Merge branch 'developer' into saurabh-dev

.slugignore

@@ -0,0 +1,5 @@
+test/*
+source/punct_model_part1.pcl
+source/punct_model_part2.pcl
+source/punct_model_part3.pcl
+source/punct_model_full.pcl

Procfile

@@ -1 +1 @@
-web: sh setup.sh && streamlit run source/scrivener_user_interface.py
+web: sh setup.sh &&  python -c "import nltk; nltk.download('punkt')" && streamlit run source/scrivener_user_interface.py

requirements.txt

@@ -6,4 +6,5 @@ pytube==11.0.1
 SpeechRecognition==3.8.1
 streamlit==0.89.0
 sumy
-punctuator
+punctuator==0.9.6
+wget

source/main/punctuation.py

@@ -1,5 +1,11 @@
+import os
+try:
+    # hack: Unfortunately Heroku uses below file and it conflicts with the Punctuator package
+    # as suggested here: https://github.com/chrisspen/punctuator2/issues/3 removing the file
+    os.remove('.heroku/python/bin/punctuator.py')
+except:
+    print("punctuator.py not found in: " + os.getcwd())
 from punctuator import Punctuator
-import os.path
 
 
 class Punctuation:

source/scrivener_user_interface.py

@@ -9,6 +9,7 @@
 import streamlit as st
 import re
 import os
+import wget
 from main.transcribe import TranscribeVideo
 from main.transcribe_yt import TranscribeYtVideo
 import secrets
@@ -62,7 +63,29 @@
 if not os.path.exists('source/punct_model_full.pcl'):
     print("Creating punct_model_full.pcl file for ML model...")
 
+
     # Path to model files parts that needs to be combined
+    # Storing these models in github causes an issue with the Heroku deployment and exceeds 500 MB (it is 618 MB)
+    # slug/payload limit. Therefore, using this alternative to get it from Github during runtime.
+    if not os.path.exists('source/punct_model_part1.pcl'):
+        print("Downloading punct_model_part1.pcl file for ML model...")
+        url1 = 'https://github.com/SN-18/scrivener/raw/developer/source/punct_model_part1.pcl'
+        filename = wget.download(url1, out='source/punct_model_part1.pcl')
+        print("\nDownloaded file: " + filename)
+
+    if not os.path.exists('source/punct_model_part2.pcl'):
+        print("Downloading punct_model_part2.pcl file for ML model...")
+        url2 = 'https://github.com/SN-18/scrivener/raw/developer/source/punct_model_part2.pcl'
+        filename = wget.download(url2, out='source/punct_model_part2.pcl')
+        print("\nDownloaded file: " + filename)
+
+    if not os.path.exists('source/punct_model_part3.pcl'):
+        print("Downloading punct_model_part3.pcl file for ML model...")
+        url3 = 'https://github.com/SN-18/scrivener/raw/developer/source/punct_model_part3.pcl'
+        filename = wget.download(url3, out='source/punct_model_part3.pcl')
+        print("\nDownloaded file: " + filename)
+
+
     first_file = os.path.abspath('source/punct_model_part1.pcl')
     second_file = os.path.abspath('source/punct_model_part2.pcl')
     third_file = os.path.abspath('source/punct_model_part3.pcl')