chat_bot_nltk.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "#load database (or preporcessing database):\n",
        "problems: \n",
        "*finding database*\n",
        "*nltk takes matrix as a input* or any machine learning model or in 2 3 array \n"
      ],
      "metadata": {
        "id": "21ItH1UO2l70"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#**Important Keynotes:**\n",
        ">* Activation function (used to make my model's answer either 0 or 1 in case when my model return probability) for that we use non-linear functions such as **relu** or **tanh** etc.\n",
        ">* dropout layer (**dropout is a one of the regularization technique that prevent model from Overfitting** (performing good on training whereas poorer on new data))\n",
        ">* some good way to use for loop while building model:\n",
        " ```\n",
        " for units in [128, 64, 32]:\n",
        "  dense_tensor = tf.keras.layers.Dense(units, activation='relu')(\n",
        "    dense_tensor)\n",
        " prediction = tf.keras.layers.Dense(1)(dense_tensor)\n",
        " ```\n",
        ">* Converting categorical features to numeric using: \n",
        "```\n",
        "keywords_embedded = tf.feature_column.embedding_column(\n",
        "     tf.feature_column.categorical_column_with_hash_bucket(\"keywords\", 10000)\n",
        "```\n",
        ">* **tf.keras.layers.DenseFetures** layer is used to convert my features into a Dense Tensor or an array\n",
        " that can later be fed into model as an input. \n",
        ">* It is my choice whether i want to use **tf.keras.layers.DenseFeatures** layer or not because i can manually separate my features and lables which gives me more freedom to tinker them. But if i am manually specifying my features while i am training my mobel i need to specify the shape of my input and output to my neural network's **_input layer_** (or first layer)\n",
        " and **_output layer_** simultaneously such as:\n",
        "```\n",
        "input_shape = (len(train_X[0]),)\n",
        "output_shape = len(train_y[0])\n",
        "epochs = 200# the deep learning model\n",
        "model = Sequential()\n",
        "model.add(Dense(128, input_shape=input_shape, activation=\"relu\"))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(64, activation=\"relu\"))\n",
        "model.add(Dropout(0.3))\n",
        "model.add(Dense(output_shape, activation = \"softmax\"))adam = tf.keras.optimizers.Adam(learning_rate=0.01, decay=1e-6)model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=adam,\n",
        "              metrics=[\"accuracy\"])\n",
        "print(model.summary())\n",
        "model.fit(x=train_X, y=train_y, epochs=200, verbose=1)\n",
        "```"
      ],
      "metadata": {
        "id": "YUOlZzkeysmr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Source Code"
      ],
      "metadata": {
        "id": "m2UbIL0ljsTD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "data = {\"intents\": [\n",
        "             {\"tag\": \"greeting\",\n",
        "              \"patterns\": [\"Hello\", \"How are you?\", \"Hi there\", \"Hi\", \"Whats up\"],\n",
        "              \"responses\": [\"Howdy Partner!\", \"Hello\", \"How are you doing?\", \"Greetings!\", \"How do you do?\"],\n",
        "             },\n",
        "             {\"tag\": \"age\",\n",
        "              \"patterns\": [\"how old are you?\", \"when is your birthday?\", \"when was you born?\"],\n",
        "              \"responses\": [\"I am 24 years old\", \"I was born in 1996\", \"My birthday is July 3rd and I was born in 1996\", \"03/07/1996\"]\n",
        "             },\n",
        "             {\"tag\": \"date\",\n",
        "              \"patterns\": [\"what are you doing this weekend?\",\n",
        "\"do you want to hang out some time?\", \"what are your plans for this week\"],\n",
        "              \"responses\": [\"I am available all week\", \"I don't have any plans\", \"I am not busy\"]\n",
        "             },\n",
        "             {\"tag\": \"name\",\n",
        "              \"patterns\": [\"what's your name?\", \"what are you called?\", \"who are you?\"],\n",
        "              \"responses\": [\"My name is Kippi\", \"I'm Kippi\", \"Kippi\"]\n",
        "             },\n",
        "             {\"tag\": \"goodbye\",\n",
        "              \"patterns\": [ \"bye\", \"g2g\", \"see ya\", \"adios\", \"cya\"],\n",
        "              \"responses\": [\"It was nice speaking to you\", \"See you later\", \"Speak soon!\"]\n",
        "             }\n",
        "]}"
      ],
      "metadata": {
        "id": "k__iLV0JxJ5a"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import json\n",
