{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "1 пункт"
      ],
      "metadata": {
        "id": "6tuecB2YaGZd"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EexxrLenVcsK"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "os.chdir('/content/drive/MyDrive/Colab Notebooks/IS_LR4')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "device_name = tf.test.gpu_device_name()\n",
        "if device_name != '/device:GPU:0':\n",
        "  raise SystemError('GPU device not found')\n",
        "print('Found GPU at: {}'.format(device_name))"
      ],
      "metadata": {
        "id": "JZQBWhlnZ3Kz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "2 пункт"
      ],
      "metadata": {
        "id": "BMVDcB8saN1d"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# загрузка датасета\n",
        "from keras.datasets import imdb\n",
        "vocabulary_size = 5000\n",
        "index_from = 3\n",
        "(X_train, y_train), (X_test, y_test) = imdb.load_data(path=\"imdb.npz\",\n",
        "num_words=vocabulary_size,\n",
        "skip_top=0,\n",
        "maxlen=None,\n",
        "seed=19,\n",
        "start_char=1,\n",
        "oov_char=2,\n",
        "index_from=index_from\n",
        ")"
      ],
      "metadata": {
        "id": "dVZOh4OjaNfT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Размер обучающего множества X_train:\", X_train.shape)\n",
        "print(\"Размер обучающих меток y_train:\", y_train.shape)\n",
        "print(\"Размер тестового множества X_test:\", X_test.shape)\n",
        "print(\"Размер тестовых меток y_test:\", y_test.shape)"
      ],
      "metadata": {
        "id": "ONS-UYoqb7VM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "3 пункт"
      ],
      "metadata": {
        "id": "vrPoEAUFcZ6P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# создание словаря для перевода индексов в слова\n",
        "# заргузка словаря \"слово:индекс\"\n",
        "word_to_id = imdb.get_word_index()\n",
        "# уточнение словаря\n",
        "word_to_id = {key:(value + index_from) for key,value in word_to_id.items()}\n",
        "word_to_id[\"<PAD>\"] = 0\n",
        "word_to_id[\"<START>\"] = 1\n",
        "word_to_id[\"<UNK>\"] = 2\n",
        "word_to_id[\"<UNUSED>\"] = 3\n",
        "# создание обратного словаря \"индекс:слово\"\n",
        "id_to_word = {value:key for key,value in word_to_id.items()}"
      ],
      "metadata": {
        "id": "ldO9EDteccDO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "idx = 19\n",
        "review_indices = X_train[idx]\n",
        "print(\"Отзыв в виде индексов:\\n\", review_indices)\n",
        "\n",
        "review_text = \" \".join(id_to_word.get(i, \"?\") for i in review_indices)\n",
        "print(\"\\nОтзыв в виде текста:\\n\", review_text)\n",
        "\n",
        "print(\"\\nДлина отзыва (количество индексов):\", len(review_indices))\n",
        "\n",
        "label = y_train[idx]\n",
        "class_name = \"Positive\" if label == 1 else \"Negative\"\n",
        "print(\"Метка класса:\", label, \"| Класс:\", class_name)\n"
      ],
      "metadata": {
        "id": "dgLjpdWVcaT1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "4 пункт"
      ],
      "metadata": {
        "id": "sJRCzgSXdyaX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Максимальная длина отзыва:\", len(max(X_train, key=len)))\n",
        "print(\"Минимальная длина отзыва:\", len(min(X_train, key=len)))"
      ],
      "metadata": {
        "id": "NKuA8LSfd4nq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "5 пункт"
      ],
      "metadata": {
        "id": "xRZqJEnkensA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# предобработка данных\n",
        "from tensorflow.keras.utils import pad_sequences\n",
        "max_words = 500\n",
        "X_train = pad_sequences(X_train, maxlen=max_words, value=0, padding='pre', truncating='post')\n",
        "X_test = pad_sequences(X_test, maxlen=max_words, value=0, padding='pre', truncating='post')\n",
        "\n"
      ],
      "metadata": {
        "id": "FcSsuqWGeqDE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "6 пункт"
      ],
      "metadata": {
        "id": "Mrh00gk2gHaS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Максимальная длина отзыва после предобработки:\", len(max(X_train, key=len)))\n",
        "print(\"Минимальная длина отзыва после предобработки:\", len(min(X_train, key=len)))"
      ],
      "metadata": {
