Merge pull request #2 from kumarks1122/main

Audio file and config file uploading feature and invalid option audio playing functionality added

.github/workflows/nodejs-docker-image.yml

@@ -0,0 +1,23 @@
+name: Docker Image CI
+
+on:
+  push:
+    branches: [ "main" ]
+  pull_request:
+    branches: [ "main" ]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+    - name: Build the Docker image
+      run: cd nodejs/voice-streaming && docker build . --file Dockerfile --tag hyd.ocir.io/ax2cel5zyviy/ekstep-djp/ivrs:latest
+    - name: Log in to Docker Hub
+      uses: docker/login-action@v1
+      with:
+        registry: hyd.ocir.io
+        username: ${{ secrets.DOCKER_USERNAME }}
+        password: ${{ secrets.DOCKER_PASSWORD }}
+    - name: Push the docker image
+      run: docker push hyd.ocir.io/ax2cel5zyviy/ekstep-djp/ivrs:latest

.gitignore

@@ -1,10 +1,13 @@
 **/node_modules
+**/bruno
 venv/**
 **/.DS_Store
 output.wav
 .vscode
 **/__pycache__
 temp_audio_files/**
+drive-files/**
 temp_converted_files/**
 !temp_audio_files/.keep
-!temp_converted_files/.keep
+!temp_converted_files/.keep
+!drive-files/.keep

README.md

@@ -1 +1,50 @@
-# ivrs-bot
+# MyJP IVRS Bot
+
+## Steps to prerequisites:
+
+### Step 1
+Install python 3.11
+
+### Step 2
+Install dependency libraries with below command
+```
+pip install -r requirements.txt
+```
+
+## Steps To Upload audio files to cloud
+
+### Step 1
+
+Get the audio files and place the audio files in the `drive-files` folder
+
+### Step 2
+
+Update the config file `config.yaml` with cloud storage credentials in below format
+
+```
+region_name: ""
+aws_secret_access_key: ""
+aws_access_key_id: ""
+endpoint_url: ""
+
+```
+
+### Step 3
+
+Create the csv file in below format.
+```
+filename,category,language
+"32-Track 1 Bangla PoemPoem Rhymes for eJadui Pitara.mp3",song,bengali
+"33-Track 2 Bangla Poem Rhymes for eJadui Pitara.mp3",song,bengali
+```
+
+### Step 4
+
+Run the below command to upload the audio files and update the ivrs_config.json file
+
+```
+python generate-config.py csv_file_name
+
+# Example
+python generate-config.py audios.csv
+```

app.py

@@ -8,7 +8,6 @@
 
 app = Flask(__name__)
 sockets = Sockets(app)
-connection_chunks = json.loads('{}')
 
 from api import api
 app.register_blueprint(api)

audio_socket.py

@@ -2,7 +2,7 @@
 from pydub.utils import make_chunks
 from flask import Blueprint
 from datetime import datetime
-from urllib.parse import parse_qs
+from urllib import parse
 from telemetry import Telemetry
 
 import os
@@ -17,42 +17,36 @@
 
 audio_socket = Blueprint('audio_socket', __name__)
 duration = 20 # no of milliseconds of each base64 string from audio file
-audio_types = ["story", "rhyme", "riddle"]
+audio_types = ["story", "song", "riddle"]
 
 AUDIO_CACHE = {}
 
-def get_audio(input_selector, language):
-
-    if input_selector > 2:
-        return None
+def get_audio(audio_key):
 
     urlData = os.environ['IVRS_CONFIG_URL']
     webURL = downloader.urlopen(urlData)
     data = webURL.read()
     encoding = webURL.info().get_content_charset('utf-8')
     config = json.loads(data.decode(encoding))
 
-    audio_type = audio_types[input_selector]
-    no_of_audios = len(config[f"{audio_type}:{language}"])
+    if audio_key not in config:
+        return None
+
+    no_of_audios = len(config[audio_key])
 
     day_of_year = datetime.today().timetuple().tm_yday
     mod_day_no = int(day_of_year % no_of_audios)
 
     audio_index = no_of_audios if mod_day_no == 0 else mod_day_no
 
-    return config[f"{audio_type}:{language}"][audio_index - 1]
+    return config[audio_key][audio_index - 1]
 