        "import string\n",
        "import random \n",
        "import nltk\n",
        "import numpy as np\n",
        "from nltk.stem import WordNetLemmatizer \n",
        "import tensorflow as tf \n",
        "from tensorflow.keras import Sequential \n",
        "from tensorflow.keras.layers import Dense, Dropout\n",
        "from matplotlib import pyplot as plt\n",
        "import pandas as pd\n",
        "\n",
        "nltk.download(\"punkt\")\n",
        "nltk.download(\"wordnet\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xQ5JYHZpxjIS",
        "outputId": "9ac7188a-e5c3-4519-b295-ecad2dd07fe1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]   Unzipping tokenizers/punkt.zip.\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "nltk.download('omw-1.4')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BdGIVIEhyp0F",
        "outputId": "b3ae0008-8aea-4e32-bdfd-7d6d225a79e0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package omw-1.4 to /root/nltk_data...\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# initializing lemmatizer to get stem of words\n",
        "lemmatizer = WordNetLemmatizer()# Each list to create\n",
        "words = []\n",
        "classes = []\n",
        "doc_X = []\n",
        "doc_y = []# Loop through all the intents\n",
        "# tokenize each pattern and append tokens to words, the patterns and\n",
        "# the associated tag to their associated list\n",
        "for intent in data[\"intents\"]:\n",
        "    for pattern in intent[\"patterns\"]:\n",
        "        tokens = nltk.word_tokenize(pattern)\n",
        "        words.extend(tokens)\n",
        "        doc_X.append(pattern)\n",
        "        doc_y.append(intent[\"tag\"])\n",
        "    \n",
        "    # add the tag to the classes if it's not there already \n",
        "    if intent[\"tag\"] not in classes:\n",
        "        classes.append(intent[\"tag\"])\n",
        "# lemmatize all the words in the vocab and convert them to lowercase\n",
        "# remove any punctuations\n",
        "words = [lemmatizer.lemmatize(word.lower()) \n",
        "        for word in words \n",
        "         if word not in string.punctuation]# sorting the vocab and classes in alphabetical order and\n",
        "                                           # taking the set to ensure no duplicates occur\n",
        "words = sorted(set(words))\n",
        "classes = sorted(set(classes))"
      ],
      "metadata": {
        "id": "Ps9xtnwOxy-H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(classes)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uTI-TdXB7u0i",
        "outputId": "4bf6b18e-fbf4-4a54-c261-a986c08189bf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['age', 'date', 'goodbye', 'greeting', 'name']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# list for training data\n",
        "training = []\n",
        "out_empty = [0] * len(classes)# creating the bag of words model\n",
        "for idx, doc in enumerate(doc_X):\n",
        "    print(idx,doc)\n",
        "    bow = []\n",
        "    text = lemmatizer.lemmatize(doc.lower())\n",
        "    for word in words:\n",
        "        bow.append(1) if word in text else bow.append(0)# mark the index of class that the current pattern is associated to\n",
        "    output_row = list(out_empty)\n",
        "    \n",
        "    output_row[classes.index(doc_y[idx])] = 1    # add the one hot encoded BoW and associated classes to training \n",
        "                                              # it will going to give me [0 0 0 0 0] list value which my model can use for creating pattern\n",
        "    \n",
        "    training.append([bow, output_row])\n",
        "    # shuffle the data and convert it to an array\n",
        "random.shuffle(training)\n",
        "training = np.array(training, dtype=object)\n",
        "# split the features and target labels\n",
        "train_X = np.array(list(training[:, 0]))\n",
        "train_y = np.array(list(training[:, 1]))"
      ],
      "metadata": {
        "id": "TISL65dU5OND",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4bf58410-a1be-4cf2-9bf0-176b5bb67e10"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0 Hello\n",
            "1 How are you?\n",
            "2 Hi there\n",
            "3 Hi\n",
            "4 Whats up\n",
            "5 how old are you?\n",
            "6 when is your birthday?\n",
            "7 when was you born?\n",
            "8 what are you doing this weekend?\n",
            "9 do you want to hang out some time?\n",
            "10 what are your plans for this week\n",
            "11 what's your name?\n",
            "12 what are you called?\n",