        "id": "woU0e9UMeqQi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "7 пункт\n"
      ],
      "metadata": {
        "id": "jvOmAItEgJsq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "idx = 19\n",
        "review_indices = X_train[idx]\n",
        "print(\"Отзыв в виде индексов:\\n\", review_indices)\n",
        "\n",
        "review_text = \" \".join(id_to_word.get(i, \"?\") for i in review_indices)\n",
        "print(\"\\nОтзыв в виде текста:\\n\", review_text)\n",
        "\n",
        "print(\"\\nДлина отзыва (количество индексов):\", len(review_indices))\n",
        "\n",
        "label = y_train[idx]\n",
        "class_name = \"Positive\" if label == 1 else \"Negative\"\n",
        "print(\"Метка класса:\", label, \"| Класс:\", class_name)"
      ],
      "metadata": {
        "id": "LGoRw4AHgJ9s"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "8 пункт"
      ],
      "metadata": {
        "id": "5CAdHfC_hVfo"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "print(\"Предобработанное обучающее множество X_train (первые 5 примеров):\")\n",
        "print(X_train[:5])\n",
        "\n",
        "print(\"\\nПредобработанное тестовое множество X_test (первые 5 примеров):\")\n",
        "print(X_test[:5])\n",
        "\n",
        "\n",
        "print(\"Размер обучающего множества X_train:\", X_train.shape)\n",
        "print(\"Размер обучающих меток y_train:\", y_train.shape)\n",
        "print(\"Размер тестового множества X_test:\", X_test.shape)\n",
        "print(\"Размер тестовых меток y_test:\", y_test.shape)"
      ],
      "metadata": {
        "id": "OAsrX4WdhYAx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "9 пункт"
      ],
      "metadata": {
        "id": "GTQnxs9AjEaz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Embedding, LSTM, Dropout, Dense\n",
        "\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Embedding(input_dim=vocabulary_size, output_dim=32, input_length=max_words))\n",
        "model.add(LSTM(64))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(1, activation='sigmoid'))\n",
        "\n",
        "\n",
        "model.compile(\n",
        "    optimizer='adam',\n",
        "    loss='binary_crossentropy',\n",
        "    metrics=['accuracy']\n",
        ")\n",
        "\n",
        "model.build(input_shape=(None, max_words))\n",
        "model.summary()"
      ],
      "metadata": {
        "id": "R_-5hCfGjibD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Обучение модели\n",
        "history = model.fit(\n",
        "    X_train,\n",
        "    y_train,\n",
        "    epochs=5,\n",
        "    batch_size=64,\n",
        "    validation_split=0.2,\n",
        "    verbose=1\n",
        ")"
      ],
      "metadata": {
        "id": "JyegJgdBlU4P"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nКачество обучения по эпохам\")\n",
        "for i in range(5):\n",
        "    train_acc = history.history['accuracy'][i]\n",
        "    val_acc = history.history['val_accuracy'][i]\n",
        "    print(f\"Эпоха {i+1}: accuracy = {train_acc:.4f}, val_accuracy = {val_acc:.4f}\")"
      ],
      "metadata": {
        "id": "9etKZpeVmeNj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "10 пункт"
      ],
      "metadata": {
        "id": "sulUG0iDmukX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_loss, test_accuracy = model.evaluate(X_test, y_test, verbose=0)\n",
        "\n",
        "print(\"Качество классификации на тестовой выборке\")\n",
        "print(f\"Test accuracy: {test_accuracy:.4f}\")"
      ],
      "metadata": {
        "id": "hVdh7SIAnWrx"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_score = model.predict(X_test)\n",
        "y_pred = [1 if y_score[i,0]>=0.5 else 0 for i in range(len(y_score))]\n",
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(y_test, y_pred, labels = [0, 1], target_names=['Negative', 'Positive']))"
      ],
      "metadata": {
        "id": "-p7pfGE7mwZi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import roc_curve, auc\n",
        "import matplotlib.pyplot as plt\n",
        "fpr, tpr, thresholds = roc_curve(y_test, y_score)\n",
        "plt.plot(fpr, tpr)\n",
        "plt.grid()\n",
        "plt.xlabel('False Positive Rate')\n",
        "plt.ylabel('True Positive Rate')\n",
        "plt.title('ROC')\n",
        "plt.show()\n",
        "print('Area under ROC is', auc(fpr, tpr))"
      ],
      "metadata": {
        "id": "A_ZMEN_YpmAq"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}