{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/emmanuelkoupoh/Documents/Github/LP_NLP/venv/lib/python3.9/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "from transformers import AutoModelForSequenceClassification\n",
    "from transformers import TFAutoModelForSequenceClassification\n",
    "from transformers import AutoTokenizer, AutoConfig\n",
    "import numpy as np\n",
    "from scipy.special import softmax"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "tokenizer = AutoTokenizer.from_pretrained('bert-base-cased')\n",
    "\n",
    "model_path = f\"test_trainer/checkpoint-1000/\"\n",
    "config = AutoConfig.from_pretrained(model_path)\n",
    "model = AutoModelForSequenceClassification.from_pretrained(model_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preprocess text (username and link placeholders)\n",
    "def preprocess(text):\n",
    "    new_text = []\n",
    "    for t in text.split(\" \"):\n",
    "        t = '@user' if t.startswith('@') and len(t) > 1 else t\n",
    "        t = 'http' if t.startswith('http') else t\n",
    "        new_text.append(t)\n",
    "    return \" \".join(new_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Input preprocessing\n",
    "text = \"Covid cases are increasing fast!\"\n",
    "text = preprocess(text)\n",
    "\n",
    "# PyTorch-based models\n",
    "encoded_input = tokenizer(text, return_tensors='pt')\n",
    "output = model(**encoded_input)\n",
    "scores = output[0][0].detach().numpy()\n",
    "scores = softmax(scores)\n",
    "\n",
    "# TensorFlow-based models\n",
    "# model = TFAutoModelForSequenceClassification.from_pretrained(model_path)\n",
    "# model.save_pretrained(model_path)\n",
    "# text = \"Covid cases are increasing fast!\"\n",
    "# encoded_input = tokenizer(text, return_tensors='tf')\n",
    "# output = model(encoded_input)\n",
    "# scores = output[0][0].numpy()\n",
    "# scores = softmax(scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "config.id2label = {0: 'NEGATIVE', 1: 'NEUTRAL', 2: 'POSITIVE'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1) NEUTRAL 0.9564\n",
      "2) POSITIVE 0.0389\n",
      "3) NEGATIVE 0.0047\n"
     ]
    }
   ],
   "source": [
    "# Print labels and scores\n",
    "ranking = np.argsort(scores)\n",
    "ranking = ranking[::-1]\n",
    "for i in range(scores.shape[0]):\n",
    "    l = config.id2label[ranking[i]]\n",
    "    s = scores[ranking[i]]\n",
    "    print(f\"{i+1}) {l} {np.round(float(s), 4)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.6 ('venv': 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": "3.9.6"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "1ab24538aa0da4b2d8c48eaca591ff7ffc54671225fb0511b432fd9e26a098ba"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}