-def get_chunks(input_selector, language, file_path):
-    audio_type = audio_types[input_selector]
+def get_chunks(audio_key, file_path):
 
     day_of_year = datetime.today().timetuple().tm_yday
-    path = Path(file_path)
-    filename = path.stem + path.suffix.split("?")[0]
-
-    cache_key = f"{audio_type}:{language}"
 
-    if cache_key in AUDIO_CACHE:
-        chunk_detail = AUDIO_CACHE[cache_key]
+    if audio_key in AUDIO_CACHE:
+        chunk_detail = AUDIO_CACHE[audio_key]
 
         if chunk_detail['cached_on'] == day_of_year:
             return chunk_detail['chunks']
@@ -84,7 +78,7 @@ def get_chunks(input_selector, language, file_path):
             }
         })
 
-    AUDIO_CACHE[cache_key] = {'cached_on': day_of_year, 'chunks': chunks_array}
+    AUDIO_CACHE[audio_key] = {'cached_on': day_of_year, 'chunks': chunks_array}
 
     remove_temp_file(local_file_path)
     remove_temp_file(local_converted_file_path)
@@ -111,6 +105,7 @@ def echo(ws, language):
         event = request_payload['event']
 
         if event == 'start':
+            print("inside start")
             did = hashlib.md5(request_payload['start']['from'].encode()).hexdigest()
             telemetry = Telemetry(request_payload['stream_sid'], did)
             request_payload['start']['from'] = did
@@ -127,11 +122,26 @@ def echo(ws, language):
 
             input_selector = int(request_payload["dtmf"]["digit"]) - 1
 
-            audio_url = get_audio(input_selector, language)
             selected_audio_type = audio_types[input_selector] if input_selector < len(audio_types) else None
+
+            audio_key = f"{selected_audio_type}:{language}"
+            audio_url = None
+            if selected_audio_type:
+                audio_url = get_audio(audio_key)
+
+                if not audio_url:
+                    audio_key = f"{selected_audio_type}:{language}:empty"
+                    audio_url = get_audio(audio_key)
+
+            if not audio_url:
+                audio_key = f"invalid_option:{language}"
+                audio_url = get_audio(audio_key)
+
             telemetry.interact(input=input_selector, language=language, audio_type=selected_audio_type,audio_name=audio_url)
+
             if audio_url:
-                chunks = get_chunks(input_selector, language, audio_url)
+                audio_url = audio_url.replace(" ", "%20");
+                chunks = get_chunks(audio_key, audio_url)
                 counter = 1
                 for chunk in chunks:
                     chunk["stream_sid"] = session_id
@@ -146,6 +156,9 @@ def echo(ws, language):
                 ws.send(json.dumps(mark_event))
                 push_telemetry_events(telemetry)
                 is_audio_sent = True
+            else:
+                pass
+
         elif event == "mark":
             # mark_event = {"event":"stop", "sequence_number": len(chunks) + 1, "stream_sid": session_id,"mark":{"name":"audio_complete"}}
             # ws.send(json.dumps(mark_event))

config.yaml

@@ -0,0 +1,4 @@
+region_name: ""
+aws_secret_access_key: ""
+aws_access_key_id: ""
+endpoint_url: ""

