{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "%pwd\n",
    "os.chdir(\"../\")\n",
    "\n",
    "%pwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from dataclasses import dataclass\n",
    "from pathlib import Path\n",
    "\n",
    "\n",
    "@dataclass(frozen=True)\n",
    "class DataTransformationConfig:\n",
    "    root_dir: str\n",
    "    data_path: str\n",
    "    tokenizer_name: str"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from box import ConfigBox\n",
    "from pathlib import Path\n",
    "from src.TextSummarizer.constants import file_path\n",
    "from src.TextSummarizer.utils.general import read_yaml, create_directories\n",
    "\n",
    "class ConfigurationManager:\n",
    "    \"\"\"\n",
    "    Class to manage the configuration files.\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self) -> None:\n",
    "        self.config: ConfigBox = read_yaml(Path(file_path.CONFIG_FILE_PATH))\n",
    "        self.params: ConfigBox = read_yaml(Path(file_path.PARAMS_FILE_PATH))\n",
    "\n",
    "        create_directories(path_to_directories=[self.config.artifacts_root])\n",
    "\n",
    "    def get_data_transformation_config(self) -> DataTransformationConfig:\n",
    "        config = self.config.data_transformation\n",
    "\n",
    "        create_directories([config.root_dir])\n",
    "\n",
    "        data_transformation_config = DataTransformationConfig(\n",
    "            root_dir=config.root_dir,\n",
    "            data_path=config.data_path,\n",
    "            tokenizer_name = config.tokenizer_name\n",
    "        )\n",
    "\n",
    "        return data_transformation_config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "import os\n",
    "from src.TextSummarizer.logger import backend_logger\n",
    "from transformers import AutoTokenizer\n",
    "from datasets import load_dataset, load_from_disk\n",
    "\n",
    "\n",
    "class DataTransformation:\n",
    "    def __init__(self, config: DataTransformationConfig):\n",
    "        self.config = config\n",
    "        self.tokenizer = AutoTokenizer.from_pretrained(config.tokenizer_name)\n",
    "\n",
    "\n",
    "\n",
    "    def convert_examples_to_features(self,example_batch):\n",
    "        input_encodings = self.tokenizer(example_batch['dialogue'] , max_length = 800, truncation = True )\n",
    "\n",
    "        with self.tokenizer.as_target_tokenizer():\n",
    "            target_encodings = self.tokenizer(example_batch['summary'], max_length = 128, truncation = True )\n",
    "\n",
    "        return {\n",
    "            'input_ids' : input_encodings['input_ids'],\n",
    "            'attention_mask': input_encodings['attention_mask'],\n",
    "            'labels': target_encodings['input_ids']\n",
    "        }\n",
    "\n",
    "\n",
    "    def convert(self):\n",
    "        dataset = load_from_disk(self.config.data_path)\n",
    "        dataset = dataset.map(self.convert_examples_to_features, batched = True)\n",
    "        dataset.save_to_disk(os.path.join(self.config.root_dir,\"dataset\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    config = ConfigurationManager()\n",
    "    data_transformation_config = config.get_data_transformation_config()\n",
    "    data_transformation = DataTransformation(config=data_transformation_config)\n",
    "    data_transformation.convert()\n",
    "except Exception as e:\n",
    "    raise e"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}