SpeechToText.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Refer to the Speech to Text 3.0 Swagger documentation: https://develop.cris.ai/swagger/ui/index?urls.primaryName=Speech%20Services%20API%20v3.0#"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from azure.storage.blob import BlobServiceClient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "import requests\n",
    "import json\n",
    "import pandas as pd\n",
    "import argparse\n",
    "import time\n",
    "import http.client, urllib.request, urllib.parse, urllib.error, base64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Speech to Text region\n",
    "region = \"westeurope\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Speech to Text - subscription key\n",
    "SUBSCRIPTION_KEY = \"a3d9ab1fb3da4a2a825689ab47ac00b2\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Storage Account connection string and SAS key\n",
    "connect_str = 'DefaultEndpointsProtocol=https;AccountName=courtcalls;AccountKey=UyZFI1qHMfKskYz0KDnXJfcS2Z3g5TFqGGhSC34Nrr6ONG7t5X9vwNRR7C8AmvBnEEwSpwxDgwQry9u8uVCP/g==;EndpointSuffix=core.windows.net'\n",
    "SAS = '?sv=2019-12-12&ss=bfqt&srt=sco&sp=rwdlacupx&se=2020-10-19T06:32:24Z&st=2020-10-11T22:32:24Z&spr=https&sig=lzQcU16q9NkDZNEgkEwEvaT8dJpCUcFNGjCAmpVcB7E%3D'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "## storage account - account name and container name. This where we store the audio files\n",
    "container_name = \"test\"\n",
    "account_name = \"courtcalls\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Speech to Text Endpoint for Transcriptions \n",
    "\n",
    "BaseURL =\"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Storage path\n",
    "STORAGE_PATH=\"https://\" + account_name + \".blob.core.windows.net/\" + container_name + \"/\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prepare HTTP POST Request Body with associated Speech to Text properties"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Speech to Text - transcription properties\n",
    "def bulkTranscription(blob_uri, blob_name):\n",
    "    BD = \"\"\"\n",
    "            {\n",
    "                \"contentUrls\": [ \n",
    "                    \"__URL__\"],\n",
    "                \"locale\": \"en-US\",\n",
    "                \"displayName\": \"blob_name\",\n",
    "                \"properties\": {\n",
    "                    \"punctuationMode\": \"Automatic\",\n",
    "                    \"profanityFilterMode\": \"Masked\",\n",
    "                    \"wordLevelTimestampsEnabled\": \"False\",\n",
    "                    \"diarizationEnabled\" : \"True\"\n",
    "                }\n",
    "            }\n",
    "            \"\"\"\n",
    "    BD = BD.replace(\"__URL__\", URL)  \n",
    "    BD = BD.replace(\"blob_name\", blobname)\n",
    "    return BD"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize HTTP POST request headers "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "headers = {\"Ocp-Apim-Subscription-Key\": SUBSCRIPTION_KEY}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize the connection to our Blob storage account and container. Store the list of blobs in blob_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "blob_service_client = BlobServiceClient.from_connection_string(connect_str)\n",
    "container_client = blob_service_client.get_container_client(container_name)\n",
    "blob_list = container_client.list_blobs()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## For every blob (audio file) in our storage account container call the Speech to Text REST API and transcribe the audio file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\tspeech17.wav\n",
      "{\n",
      "  \"self\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f\",\n",
      "  \"model\": {\n",
      "    \"self\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/models/base/3f61c41d-9ba7-47f4-882d-620a590febfc\"\n",