generate-config.py

@@ -0,0 +1,93 @@
+import sys
+import csv
+import re
+import json
+# import gdown
+import boto3
+import yaml
+
+
+from pathlib import Path
+
+def get_upload_path(file_name, category, language):
+    if category != "" and language != "": 
+        return f'audio/{category}/{language}/{file_name}'
+    else:
+        return f'audio/{file_name}'
+
+def upload_to_oci_storage(s3, file_name, category, language):
+    try:
+        upload_path = get_upload_path(file_name, category, language)
+        s3.meta.client.upload_file(f"drive-files/{file_name}", 'sbdjp-ivrs', upload_path)
+        print(upload_path, "uploaded successfully")
+    except Exception as e:
+        print(e)
+
+def create_config(s3):
+    config = {}
+    categories = set()
+    languages = set()
+
+    invalid_option_link = ["https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/invalid_option_english.wav"]
+
+    for obj in s3.Bucket('sbdjp-ivrs').objects.all():
+        path = Path(obj.key)
+
+        if len(path.parts) < 4: continue
+
+        if path.parts[0] != "audio": continue
+
+        category = path.parts[1].strip().lower()
+        language = path.parts[2].strip().lower()
+
+        categories.add(category)
+        languages.add(language)
+        audio_key = f"{category}:{language}"
+        if audio_key not in config: config[audio_key] = []
+
+        url = s3.meta.client.generate_presigned_url(ClientMethod = 'get_object', Params = { 'Bucket': 'sbdjp-ivrs', 'Key': obj.key })
+        url = url.split("?")[0]
+
+        config[audio_key].append(url)
+
+    for lang in languages:
+        for cat in categories:
+            audio_key = f"{cat}:{lang}:empty"
+            config[audio_key] = invalid_option_link
+
+        audio_key = f"invalid_option:{lang}"
+        config[audio_key] = invalid_option_link
+
+    with open('drive-files/ivrs_config.json', 'w') as f:
+        f.write(json.dumps(config))
+
+if __name__ == '__main__':
+    n = len(sys.argv)
+    if n < 1:
+        print('Usage: generate-config.py <csv file>')
+        exit(1)
+
+    with open("config.yaml", "r") as stream:
+        config = yaml.safe_load(stream)
+        s3 = boto3.resource(
+            's3',
+            region_name=config["region_name"],
+            aws_secret_access_key=config["aws_secret_access_key"],
+            aws_access_key_id=config["aws_access_key_id"],
+            endpoint_url=config["endpoint_url"]
+        )
+
+    with open(sys.argv[1]) as csv_file:
+        files_list = [d for d in csv.DictReader(csv_file)]
+
+        for file in files_list:
+            category = file["category"].strip().lower()
+            language = file["language"].strip().lower()
+            try:
+                upload_to_oci_storage(s3, file["filename"], category, language)
+            except Exception as inst:
+                print(inst)
+
+    create_config(s3)
+    upload_to_oci_storage(s3, "ivrs_config.json", "", "")
+

ivrs_config.json

@@ -1,25 +1,21 @@
 {
-  "story:english":[
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/story/English/Fur-Feather-Africa.mp3",
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/story/English/Windswept.mp3"
+  "song:english": [
+    "https://ax2cel5zyviy.compat.objectstorage.ap-hyderabad-1.oraclecloud.com/sbdjp-ivrs/audio/song/English/100-The%20Swing.mp3"
   ],
-  "rhyme:english":[
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/rhyme/English/rabbit.m4a"
+  "song:hindi": [
+    "https://ax2cel5zyviy.compat.objectstorage.ap-hyderabad-1.oraclecloud.com/sbdjp-ivrs/audio/song/Hindi/39-Track%208%20Hindi%20Poem%20Rhymes%20for%20eJadui%20Pitara%20Pd.mp3"
   ],
-  "riddle:english":[
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/riddle/riddle.mp3"
+  "story:english": [
+    "https://ax2cel5zyviy.compat.objectstorage.ap-hyderabad-1.oraclecloud.com/sbdjp-ivrs/audio/story/English/1-Birbal%20finds%20the%20thief.mp3",
+    "https://ax2cel5zyviy.compat.objectstorage.ap-hyderabad-1.oraclecloud.com/sbdjp-ivrs/audio/story/English/10-Goose%20with%20the%20golden%20egg.mp3"
   ],
-  "story:hindi":[
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/story/Hindi/bhediyabakri.mp3",
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/story/Hindi/pakshiparakh.mp3",
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/story/Hindi/darpokkharghosh.mp3"
+  "story:hindi:empty": [
+    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/invalid_option_english.wav"
   ],
-  "rhyme:hindi":[
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/rhyme/Hindi/hathiraja.m4a",
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/rhyme/Hindi/janmashtami.m4a",
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/rhyme/Hindi/words.m4a"
+  "invalid_option:english": [
+    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/invalid_option_english.wav"
   ],
-  "riddle:hindi":[
-    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/riddle/riddle.mp3"
+  "invalid_option:hindi": [
+    "https://objectstorage.ap-hyderabad-1.oraclecloud.com/n/ax2cel5zyviy/b/sbdjp-ivrs/o/audio/invalid_option_english.wav"
   ]
 }