Алиса.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Скачайте модель для распознавания русского или другого языка распознавания\n",
    "\n",
    "https://alphacephei.com/vosk/models -- с этого сайта. Там есть лёгкая (быстрее но хуже) и \"тяжёлая\" (медленнее но качественнее) версия. Распакуйте её в какую-нибудь папку и укажите путь к модели в переменной в следующем блоке."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_path = \"C:/dev/postnauka/models/vosk-model-ru-0.10\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Установите библиотеки, можно через `pip install`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install wave pyaudio wikipedia numpy matplotlib pyttsx3 vosk natasha yandex_music"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Импортируем все зависимости, загрузим модели распознавания и генерации"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json, os, sys, wave, pyaudio, wikipedia, numpy as np, matplotlib.pyplot as plt, pyttsx3\n",
    "from vosk import Model, KaldiRecognizer\n",
    "from natasha import *\n",
    "from IPython.display import HTML, clear_output, IFrame\n",
    "from yandex_music.client import Client\n",
    "\n",
    "kaldi_recognizer = KaldiRecognizer(\n",
    "        Model(model_path), 22050)\n",
    "wikipedia.set_lang(\"ru\")\n",
    "\n",
    "segmenter = Segmenter()\n",
    "emb = NewsEmbedding()\n",
    "morph_tagger = NewsMorphTagger(emb)\n",
    "syntax_parser = NewsSyntaxParser(emb)\n",
    "ner_tagger = NewsNERTagger(emb)\n",
    "morph_vocab = MorphVocab()\n",
    "names_extractor = NamesExtractor(morph_vocab)\n",
    "\n",
    "mus_client = Client()\n",
    "engine = pyttsx3.init()\n",
    "voices = engine.getProperty('voices')\n",
    "for voice in voices:\n",
    "    if 'RU' in voice.id:\n",
    "        engine.setProperty('voice', voice.id)\n",
    "        break\n",
    "clear_output()\n",
    "pya = pyaudio.PyAudio()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Функции записи и распознавания"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def записать(player, seconds=7):\n",
    "    print(\"Начинаем запись!\", end='\\t')\n",
    "    leng = (22050 * seconds)\n",
    "    stream = player.open(input=True, format=pyaudio.paInt16, channels=1, rate=22050,\n",
    "                         frames_per_buffer=leng)\n",
    "    data = stream.read(leng)\n",
    "    stream.stop_stream()\n",
    "    stream.close()\n",
    "    print(\"Записали!\")\n",
    "    return data\n",
    "\n",
    "def распознать(frames):\n",
    "    global kaldi_recognizer\n",
    "    words = {}\n",
    "    if kaldi_recognizer.AcceptWaveform(frames): words = json.loads(kaldi_recognizer.Result())\n",
    "    else: words = json.loads(kaldi_recognizer.PartialResult())\n",
    "    text = words['text'] if 'text' in words else words ['partial']\n",
    "    return text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Синтаксический и семантический анализ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def синтаксический_анализ(text):\n",
    "    doc = Doc(text)\n",
    "    doc.segment(segmenter)\n",
    "    doc.tag_morph(morph_tagger)\n",
    "    doc.tag_ner(ner_tagger)\n",
    "    for token in doc.tokens: token.lemmatize(morph_vocab)   \n",
    "    doc.parse_syntax(syntax_parser)\n",
    "    doc.sents[0].syntax.print()\n",
    "    action, dep = None, []\n",
    "    for tok in doc.tokens:\n",
    "        if tok.pos == 'VERB' or tok.rel == 'root': \n",
    "            action = tok\n",
    "            break\n",
    "    for tok in doc.tokens:\n",
    "        if tok != action: dep.append(tok)\n",
    "    return action.lemma if action else None, [d.text for d in dep]\n",
    "\n",
    "def embed(words):\n",
    "    n = 0\n",
    "    e = None\n",
    "    for word in words:\n",
    "        if word in emb:\n",
    "            n += 1\n",
    "            if e is None: e = emb.get(word)\n",
    "            else: e = e + emb.get(word)\n",
    "    if n > 0: return e / n\n",
    "    else: return None\n",
    "    \n",
    "def интент(action):\n",
    "    actions = {\n",
    "        'включить': 'play',\n",
    "        'играть': 'play',\n",
    "        'поставить': 'play',\n",
    "        'прочитать': 'wiki',\n",
    "        'рассказать': 'wiki',\n",
    "        'кто': 'wiki',\n",
    "        'какой': 'wiki'\n",
    "    }\n",
    "    res = []\n",
    "    for k in actions:\n",
    "        res.append([emb.sim(action, k), actions[k], k])\n",
    "    res.sort()\n",
    "    print(\"Интент:\", res[-1])\n",
    "    return res[-1][1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Функции действий с сервисами"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def play(text):\n",
    "    searchres = mus_client.search(text)\n",
    "    clear_output()\n",
    "    display(IFrame(f\"https://music.yandex.ru/iframe/#track/{searchres['best']['result']['id']}\", width=400, height=80))\n",
    "\n",
    "def wiki(text):\n",
    "    wikipedia.set_lang(\"ru\")\n",
    "    summ = wikipedia.summary(wikipedia.search(text)[0], sentences=3)\n",
    "    print(summ)\n",
    "    engine.say(summ)\n",
    "    engine.runAndWait()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Запись, распознавание и анализ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Начинаем запись!\tЗаписали!\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Вы сказали: сыграй город под подошвой\n",
      "     сыграй   \n",
      "┌─── город    \n",
      "│ ┌► под      case\n",
      "└►└─ подошвой nmod\n",
      "действие = сыграй\tобъект = город под подошвой\n",
      "Интент: [-0.027169278, 'play', 'играть']\n"
     ]
    }
   ],
   "source": [
    "сигнал = записать(pya, 5)\n",
    "plt.figure(figsize=(13, 2))\n",
    "plt.plot(np.frombuffer(сигнал, np.int16))\n",
    "plt.show()\n",
    "text = распознать(сигнал)\n",
    "print(\"Вы сказали:\", text)\n",
    "что_делать, с_чем = синтаксический_анализ(text)\n",
    "print(f\"действие = {что_делать}\\tобъект = {' '.join(с_чем)}\")\n",
    "intent_ = интент(что_делать)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Выполнение действий (обращение к сервисам)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "        <iframe\n",
       "            width=\"400\"\n",
       "            height=\"80\"\n",
       "            src=\"https://music.yandex.ru/iframe/#track/28420997\"\n",
       "            frameborder=\"0\"\n",
       "            allowfullscreen\n",
       "        ></iframe>\n",
       "        "
      ],
      "text/plain": [
       "<IPython.lib.display.IFrame at 0x1a3bc4fab50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "if intent_ == 'wiki': wiki(' '.join(с_чем))\n",
    "elif intent_ == 'play': play(' '.join(с_чем))"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}