4, shuffle=True, stratify=df.loc[df.index.isin(x_train)].category)\n",
    "x_test = list(x_test) + list(certs_from_rare_categories)\n",
    "\n",
    "with open(\"../../../data/reference_annotations_split/train.json\", \"w\") as handle:\n",
    "    json.dump(x_train.tolist(), handle, indent=4)\n",
    "\n",
    "with open(\"../../../data/reference_annotations_split/valid.json\", \"w\") as handle:\n",
    "    json.dump(x_valid.tolist(), handle, indent=4)\n",
    "\n",
    "with open(\"../../../data/reference_annotations_split/test.json\", \"w\") as handle:\n",
    "    json.dump(x_test, handle, indent=4)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.8.13"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "a5b8c5b127d2cfe5bc3a1c933e197485eb9eba25154c3661362401503b4ef9d4"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}