            "13 who are you?\n",
            "14 bye\n",
            "15 g2g\n",
            "16 see ya\n",
            "17 adios\n",
            "18 cya\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def plot_the_loss_curve(epochs, mse):\n",
        "  \"\"\"Plot a curve of loss vs. epoch.\"\"\"\n",
        "\n",
        "  plt.figure()\n",
        "  plt.xlabel(\"Epoch\")\n",
        "  plt.ylabel(\"Mean Squared Error\")\n",
        "\n",
        "  plt.plot(epochs, mse, label=\"Loss\")\n",
        "  plt.legend()\n",
        "  plt.ylim([mse.min()*0.95, mse.max() * 1.03])\n",
        "  plt.show()  \n",
        "\n",
        "print(\"Defined the plot_the_loss_curve function.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e_-gTedaxnLy",
        "outputId": "3beb4bb2-4321-4366-e112-ce8e15fd89c7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Defined the plot_the_loss_curve function.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from nltk.translate import metrics\n",
        "# defining some parameters\n",
        "input_shape = (len(train_X[0]),)\n",
        "output_shape = len(train_y[0])\n",
        "def create_model(my_learning_rate):\n",
        "  model = Sequential()\n",
        "  model.add(Dense(128, input_shape=input_shape, activation=\"relu\"))\n",
        "  model.add(Dropout(0.5))\n",
        "  model.add(Dense(64, activation=\"relu\"))\n",
        "  model.add(Dropout(0.3))\n",
        "  model.add(Dense(output_shape, activation = \"softmax\"))\n",
        "  adam = tf.keras.optimizers.Adam(learning_rate=0.01, decay=1e-6)\n",
        "  model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=adam,\n",
        "              metrics=[\"accuracy\"])\n",
        "  return model"
      ],
      "metadata": {
        "id": "hW4Yu8rPqOeK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def train_model(model, batch_size, epochs):\n",
        "  featurs = train_X\n",
        "  labels = train_y\n",
        "  history = model.fit(x=featurs, y=labels, batch_size=batch_size,\n",
        "            epochs=epochs, shuffle=True)\n",
        "  epochs = history.epoch\n",
        "\n",
        "  # To track the progression of training, gather a snapshot\n",
        "  # of the model's mean squared error at each epoch. \n",
        "  hist = pd.DataFrame(history.history)\n",
        "  accuracy = hist[\"Accuracy\"]\n",
        "\n",
        "  return epochs, accuracy"
      ],
      "metadata": {
        "id": "GPeFgJXTwV7H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# defining some parameters\n",
        "input_shape = (len(train_X[0]),)\n",
        "output_shape = len(train_y[0])\n",
        "epochs = 200# the deep learning model\n",
        "model = Sequential()\n",
        "model.add(Dense(128, input_shape=input_shape, activation=\"relu\"))\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(64, activation=\"relu\"))\n",
        "model.add(Dropout(0.3))\n",
        "model.add(Dense(output_shape, activation = \"softmax\"))\n",
        "adam = tf.keras.optimizers.Adam(learning_rate=0.01, decay=1e-6)\n",
        "model.compile(loss='categorical_crossentropy',\n",
        "              optimizer=adam,\n",
        "              metrics=[\"accuracy\"])\n",
        "print(model.summary())\n",
        "model.fit(x=train_X, y=train_y, epochs=200, verbose=1)"
      ],
      "metadata": {
        "id": "hxMTJYO2z0Jn",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1ae1ab35-8139-4d91-ef1c-404c03ec2713"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 128)               5120      \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 128)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 64)                8256      \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 64)                0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 5)                 325       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 13,701\n",
            "Trainable params: 13,701\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Epoch 1/200\n",
            "1/1 [==============================] - 2s 2s/step - loss: 1.6258 - accuracy: 0.1579\n",
            "Epoch 2/200\n",
            "1/1 [==============================] - 0s 30ms/step - loss: 1.4896 - accuracy: 0.3158\n",
            "Epoch 3/200\n",
            "1/1 [==============================] - 0s 36ms/step - loss: 1.5141 - accuracy: 0.4737\n",
            "Epoch 4/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 1.4247 - accuracy: 0.4211\n",
            "Epoch 5/200\n",
            "1/1 [==============================] - 0s 28ms/step - loss: 1.1869 - accuracy: 0.7368\n",
