{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "_dev = pd.read_json(\"EN_dev_Anon.jsonl\", lines=True)\n",
    "_test = pd.read_json(\"EN_test_Anon.jsonl\", lines=True)\n",
    "_train = pd.read_json(\"EN_train_Anon.jsonl\", lines=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process(df):\n",
    "    df = df.copy()\n",
    "    df.columns = df.columns.str.lower()\n",
    "    df[\"violated\"] = df.violated_articles.str.len() != 0\n",
    "    return df\n",
    "\n",
    "dev = process(_dev)\n",
    "test = process(_test)\n",
    "train = process(_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "d = {k: list(v) for k, v in train.groupby(\"violated\").indices.items()}\n",
    "dist = dict(dev.violated.value_counts().items())\n",
    "d_train = {k: v[0:dist[k]] for k, v in d.items()}\n",
    "d_remaining = {k: v[dist[k]:] for k, v in d.items()}\n",
    "\n",
    "new_rows = []\n",
    "for i in range(len(dev[\"violated\"])):\n",
    "    label = i % 2 == 0\n",
    "    new_rows.append(train.iloc[d_train[label].pop()])\n",
    "\n",
    "new_train = pd.concat([pd.DataFrame(new_rows), pd.DataFrame(train.iloc[i] for l in d_remaining.values() for i in l).sample(frac=1, random_state=42)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "new_train.to_json(\"train.jsonl\", lines=True, orient=\"records\")\n",
    "test.to_json(\"test.jsonl\", lines=True, orient=\"records\")\n",
    "dev.to_json(\"dev.jsonl\", lines=True, orient=\"records\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_json(\"data/train.jsonl\", lines=True, orient=\"records\")\n",
    "test = pd.read_json(\"data/test.jsonl\", lines=True, orient=\"records\")\n",
    "dev = pd.read_json(\"data/dev.jsonl\", lines=True, orient=\"records\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "\n",
    "number = re.compile(\"^([0-9]+|CARDINAL)\\s?\\. \")\n",
    "train[\"text\"] = train[\"text\"].map(lambda r: [re.sub(number, \"\", line) for line in r])\n",
    "test[\"text\"] = test[\"text\"].map(lambda r: [re.sub(number, \"\", line) for line in r])\n",
    "dev[\"text\"] = dev[\"text\"].map(lambda r: [re.sub(number, \"\", line) for line in r])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "train.to_json(\"data/train.jsonl\", lines=True, orient=\"records\")\n",
    "test.to_json(\"data/test.jsonl\", lines=True, orient=\"records\")\n",
    "dev.to_json(\"data/dev.jsonl\", lines=True, orient=\"records\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.10.5 64-bit",
   "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": "3.10.5"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "e7370f93d1d0cde622a1f8e1c04877d8463912d04d973331ad4851f04de6915a"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}