      "  },\n",
      "  \"links\": {\n",
      "    \"files\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f/files\"\n",
      "  },\n",
      "  \"properties\": {\n",
      "    \"diarizationEnabled\": true,\n",
      "    \"wordLevelTimestampsEnabled\": false,\n",
      "    \"channels\": [\n",
      "      0,\n",
      "      1\n",
      "    ],\n",
      "    \"punctuationMode\": \"Automatic\",\n",
      "    \"profanityFilterMode\": \"Masked\"\n",
      "  },\n",
      "  \"lastActionDateTime\": \"2020-10-12T02:52:52Z\",\n",
      "  \"status\": \"NotStarted\",\n",
      "  \"createdDateTime\": \"2020-10-12T02:52:52Z\",\n",
      "  \"locale\": \"en-US\",\n",
      "  \"displayName\": \"speech17.wav\"\n",
      "}\n",
      "{'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f', 'model': {'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/models/base/3f61c41d-9ba7-47f4-882d-620a590febfc'}, 'links': {'files': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f/files'}, 'properties': {'diarizationEnabled': True, 'wordLevelTimestampsEnabled': False, 'channels': [0, 1], 'punctuationMode': 'Automatic', 'profanityFilterMode': 'Masked'}, 'lastActionDateTime': '2020-10-12T02:52:52Z', 'status': 'NotStarted', 'createdDateTime': '2020-10-12T02:52:52Z', 'locale': 'en-US', 'displayName': 'speech17.wav'}\n",
      "\tspeech18.wav\n",
      "{\n",
      "  \"self\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/88bac45e-207f-4e5e-911d-dbb85dfbb1a2\",\n",
      "  \"model\": {\n",
      "    \"self\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/models/base/3f61c41d-9ba7-47f4-882d-620a590febfc\"\n",
      "  },\n",
      "  \"links\": {\n",
      "    \"files\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/88bac45e-207f-4e5e-911d-dbb85dfbb1a2/files\"\n",
      "  },\n",
      "  \"properties\": {\n",
      "    \"diarizationEnabled\": true,\n",
      "    \"wordLevelTimestampsEnabled\": false,\n",
      "    \"channels\": [\n",
      "      0,\n",
      "      1\n",
      "    ],\n",
      "    \"punctuationMode\": \"Automatic\",\n",
      "    \"profanityFilterMode\": \"Masked\"\n",
      "  },\n",
      "  \"lastActionDateTime\": \"2020-10-12T02:52:53Z\",\n",
      "  \"status\": \"NotStarted\",\n",
      "  \"createdDateTime\": \"2020-10-12T02:52:53Z\",\n",
      "  \"locale\": \"en-US\",\n",
      "  \"displayName\": \"speech18.wav\"\n",
      "}\n",
      "{'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/88bac45e-207f-4e5e-911d-dbb85dfbb1a2', 'model': {'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/models/base/3f61c41d-9ba7-47f4-882d-620a590febfc'}, 'links': {'files': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/88bac45e-207f-4e5e-911d-dbb85dfbb1a2/files'}, 'properties': {'diarizationEnabled': True, 'wordLevelTimestampsEnabled': False, 'channels': [0, 1], 'punctuationMode': 'Automatic', 'profanityFilterMode': 'Masked'}, 'lastActionDateTime': '2020-10-12T02:52:53Z', 'status': 'NotStarted', 'createdDateTime': '2020-10-12T02:52:53Z', 'locale': 'en-US', 'displayName': 'speech18.wav'}\n"
     ]
    }
   ],
   "source": [
    "for blob in blob_list:\n",
    "    print(\"\\t\" + blob.name)\n",
    "    RECORDINGS_BLOB_URI = STORAGE_PATH + blob.name + SAS\n",
    "    URL = RECORDINGS_BLOB_URI\n",
    "    blobname = blob.name\n",
    "    r = requests.post(BaseURL, headers=headers,data=bulkTranscription(URL,blobname))\n",
    "    print(r.text)\n",
    "    js = json.loads(r.text)\n",
    "    print(js)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "headers = {\"Ocp-Apim-Subscription-Key\": SUBSCRIPTION_KEY}\n",
    "r = requests.get(BaseURL, headers=headers)\n",