            "Epoch 6/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 1.0457 - accuracy: 0.7368\n",
            "Epoch 7/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.9362 - accuracy: 0.9474\n",
            "Epoch 8/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.8702 - accuracy: 0.9474\n",
            "Epoch 9/200\n",
            "1/1 [==============================] - 0s 26ms/step - loss: 0.8093 - accuracy: 0.9474\n",
            "Epoch 10/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 0.6292 - accuracy: 1.0000\n",
            "Epoch 11/200\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 0.6209 - accuracy: 0.8947\n",
            "Epoch 12/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.4189 - accuracy: 1.0000\n",
            "Epoch 13/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.4676 - accuracy: 0.8947\n",
            "Epoch 14/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.2977 - accuracy: 0.9474\n",
            "Epoch 15/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.3579 - accuracy: 0.8947\n",
            "Epoch 16/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.1820 - accuracy: 1.0000\n",
            "Epoch 17/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.1100 - accuracy: 1.0000\n",
            "Epoch 18/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.1586 - accuracy: 1.0000\n",
            "Epoch 19/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.1901 - accuracy: 0.8947\n",
            "Epoch 20/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.1876 - accuracy: 0.9474\n",
            "Epoch 21/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0733 - accuracy: 1.0000\n",
            "Epoch 22/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 0.1062 - accuracy: 1.0000\n",
            "Epoch 23/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0568 - accuracy: 1.0000\n",
            "Epoch 24/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0690 - accuracy: 0.9474\n",
            "Epoch 25/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0234 - accuracy: 1.0000\n",
            "Epoch 26/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0349 - accuracy: 1.0000\n",
            "Epoch 27/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0112 - accuracy: 1.0000\n",
            "Epoch 28/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0369 - accuracy: 1.0000\n",
            "Epoch 29/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0308 - accuracy: 1.0000\n",
            "Epoch 30/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0104 - accuracy: 1.0000\n",
            "Epoch 31/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0274 - accuracy: 1.0000\n",
            "Epoch 32/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0785 - accuracy: 0.9474\n",
            "Epoch 33/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0039 - accuracy: 1.0000\n",
            "Epoch 34/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 0.0030 - accuracy: 1.0000\n",
            "Epoch 35/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0024 - accuracy: 1.0000\n",
            "Epoch 36/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 0.0210 - accuracy: 1.0000\n",
            "Epoch 37/200\n",
            "1/1 [==============================] - 0s 24ms/step - loss: 0.0079 - accuracy: 1.0000\n",
            "Epoch 38/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0480 - accuracy: 0.9474\n",
            "Epoch 39/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0046 - accuracy: 1.0000\n",
            "Epoch 40/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0527 - accuracy: 0.9474\n",
            "Epoch 41/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0163 - accuracy: 1.0000\n",
            "Epoch 42/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.0023 - accuracy: 1.0000\n",
            "Epoch 43/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0063 - accuracy: 1.0000\n",
            "Epoch 44/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0901 - accuracy: 0.9474\n",
            "Epoch 45/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0253 - accuracy: 1.0000\n",
            "Epoch 46/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0123 - accuracy: 1.0000\n",
            "Epoch 47/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0168 - accuracy: 1.0000\n",
            "Epoch 48/200\n",
            "1/1 [==============================] - 0s 40ms/step - loss: 4.3355e-04 - accuracy: 1.0000\n",
            "Epoch 49/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0021 - accuracy: 1.0000\n",
            "Epoch 50/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0315 - accuracy: 1.0000\n",
