{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "DEFAULT_BASE_URL = \"http://pipeline:8080/%s\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import os.path\n",
    "import requests\n",
    "from urllib.parse import urlencode\n",
    "import json\n",
    "# things we need for NLP\n",
    "import nltk\n",
    "from nltk.stem.lancaster import LancasterStemmer\n",
    "stemmer = LancasterStemmer()\n",
    "import tensorflow.python.framework.dtypes\n",
    "# things we need for Tensorflow\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import random\n",
    "\n",
    "from flask import Flask, jsonify, request\n",
    "from flask_cors import CORS, cross_origin\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# restore all of our data structures\n",
    "import pickle\n",
    "data = pickle.load( open( \"training_data.parquet\", \"rb\" ) )\n",
    "# data = pd.read_parquet(os.path.join(\"data\", \"training_data.parquet\"))\n",
    "words = data['words']\n",
    "classes = data['classes']\n",
    "train_x = data['train_x']\n",
    "train_y = data['train_y']\n",
    "\n",
    "# import our chat-bot intents file\n",
    "import json\n",
    "with open('intents.json') as json_data:\n",
    "    intents = json.load(json_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 1000\n",
    "batch_size = 8\n",
    "\n",
    "# Build neural network\n",
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Dense(8, input_shape=(len(train_x[0]),)),\n",
    "    tf.keras.layers.Dense(8),\n",
    "    tf.keras.layers.Dense(len(train_y[0]), activation=\"softmax\"),\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def clean_up_sentence(sentence):\n",
    "    # tokenize the pattern\n",
    "    sentence_words = nltk.word_tokenize(sentence)\n",
    "    # stem each word\n",
    "    sentence_words = [stemmer.stem(word.lower()) for word in sentence_words]\n",
    "    return sentence_words\n",
    "\n",
    "# return bag of words array: 0 or 1 for each word in the bag that exists in the sentence\n",
    "def bow(sentence, words, show_details=False):\n",
    "    # tokenize the pattern\n",
    "    sentence_words = clean_up_sentence(sentence)\n",
    "    # bag of words\n",
    "    bag = [0]*len(words)  \n",
    "    for s in sentence_words:\n",
    "        for i,w in enumerate(words):\n",
    "            if w == s: \n",
    "                bag[i] = 1\n",
    "                if show_details:\n",
    "                    print (\"found in bag: %s\" % w)\n",
    "\n",
    "    return(np.array(bag))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0\n",
      " 0 0 0 0 0 0 0 0]\n",
      "['adverse_drug', 'blood_pressure', 'blood_pressure_search', 'goodbye', 'greeting', 'hospital_search', 'options', 'pharmacy_search', 'thanks']\n"
     ]
    }
   ],
   "source": [
    "p = bow(\"looking for help\", words)\n",
    "print (p)\n",
    "print (classes)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a data structure to hold user context\n",
    "context = {}\n",
    "\n",
    "ERROR_THRESHOLD = 0.25\n",
    "userid =123\n",
    "\n",
    "\n",
    "def score_chat(text, url = None):\n",
    "    d = np.array([bow(sentence, words)])\n",
    "    payload = urlencode({\"json_args\" : json.dumps(d)})\n",
    "    headers = {'content-type': 'application/x-www-form-urlencoded'}\n",
    "    response = requests.request(\"POST\", url, data=payload, headers=headers)\n",
    "    try:\n",
    "        return json.loads(response.text)\n",
    "    except BaseException as e:\n",
    "        raise RuntimeError(\"Error: caught %r while processing %r (%r)\" % (e, response, response.text))\n",
    "\n",
    "\n",
    "# @app.route(\"/redsam/api/v1.0/classify\", methods=['POST'])\n",
    "def classify(sentence):\n",
    "    # generate probabilities from the model\n",
    "   \n",
    "    results = score_chat(sentence)[0]\n",
    "    # filter out predictions below a threshold\n",
    "    results = [[i,r] for i,r in enumerate(results) if r>ERROR_THRESHOLD]\n",
    "    # sort by strength of probability\n",
    "    results.sort(key=lambda x: x[1], reverse=True)\n",
    "    return_list = []\n",
    "    for r in results:\n",
    "        return_list.append((classes[r[0]], r[1]))\n",
    "    # return tuple of intent and probability\n",
    "    return return_list\n",
    "\n",
    "def response(sentence, userID=userid, show_details=False):\n",
    "    results = classify(sentence)\n",
    "    # if we have a classification then find the matching intent tag\n",
    "    if results:\n",
    "        # loop as long as there are matches to process\n",
    "        while results:\n",
    "            for i in intents['intents']:\n",
    "                # find a tag matching the first result\n",
    "                if i['tag'] == results[0][0]:\n",
    "                    # set context for this intent if necessary\n",
    "                    if 'context_set' in i:\n",
    "                        if show_details: print ('context:', i['context_set'])\n",
    "                        context[userID] = i['context_set']\n",
    "\n",
    "                    # check if this intent is contextual and applies to this user's conversation\n",
    "                    if not 'context_filter' in i or \\\n",
    "                        (userID in context and 'context_filter' in i and i['context_filter'] == context[userID]):\n",
    "                        if show_details: print ('tag:', i['tag'])\n",
    "                        # a random response from the intent\n",
    "                        print(random.choice(i['responses']))\n",
    "                    return\n",
    "\n",
    "            results.pop(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "hello I don't know how to score the args (['what help you offer?']) you supplied\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "\"I don't know how to score the args (['what help you offer?']) you supplied\""
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "score_chat(\"what help you offer?\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}