    "js = json.loads(r.text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'values': [{'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f', 'model': {'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/models/base/3f61c41d-9ba7-47f4-882d-620a590febfc'}, 'links': {'files': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f/files'}, 'properties': {'diarizationEnabled': True, 'wordLevelTimestampsEnabled': False, 'channels': [0, 1], 'punctuationMode': 'Automatic', 'profanityFilterMode': 'Masked'}, 'lastActionDateTime': '2020-10-12T02:52:52Z', 'status': 'NotStarted', 'createdDateTime': '2020-10-12T02:52:52Z', 'locale': 'en-US', 'displayName': 'speech17.wav'}, {'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/88bac45e-207f-4e5e-911d-dbb85dfbb1a2', 'model': {'self': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/models/base/3f61c41d-9ba7-47f4-882d-620a590febfc'}, 'links': {'files': 'https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/88bac45e-207f-4e5e-911d-dbb85dfbb1a2/files'}, 'properties': {'diarizationEnabled': True, 'wordLevelTimestampsEnabled': False, 'channels': [0, 1], 'punctuationMode': 'Automatic', 'profanityFilterMode': 'Masked'}, 'lastActionDateTime': '2020-10-12T02:52:53Z', 'status': 'NotStarted', 'createdDateTime': '2020-10-12T02:52:53Z', 'locale': 'en-US', 'displayName': 'speech18.wav'}]}\n"
     ]
    }
   ],
   "source": [
    "print(js)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## List all transcriptions in either of the following states (Not Started, Running, Succeeded, Failed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def listAllTranscriptions() :\n",
    "    import datetime\n",
    "    df = pd.DataFrame(columns=[\"id\",\"name\", \"status\",\"createdDateTime\",\"lastActionDateTime\",\"CallDuration\"])\n",
    "    headers = {\"Ocp-Apim-Subscription-Key\": SUBSCRIPTION_KEY}\n",
    "    r = requests.get(BaseURL, headers=headers)\n",
    "    js = json.loads(r.text)\n",
    "    #print(js)\n",
    "    for TR in js['values']:\n",
    "        if TR[\"status\"] not in (\"Succeeded\")  :\n",
    "            Dur=\"-\"\n",
    "        else :\n",
    "            Dur = TR[\"properties\"][\"duration\"]\n",
    "        df = df.append({\n",
    "         \"id\":                  TR[\"self\"].split('/')[-1],\n",
    "         \"name\":                TR[\"displayName\"],\n",
    "         \"status\":              TR[\"status\"],\n",
    "         \"createdDateTime\":     datetime.datetime.strptime(TR[\"createdDateTime\"],\"%Y-%m-%dT%H:%M:%SZ\"),            \n",
    "         \"lastActionDateTime\":  datetime.datetime.strptime(TR[\"lastActionDateTime\"],\"%Y-%m-%dT%H:%M:%SZ\"),\n",
    "         \"CallDuration\":            Dur      \n",
    "          }, ignore_index=True)\n",
    "        df[\"ProcTime\"] = df[\"lastActionDateTime\"] - df[\"createdDateTime\"]\n",
    "        \n",
    "    return(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "headers = {\"Ocp-Apim-Subscription-Key\": SUBSCRIPTION_KEY}\n",
    "r = requests.get(BaseURL, headers=headers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>name</th>\n",
       "      <th>status</th>\n",
       "      <th>createdDateTime</th>\n",
       "      <th>lastActionDateTime</th>\n",
       "      <th>CallDuration</th>\n",
       "      <th>ProcTime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>f056993a-7a7d-4f47-b9d8-122c61642f7f</td>\n",
       "      <td>speech17.wav</td>\n",
       "      <td>Succeeded</td>\n",
       "      <td>2020-10-12 02:52:52</td>\n",
       "      <td>2020-10-12 02:54:22</td>\n",
       "      <td>PT49S</td>\n",
       "      <td>00:01:30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>88bac45e-207f-4e5e-911d-dbb85dfbb1a2</td>\n",
       "      <td>speech18.wav</td>\n",
       "      <td>Succeeded</td>\n",
       "      <td>2020-10-12 02:52:53</td>\n",
       "      <td>2020-10-12 02:54:16</td>\n",
       "      <td>PT49S</td>\n",
       "      <td>00:01:23</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     id          name     status  \\\n",