            "Epoch 51/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0105 - accuracy: 1.0000\n",
            "Epoch 52/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0025 - accuracy: 1.0000\n",
            "Epoch 53/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0019 - accuracy: 1.0000\n",
            "Epoch 54/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 6.6720e-04 - accuracy: 1.0000\n",
            "Epoch 55/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0964 - accuracy: 0.9474\n",
            "Epoch 56/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.0881 - accuracy: 0.9474\n",
            "Epoch 57/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.0020 - accuracy: 1.0000\n",
            "Epoch 58/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 2.6864e-04 - accuracy: 1.0000\n",
            "Epoch 59/200\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 0.0218 - accuracy: 1.0000\n",
            "Epoch 60/200\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 0.0049 - accuracy: 1.0000\n",
            "Epoch 61/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 8.8331e-04 - accuracy: 1.0000\n",
            "Epoch 62/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 0.0063 - accuracy: 1.0000\n",
            "Epoch 63/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0140 - accuracy: 1.0000\n",
            "Epoch 64/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0047 - accuracy: 1.0000\n",
            "Epoch 65/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0046 - accuracy: 1.0000\n",
            "Epoch 66/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 4.7832e-04 - accuracy: 1.0000\n",
            "Epoch 67/200\n",
            "1/1 [==============================] - 0s 33ms/step - loss: 3.2855e-04 - accuracy: 1.0000\n",
            "Epoch 68/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 5.2661e-04 - accuracy: 1.0000\n",
            "Epoch 69/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 0.0013 - accuracy: 1.0000\n",
            "Epoch 70/200\n",
            "1/1 [==============================] - 0s 29ms/step - loss: 0.0036 - accuracy: 1.0000\n",
            "Epoch 71/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.0027 - accuracy: 1.0000\n",
            "Epoch 72/200\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 7.5311e-05 - accuracy: 1.0000\n",
            "Epoch 73/200\n",
            "1/1 [==============================] - 0s 31ms/step - loss: 0.0012 - accuracy: 1.0000\n",
            "Epoch 74/200\n",
            "1/1 [==============================] - 0s 26ms/step - loss: 0.0158 - accuracy: 1.0000\n",
            "Epoch 75/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.0011 - accuracy: 1.0000\n",
            "Epoch 76/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.0059 - accuracy: 1.0000\n",
            "Epoch 77/200\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 0.0026 - accuracy: 1.0000\n",
            "Epoch 78/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 3.0485e-04 - accuracy: 1.0000\n",
            "Epoch 79/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 7.0142e-04 - accuracy: 1.0000\n",
            "Epoch 80/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 4.2963e-04 - accuracy: 1.0000\n",
            "Epoch 81/200\n",
            "1/1 [==============================] - 0s 30ms/step - loss: 1.9053e-04 - accuracy: 1.0000\n",
            "Epoch 82/200\n",
            "1/1 [==============================] - 0s 30ms/step - loss: 0.0010 - accuracy: 1.0000\n",
            "Epoch 83/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0019 - accuracy: 1.0000\n",
            "Epoch 84/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 4.3027e-04 - accuracy: 1.0000\n",
            "Epoch 85/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.0027 - accuracy: 1.0000\n",
            "Epoch 86/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.0836 - accuracy: 0.9474\n",
            "Epoch 87/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.1337 - accuracy: 0.9474\n",
            "Epoch 88/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0012 - accuracy: 1.0000\n",
            "Epoch 89/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0029 - accuracy: 1.0000\n",
            "Epoch 90/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0071 - accuracy: 1.0000\n",
            "Epoch 91/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 9.2697e-04 - accuracy: 1.0000\n",
            "Epoch 92/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.0069 - accuracy: 1.0000\n",
            "Epoch 93/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0019 - accuracy: 1.0000\n",
