{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "d7710e8b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Xet Storage is enabled for this repo, but the 'hf_xet' package is not installed. Falling back to regular HTTP download. For better performance, install the package with: `pip install huggingface_hub[hf_xet]` or `pip install hf_xet`\n",
      "c:\\Users\\User\\Chatbot\\.venv\\Lib\\site-packages\\huggingface_hub\\file_download.py:143: UserWarning: `huggingface_hub` cache-system uses symlinks by default to efficiently store duplicated files but your machine does not support them in C:\\Users\\User\\.cache\\huggingface\\hub\\models--facebook--bart-base. Caching files will still work but in a degraded version that might require more space on your disk. This warning can be disabled by setting the `HF_HUB_DISABLE_SYMLINKS_WARNING` environment variable. For more details, see https://huggingface.co/docs/huggingface_hub/how-to-cache#limitations.\n",
      "To support symlinks on Windows, you either need to activate Developer Mode or to run Python as an administrator. In order to activate developer mode, see this article: https://docs.microsoft.com/en-us/windows/apps/get-started/enable-your-device-for-development\n",
      "  warnings.warn(message)\n"
     ]
    }
   ],
   "source": [
    "from transformers import AutoTokenizer, AutoModelForSeq2SeqLM\n",
    "\n",
    "facebook_model_name = \"facebook/blenderbot-400M-distill\"\n",
    "google_model_name = \"google/flan-t5-base\"\n",
    "bart_model_name = \"facebook/bart-base\"\n",
    "\n",
    "\n",
    "# Load the models and tokenizers\n",
    "fb_model = AutoModelForSeq2SeqLM.from_pretrained(facebook_model_name)\n",
    "google_model = AutoModelForSeq2SeqLM.from_pretrained(google_model_name)\n",
    "bart_model = AutoModelForSeq2SeqLM.from_pretrained(bart_model_name)\n",
    "\n",
    "# Tokenizers\n",
    "fb_tokenizer = AutoTokenizer.from_pretrained(facebook_model_name)\n",
    "google_tokenizer = AutoTokenizer.from_pretrained(google_model_name)\n",
    "bart_tokenizer= AutoTokenizer.from_pretrained(bart_model_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "2f852aa5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define chat function\n",
    "def chat_with_bots(input_text, model_name=\"google\"):\n",
    "    if model_name == \"google\":\n",
    "        inputs = google_tokenizer.encode(input_text, return_tensors='pt')\n",
    "        outputs = google_model.generate(inputs, max_new_tokens=1000)\n",
    "        response = google_tokenizer.decode(\n",
    "            outputs[0], \n",
    "            skip_special_tokens=True\n",
    "        ).strip()\n",
    "        \n",
    "        return response\n",
    "    elif model_name == \"facebook\":\n",
    "        inputs = fb_tokenizer.encode(input_text, return_tensors='pt')\n",
    "        outputs = fb_model.generate(inputs, max_new_tokens=1000)\n",
    "        response = fb_tokenizer.decode(\n",
    "            outputs[0], \n",
    "            skip_special_tokens=True\n",
    "        ).strip()\n",
    "        return response\n",
    "        \n",
    "    elif model_name == \"bart\":\n",
    "        inputs = bart_tokenizer.encode(input_text, return_tensors='pt')\n",
    "        outputs = bart_model.generate(inputs, max_new_tokens=1000)\n",
    "        response = bart_tokenizer.decode(\n",
    "            outputs[0], \n",
    "            skip_special_tokens=True\n",
    "        ).strip()\n",
    "        return response\n",
    "    else: \n",
    "        response = \"No model selected\"\n",
    "        return response"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "568e3555",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Hi, my name is samantha. How are you doing this fine evening?\n"
     ]
    }
   ],
   "source": [
    "print(chat_with_bots(\"hi, what your name\", model_name=\"facebook\"))"
   ]
  }
 ],
 "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}