       "0  f056993a-7a7d-4f47-b9d8-122c61642f7f  speech17.wav  Succeeded   \n",
       "1  88bac45e-207f-4e5e-911d-dbb85dfbb1a2  speech18.wav  Succeeded   \n",
       "\n",
       "      createdDateTime  lastActionDateTime CallDuration ProcTime  \n",
       "0 2020-10-12 02:52:52 2020-10-12 02:54:22        PT49S 00:01:30  \n",
       "1 2020-10-12 02:52:53 2020-10-12 02:54:16        PT49S 00:01:23  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "listAllTranscriptions()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "transcriptions = listAllTranscriptions()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Go through each transcription that finished and is in \"Succeeded\" status and obtain the json output of the transcription. Save the json file locally"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "headers = {\n",
    "    # Request headers\n",
    "    'Ocp-Apim-Subscription-Key': SUBSCRIPTION_KEY,\n",
    "}\n",
    "\n",
    "params = urllib.parse.urlencode({\n",
    "    # Request parameters\n",
    "})\n",
    "number_of_rows = len(transcriptions)\n",
    "for index, row in transcriptions.iterrows():\n",
    "    if row[\"status\"] in (\"Succeeded\"):\n",
    "        conn = http.client.HTTPSConnection('westeurope.api.cognitive.microsoft.com')\n",
    "        conn.request(\"GET\", \"/speechtotext/v3.0/transcriptions/\" + row[\"id\"] + \"/files?%s\" % params, \"{body}\", headers)\n",
    "        response = conn.getresponse()\n",
    "        data = response.read()\n",
    "        contentUrl = json.loads(data)['values'][0]['links']['contentUrl']\n",
    "        r = requests.get(contentUrl, headers=headers)\n",
    "        json_text = json.loads(r.text)\n",
    "        file_name = row[\"name\"].split('.')[0] + \".json\"\n",
    "        with open(file_name, 'w') as f:\n",
    "            json.dump(json_text, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Obtain the transcription json of a specific transcription id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "b'{\\r\\n  \"values\": [\\r\\n    {\\r\\n      \"self\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f/files/5136e2e5-6e69-43ed-ade8-43ecdee3ada2\",\\r\\n      \"name\": \"contenturl_0.json\",\\r\\n      \"kind\": \"Transcription\",\\r\\n      \"properties\": {\\r\\n        \"size\": 46869\\r\\n      },\\r\\n      \"createdDateTime\": \"2020-10-12T02:54:22Z\",\\r\\n      \"links\": {\\r\\n        \"contentUrl\": \"https://spsvcprodweu.blob.core.windows.net/bestor-acc02701-cb45-42dc-9d32-2843911017ca/TranscriptionData/5136e2e5-6e69-43ed-ade8-43ecdee3ada2.json?sv=2019-07-07&sr=b&sig=JCMmJCWxPDolOVfqvHf4oRYhvBLCsoIb%2Bdb4cv1SJeY%3D&st=2020-10-12T03:05:17Z&se=2020-10-12T15:10:17Z&sp=rl\"\\r\\n      }\\r\\n    },\\r\\n    {\\r\\n      \"self\": \"https://westeurope.api.cognitive.microsoft.com/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f/files/236994e4-1579-406d-96dc-3277bb1d1d85\",\\r\\n      \"name\": \"report.json\",\\r\\n      \"kind\": \"TranscriptionReport\",\\r\\n      \"properties\": {\\r\\n        \"size\": 510\\r\\n      },\\r\\n      \"createdDateTime\": \"2020-10-12T02:54:22Z\",\\r\\n      \"links\": {\\r\\n        \"contentUrl\": \"https://spsvcprodweu.blob.core.windows.net/bestor-acc02701-cb45-42dc-9d32-2843911017ca/TranscriptionData/236994e4-1579-406d-96dc-3277bb1d1d85.json?sv=2019-07-07&sr=b&sig=WqSi8UzHx5d1xjlW8EeKaNH9SQzKxndBHfVDNZElfD0%3D&st=2020-10-12T03:05:17Z&se=2020-10-12T15:10:17Z&sp=rl\"\\r\\n      }\\r\\n    }\\r\\n  ]\\r\\n}'\n"
     ]
    }
   ],
   "source": [
    "headers = {\n",
    "    # Request headers\n",
    "    'Ocp-Apim-Subscription-Key': SUBSCRIPTION_KEY,\n",
    "}\n",
    "\n",
    "params = urllib.parse.urlencode({\n",
    "    # Request parameters\n",
    "})\n",
    "\n",
    "try:\n",
    "    conn = http.client.HTTPSConnection('westeurope.api.cognitive.microsoft.com')\n",