            "Epoch 94/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 2.0313e-04 - accuracy: 1.0000\n",
            "Epoch 95/200\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 0.0027 - accuracy: 1.0000\n",
            "Epoch 96/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 4.2215e-04 - accuracy: 1.0000\n",
            "Epoch 97/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0363 - accuracy: 1.0000\n",
            "Epoch 98/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.9093e-04 - accuracy: 1.0000\n",
            "Epoch 99/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0017 - accuracy: 1.0000\n",
            "Epoch 100/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 0.0012 - accuracy: 1.0000\n",
            "Epoch 101/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0011 - accuracy: 1.0000\n",
            "Epoch 102/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0209 - accuracy: 1.0000\n",
            "Epoch 103/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0040 - accuracy: 1.0000\n",
            "Epoch 104/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0068 - accuracy: 1.0000\n",
            "Epoch 105/200\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 0.0027 - accuracy: 1.0000\n",
            "Epoch 106/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 0.0075 - accuracy: 1.0000\n",
            "Epoch 107/200\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 0.0069 - accuracy: 1.0000\n",
            "Epoch 108/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 5.3557e-04 - accuracy: 1.0000\n",
            "Epoch 109/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0208 - accuracy: 1.0000\n",
            "Epoch 110/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0011 - accuracy: 1.0000\n",
            "Epoch 111/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0200 - accuracy: 1.0000\n",
            "Epoch 112/200\n",
            "1/1 [==============================] - 0s 30ms/step - loss: 0.0088 - accuracy: 1.0000\n",
            "Epoch 113/200\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 5.7127e-04 - accuracy: 1.0000\n",
            "Epoch 114/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 3.0202e-04 - accuracy: 1.0000\n",
            "Epoch 115/200\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 0.0071 - accuracy: 1.0000\n",
            "Epoch 116/200\n",
            "1/1 [==============================] - 0s 29ms/step - loss: 0.0019 - accuracy: 1.0000\n",
            "Epoch 117/200\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 0.0290 - accuracy: 1.0000\n",
            "Epoch 118/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 0.0014 - accuracy: 1.0000\n",
            "Epoch 119/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 0.0026 - accuracy: 1.0000\n",
            "Epoch 120/200\n",
            "1/1 [==============================] - 0s 27ms/step - loss: 9.7966e-04 - accuracy: 1.0000\n",
            "Epoch 121/200\n",
            "1/1 [==============================] - 0s 24ms/step - loss: 0.1398 - accuracy: 0.9474\n",
            "Epoch 122/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 4.3225e-04 - accuracy: 1.0000\n",
            "Epoch 123/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0010 - accuracy: 1.0000\n",
            "Epoch 124/200\n",
            "1/1 [==============================] - 0s 51ms/step - loss: 4.2126e-04 - accuracy: 1.0000\n",
            "Epoch 125/200\n",
            "1/1 [==============================] - 0s 31ms/step - loss: 3.9953e-04 - accuracy: 1.0000\n",
            "Epoch 126/200\n",
            "1/1 [==============================] - 0s 27ms/step - loss: 0.0109 - accuracy: 1.0000\n",
            "Epoch 127/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.0016 - accuracy: 1.0000\n",
            "Epoch 128/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0011 - accuracy: 1.0000\n",
            "Epoch 129/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 3.0659e-04 - accuracy: 1.0000\n",
            "Epoch 130/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0025 - accuracy: 1.0000\n",
            "Epoch 131/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.0022 - accuracy: 1.0000\n",
            "Epoch 132/200\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 5.2397e-04 - accuracy: 1.0000\n",
            "Epoch 133/200\n",
            "1/1 [==============================] - 0s 28ms/step - loss: 0.0015 - accuracy: 1.0000\n",
            "Epoch 134/200\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0015 - accuracy: 1.0000\n",
            "Epoch 135/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 9.4395e-04 - accuracy: 1.0000\n",
