{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "5d0842c8-c292-41ce-a27b-73986bf43e1c",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\obses\\AppData\\Local\\Programs\\Python\\Python310\\lib\\site-packages\\torchtext\\vocab\\__init__.py:4: UserWarning: \n",
      "/!\\ IMPORTANT WARNING ABOUT TORCHTEXT STATUS /!\\ \n",
      "Torchtext is deprecated and the last released version will be 0.18 (this one). You can silence this warning by calling the following at the beginnign of your scripts: `import torchtext; torchtext.disable_torchtext_deprecation_warning()`\n",
      "  warnings.warn(torchtext._TORCHTEXT_DEPRECATION_MSG)\n",
      "C:\\Users\\obses\\AppData\\Local\\Programs\\Python\\Python310\\lib\\site-packages\\torchtext\\utils.py:4: UserWarning: \n",
      "/!\\ IMPORTANT WARNING ABOUT TORCHTEXT STATUS /!\\ \n",
      "Torchtext is deprecated and the last released version will be 0.18 (this one). You can silence this warning by calling the following at the beginnign of your scripts: `import torchtext; torchtext.disable_torchtext_deprecation_warning()`\n",
      "  warnings.warn(torchtext._TORCHTEXT_DEPRECATION_MSG)\n"
     ]
    }
   ],
   "source": [
    "from collections import Counter\n",
    "import torch\n",
    "from datasets import load_dataset\n",
    "from torchtext.vocab import vocab\n",
    "from tqdm.notebook import tqdm\n",
    "import pandas as pd\n",
    "from nltk.tokenize import word_tokenize\n",
    "import string\n",
    "\n",
    "from transformers import pipeline, AutoModelForTokenClassification, AutoTokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "4ccf08e2-fec1-4d68-a1fd-d33af6bd54bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv('train.tsv', sep='\\t', header=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c027b2f3-48eb-442a-a212-c5fa9d1cfba5",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = pd.read_csv('../dev-0/in.tsv', sep='\\t', header=None)\n",
    "Y_test = pd.read_csv('../dev-0/expected.tsv', sep='\\t', header=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "9fc7a069-52ce-4343-a8a2-dd77619f5d25",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = dataset[dataset.columns[1]]\n",
    "Y_train = dataset[dataset.columns[0]]\n",
    "\n",
    "X_test = X_test[X_test.columns[0]]\n",
    "Y_test = Y_test[Y_test.columns[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7cd49893-c730-4089-9b21-8683c11fd6cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = [text.split() for text in X_train]\n",
    "X_test = [text.split() for text in X_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "bd029e57-f5c3-4b2b-bcf0-ea25d71ee952",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of the model checkpoint at dbmdz/bert-large-cased-finetuned-conll03-english were not used when initializing BertForTokenClassification: ['bert.pooler.dense.bias', 'bert.pooler.dense.weight']\n",
      "- This IS expected if you are initializing BertForTokenClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing BertForTokenClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9f4a002be4b64bfdba0a227b9a6815f2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0it [00:00, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "215\n"
     ]
    }
   ],
   "source": [
    "model = AutoModelForTokenClassification.from_pretrained(\"dbmdz/bert-large-cased-finetuned-conll03-english\")\n",
    "tokenizer = AutoTokenizer.from_pretrained(\"google-bert/bert-base-cased\")\n",
    "\n",
    "recognizer = pipeline(\"ner\", model=model, tokenizer=tokenizer)\n",
    "full_tags = []\n",
    "for idx, el in tqdm(enumerate(X_test)):\n",
    "    out_tags = []\n",
    "    tags_corrected = []\n",
    "\n",
    "    temp = []\n",
    "    tags_joined = []\n",
    "    tag = 'NULL'\n",
    "    t = \"\"\n",
    "    for r in recognizer(\" \".join(el)):\n",
    "        \n",
    "        if len(t) == 0:\n",
    "            t = r['word']\n",
    "            tag = r['entity']\n",
    "            continue\n",
    "        if \"#\" in r['word']:\n",
    "            t = t + str(r['word']).replace(\"#\",\"\")\n",
    "            continue\n",
    "        if \"#\" not in r['word'] and len(t) != 0:\n",
    "            temp.append(t)\n",
    "            tags_joined.append(tag)\n",
    "            t = r['word']\n",
    "            tag = r['entity']\n",
    "\n",
    "    for tag in Y_test[idx].split():\n",
    "        if tag == \"O\":\n",
    "            out_tags.append(\"O\")\n",
    "        if tag != \"O\" and len(tags_joined) > 0:\n",
    "            out_tags.append(tags_joined[0])\n",
    "            tags_joined = tags_joined[1:]\n",
    "            continue\n",
    "        if tag != \"O\" and len(tags_joined) == 0:\n",
    "            out_tags.append(\"O\")\n",
    "\n",
    "    #print(len(Y_test[idx].split()), len(out_tags))\n",
    "\n",
    "    out_tags = \" \".join(out_tags).replace(\"I-\",\"B-\").split()\n",
    "        \n",
    "    last_tag = out_tags[0]\n",
    "    tags_corrected.append(last_tag)\n",
    "    \n",
    "    for tag in out_tags[1:]:\n",
    "\n",
    "        if tag == last_tag:\n",
    "            tags_corrected.append(tag.replace(\"B-\",\"I-\"))\n",
    "            last_tag = tag\n",
    "        else:\n",
    "            last_tag = tag\n",
    "            tags_corrected.append(tag)\n",
    "            \n",
    "    #print(len(Y_test[idx].split()), len(tags_corrected))\n",
    "\n",
    "    full_tags.append(tags_corrected)\n",
    "\n",
    "print(len(full_tags))\n",
    "\n",
    "# for idx, el in tqdm(enumerate(full_tags)):\n",
    "#     if len(el) != len(len(Y_test[idx].split())):\n",
    "#         print(\"Somethings wrong sir\")\n",
    "   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "0feb1763-b28b-4cee-a996-f493c3f3d037",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"out.tsv\", 'w') as file:\n",
    "    for el in full_tags:\n",
    "    \n",
    "        file.write(\" \".join(el)) \n",
    "        file.write(f\"\\n\")\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}