    "    conn.request(\"GET\", \"/speechtotext/v3.0/transcriptions/f056993a-7a7d-4f47-b9d8-122c61642f7f/files?%s\" % params, \"{body}\", headers)\n",
    "    response = conn.getresponse()\n",
    "    data = response.read()\n",
    "    print(data)\n",
    "    conn.close()\n",
    "except Exception as e:\n",
    "    print(\"[Errno {0}] {1}\".format(e.errno, e.strerror))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Truncate the transcriptions log "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def deleteAllTranscriptions() :\n",
    "    trs=listAllTranscriptions()\n",
    "    for x in trs[\"id\"] :\n",
    "        print(x)\n",
    "        headers = {\"Ocp-Apim-Subscription-Key\": SUBSCRIPTION_KEY}\n",
    "        r = requests.delete(BaseURL+x, headers=headers)\n",
    "        print(r)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "deleteAllTranscriptions() "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Obtain sentiment scores using Text Analytics for each of the utterances from the transcription"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "headers = {\"Ocp-Apim-Subscription-Key\": SUBSCRIPTION_KEY}\n",
    "r = requests.get(json.loads(data)['values'][0]['links']['contentUrl'], headers=headers)\n",
    "json_text = json.loads(r.text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"Asset, why can't that reflect on be about your overall financial condition? It's. Iona genuine Premier. I think I mean I get 2 answers that one is my first answer is if it's a statement about a single asset. It's just not a statement about overall financial condition category categorically. I said I would win the Lotto lottery, go broke too, so that's my first answer. What's your second? Second answer is the one I gave to justice leader, which is that if one. The court rejected the first answer. One could say that you'd look to whether a reasonable person, in context, would view the statement as being about one's overall financial situation. Depends in context, doesn't it? It went to the bank, says Schmidt. You are broke? Are you kidding, says Schmidt. I have a genuine Verma.\""
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "json_text['combinedRecognizedPhrases'][0]['display']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "TEXT_SUBSCRIPTION_KEY = \"f920698ffb974dbd9adf129775be9206\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "text_headers = {\n",
    "    'Content-Type': 'application/json',\n",
    "    'Ocp-Apim-Subscription-Key': TEXT_SUBSCRIPTION_KEY,\n",
    "}\n",
    "\n",
    "text_params = urllib.parse.urlencode({\n",
    "    # Request parameters\n",
    "})\n",
    "\n",
    "docs = {\"documents\": [\n",
    "           {\n",
    "               \"language\": \"en\",\n",
    "               \"id\" : \"1\",\n",
    "               \"text\": json_text['combinedRecognizedPhrases'][0]['display']\n",
    "               \n",
    "           }\n",
    "]}\n",
    "\n",
    "\n",
    "try:\n",
    "    conn = http.client.HTTPSConnection('westeurope.api.cognitive.microsoft.com')\n",
    "    conn.request(\"POST\", \"/text/analytics/v3.1-preview.1/sentiment?%s\" % text_params, json.dumps(docs), text_headers)\n",
    "    response = conn.getresponse()\n",
    "    data = response.read()\n",
    "    \n",
    "    conn.close()\n",
    "except Exception as e:\n",
    "    print(\"[Errno {0}] {1}\".format(e.errno, e.strerror))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'documents': [{'id': '1',\n",
       "   'sentiment': 'mixed',\n",
       "   'confidenceScores': {'positive': 0.37, 'neutral': 0.03, 'negative': 0.6},\n",
       "   'sentences': [{'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.08, 'neutral': 0.86, 'negative': 0.06},\n",
       "     'offset': 0,\n",
       "     'length': 75,\n",
       "     'text': \"Asset, why can't that reflect on be about your overall financial condition?\"},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.13, 'neutral': 0.83, 'negative': 0.04},\n",