            "Epoch 136/200\n",
            "1/1 [==============================] - 0s 30ms/step - loss: 7.6673e-04 - accuracy: 1.0000\n",
            "Epoch 137/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 7.5254e-04 - accuracy: 1.0000\n",
            "Epoch 138/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0020 - accuracy: 1.0000\n",
            "Epoch 139/200\n",
            "1/1 [==============================] - 0s 28ms/step - loss: 0.0015 - accuracy: 1.0000\n",
            "Epoch 140/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.1078e-04 - accuracy: 1.0000\n",
            "Epoch 141/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 8.7029e-04 - accuracy: 1.0000\n",
            "Epoch 142/200\n",
            "1/1 [==============================] - 0s 28ms/step - loss: 0.0021 - accuracy: 1.0000\n",
            "Epoch 143/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0054 - accuracy: 1.0000\n",
            "Epoch 144/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 8.6731e-04 - accuracy: 1.0000\n",
            "Epoch 145/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 7.0832e-04 - accuracy: 1.0000\n",
            "Epoch 146/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 3.5030e-04 - accuracy: 1.0000\n",
            "Epoch 147/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0011 - accuracy: 1.0000\n",
            "Epoch 148/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0033 - accuracy: 1.0000\n",
            "Epoch 149/200\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 6.5598e-04 - accuracy: 1.0000\n",
            "Epoch 150/200\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 1.6478e-04 - accuracy: 1.0000\n",
            "Epoch 151/200\n",
            "1/1 [==============================] - 0s 23ms/step - loss: 1.8611e-04 - accuracy: 1.0000\n",
            "Epoch 152/200\n",
            "1/1 [==============================] - 0s 53ms/step - loss: 4.7155e-05 - accuracy: 1.0000\n",
            "Epoch 153/200\n",
            "1/1 [==============================] - 0s 40ms/step - loss: 0.0039 - accuracy: 1.0000\n",
            "Epoch 154/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 1.9009e-05 - accuracy: 1.0000\n",
            "Epoch 155/200\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 1.6303e-04 - accuracy: 1.0000\n",
            "Epoch 156/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 8.6456e-05 - accuracy: 1.0000\n",
            "Epoch 157/200\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 7.7713e-05 - accuracy: 1.0000\n",
            "Epoch 158/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.0038 - accuracy: 1.0000\n",
            "Epoch 159/200\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 0.0012 - accuracy: 1.0000\n",
            "Epoch 160/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 7.1746e-05 - accuracy: 1.0000\n",
            "Epoch 161/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 3.7708e-05 - accuracy: 1.0000\n",
            "Epoch 162/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 2.1580e-04 - accuracy: 1.0000\n",
            "Epoch 163/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 4.2273e-04 - accuracy: 1.0000\n",
            "Epoch 164/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 9.6495e-04 - accuracy: 1.0000\n",
            "Epoch 165/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 4.3556e-04 - accuracy: 1.0000\n",
            "Epoch 166/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.6098e-05 - accuracy: 1.0000\n",
            "Epoch 167/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 5.8766e-05 - accuracy: 1.0000\n",
            "Epoch 168/200\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 4.4648e-04 - accuracy: 1.0000\n",
            "Epoch 169/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 3.8868e-05 - accuracy: 1.0000\n",
            "Epoch 170/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 1.8453e-04 - accuracy: 1.0000\n",
            "Epoch 171/200\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 4.9421e-04 - accuracy: 1.0000\n",
            "Epoch 172/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 2.2690e-04 - accuracy: 1.0000\n",
            "Epoch 173/200\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 0.0096 - accuracy: 1.0000\n",
            "Epoch 174/200\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 0.0013 - accuracy: 1.0000\n",
            "Epoch 175/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 3.5758e-04 - accuracy: 1.0000\n",
            "Epoch 176/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 8.9844e-06 - accuracy: 1.0000\n",
            "Epoch 177/200\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.0011 - accuracy: 1.0000\n",
            "Epoch 178/200\n",
            "1/1 [==============================] - 0s 29ms/step - loss: 0.0030 - accuracy: 1.0000\n",
            "Epoch 179/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.4579e-04 - accuracy: 1.0000\n",
            "Epoch 180/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0012 - accuracy: 1.0000\n",
            "Epoch 181/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 1.7831e-04 - accuracy: 1.0000\n",
            "Epoch 182/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0014 - accuracy: 1.0000\n",
            "Epoch 183/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 6.4133e-04 - accuracy: 1.0000\n",
            "Epoch 184/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0020 - accuracy: 1.0000\n",
            "Epoch 185/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.1668e-04 - accuracy: 1.0000\n",
            "Epoch 186/200\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 5.9898e-04 - accuracy: 1.0000\n",
            "Epoch 187/200\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.7611e-05 - accuracy: 1.0000\n",
            "Epoch 188/200\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 7.2501e-04 - accuracy: 1.0000\n",
            "Epoch 189/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 5.4773e-06 - accuracy: 1.0000\n",
            "Epoch 190/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0251 - accuracy: 1.0000\n",
            "Epoch 191/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 5.1368e-04 - accuracy: 1.0000\n",
            "Epoch 192/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0014 - accuracy: 1.0000\n",
            "Epoch 193/200\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 4.4066e-05 - accuracy: 1.0000\n",
            "Epoch 194/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 7.9371e-05 - accuracy: 1.0000\n",
            "Epoch 195/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 5.8696e-04 - accuracy: 1.0000\n",
            "Epoch 196/200\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 7.5100e-06 - accuracy: 1.0000\n",
            "Epoch 197/200\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 3.8940e-04 - accuracy: 1.0000\n",
            "Epoch 198/200\n",
            "1/1 [==============================] - 0s 18ms/step - loss: 0.0120 - accuracy: 1.0000\n",
            "Epoch 199/200\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0019 - accuracy: 1.0000\n",
            "Epoch 200/200\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 1.5236e-04 - accuracy: 1.0000\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f9af1bd2990>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def clean_text(text): \n",
        "  tokens = nltk.word_tokenize(text)\n",
        "  tokens = [lemmatizer.lemmatize(word) for word in tokens]\n",
        "  return tokens\n",
        "\n",
        "def bag_of_words(text, vocab): \n",
        "  tokens = clean_text(text)\n",
        "  bow = [0] * len(vocab)\n",
        "  for w in tokens: \n",
        "    for idx, word in enumerate(vocab):\n",
        "      if word == w: \n",
        "        bow[idx] = 1\n",
        "  return np.array(bow)\n",
        "\n",
        "def pred_class(text, vocab, labels): \n",
        "  bow = bag_of_words(text, vocab)\n",
        "  result = model.predict(np.array([bow]))[0]\n",
        "  thresh = 0.2\n",
        "  y_pred = [[idx, res] for idx, res in enumerate(result) if res > thresh]\n",
        "\n",
        "  y_pred.sort(key=lambda x: x[1], reverse=True)\n",
        "  return_list = []\n",
        "  for r in y_pred:\n",
        "    return_list.append(labels[r[0]])\n",
        "  return return_list\n",
        "\n",
        "def get_response(intents_list, intents_json): \n",
        "  tag = intents_list[0]\n",
        "  list_of_intents = intents_json[\"intents\"]\n",
        "  for i in list_of_intents: \n",
        "    if i[\"tag\"] == tag:\n",
        "      result = random.choice(i[\"responses\"])\n",
        "      break\n",
        "  return result"
      ],
      "metadata": {
        "id": "LRK4cxg61bmB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# running the chatbot\n",
        "while True:\n",
        "    message = input(\"\")\n",
        "    intents = pred_class(message, words, classes)\n",
        "    result = get_response(intents, data)\n",
        "    print(result)\n",
        "    if message == \"bye\":\n",
        "      break\n",
        "  "
      ],
      "metadata": {
        "id": "Nvgy88mk1hxE"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}