       "     'offset': 76,\n",
       "     'length': 5,\n",
       "     'text': \"It's.\"},\n",
       "    {'sentiment': 'positive',\n",
       "     'confidenceScores': {'positive': 0.93, 'neutral': 0.07, 'negative': 0.0},\n",
       "     'offset': 82,\n",
       "     'length': 21,\n",
       "     'text': 'Iona genuine Premier.'},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.41, 'neutral': 0.58, 'negative': 0.01},\n",
       "     'offset': 104,\n",
       "     'length': 103,\n",
       "     'text': \"I think I mean I get 2 answers that one is my first answer is if it's a statement about a single asset.\"},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.01, 'neutral': 0.65, 'negative': 0.34},\n",
       "     'offset': 208,\n",
       "     'length': 83,\n",
       "     'text': \"It's just not a statement about overall financial condition category categorically.\"},\n",
       "    {'sentiment': 'negative',\n",
       "     'confidenceScores': {'positive': 0.0, 'neutral': 0.0, 'negative': 1.0},\n",
       "     'offset': 292,\n",
       "     'length': 78,\n",
       "     'text': \"I said I would win the Lotto lottery, go broke too, so that's my first answer.\"},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.07, 'neutral': 0.89, 'negative': 0.04},\n",
       "     'offset': 371,\n",
       "     'length': 19,\n",
       "     'text': \"What's your second?\"},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.0, 'neutral': 0.99, 'negative': 0.01},\n",
       "     'offset': 391,\n",
       "     'length': 72,\n",
       "     'text': 'Second answer is the one I gave to justice leader, which is that if one.'},\n",
       "    {'sentiment': 'negative',\n",
       "     'confidenceScores': {'positive': 0.0, 'neutral': 0.0, 'negative': 1.0},\n",
       "     'offset': 464,\n",
       "     'length': 36,\n",
       "     'text': 'The court rejected the first answer.'},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.15, 'neutral': 0.78, 'negative': 0.07},\n",
       "     'offset': 501,\n",
       "     'length': 148,\n",
       "     'text': \"One could say that you'd look to whether a reasonable person, in context, would view the statement as being about one's overall financial situation.\"},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.01, 'neutral': 0.97, 'negative': 0.02},\n",
       "     'offset': 650,\n",
       "     'length': 31,\n",
       "     'text': \"Depends in context, doesn't it?\"},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.01, 'neutral': 0.98, 'negative': 0.01},\n",
       "     'offset': 682,\n",
       "     'length': 34,\n",
       "     'text': 'It went to the bank, says Schmidt.'},\n",
       "    {'sentiment': 'negative',\n",
       "     'confidenceScores': {'positive': 0.0, 'neutral': 0.0, 'negative': 1.0},\n",
       "     'offset': 717,\n",
       "     'length': 14,\n",
       "     'text': 'You are broke?'},\n",
       "    {'sentiment': 'neutral',\n",
       "     'confidenceScores': {'positive': 0.02, 'neutral': 0.97, 'negative': 0.01},\n",
       "     'offset': 732,\n",
       "     'length': 30,\n",
       "     'text': 'Are you kidding, says Schmidt.'},\n",
       "    {'sentiment': 'positive',\n",
       "     'confidenceScores': {'positive': 0.94, 'neutral': 0.05, 'negative': 0.01},\n",
       "     'offset': 763,\n",
       "     'length': 23,\n",
       "     'text': 'I have a genuine Verma.'}],\n",
       "   'warnings': []}],\n",
       " 'errors': [],\n",
       " 'modelVersion': '2020-04-01'}"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "json.loads(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.7.4 64-bit (conda)",
   "language": "python",
   "name": "python37464bitconda92af5ceb5c3446b4bef67bc2a88383b5"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}