Update functions.py

functions.py

@@ -217,7 +217,7 @@ def gen_embeddings(embedding_model):
     return embeddings
     
 @st.cache_data
-def embed_text(query,title,embedding_model,_docsearch):
+def embed_text(query,embedding_model,_docsearch):
     
     '''Embed text and generate semantic search scores'''
 
@@ -230,8 +230,6 @@ def embed_text(query,title,embedding_model,_docsearch):
                           temperature=0
                          )
 
-    title = title.split()[0].lower()
-
     chain = ConversationalRetrievalChain.from_llm(chat_llm, 
                                                   retriever= _docsearch.as_retriever(),
                                                   qa_prompt = load_prompt(),
@@ -304,8 +302,7 @@ def inference(link, upload, _asr_model):
         
             audio_file = get_yt_audio(link)
             # title = yt.title
-            print(audio_file)
-            print(len(audio_file))
+            
             if 'audio' not in st.session_state:
                 st.session_state['audio'] = audio_file
     
@@ -317,13 +314,12 @@ def inference(link, upload, _asr_model):
     
                 #Use whisper API
                 results = load_whisper_api(audio_file)['text']
-                print(results)
     
             else:
     
-                st.write('File size larger than 24mb, applying chunking and transcription')
+                st.warning('File size larger than 24mb, applying chunking and transcription',icon="⚠️")
     
-                song = AudioSegment.from_file(audio_file, format='mp4')
+                song = AudioSegment.from_file(audio_file, format='mp3')
     
                 # PyDub handles time in milliseconds
                 twenty_minutes = 20 * 60 * 1000
@@ -333,8 +329,8 @@ def inference(link, upload, _asr_model):
                 transcriptions = []
                 
                 for i, chunk in enumerate(chunks):
-                    chunk.export(f'output/chunk_{i}.mp4', format='mp4')
-                    transcriptions.append(load_whisper_api('output/chunk_{i}.mp4')['text'])
+                    chunk.export(f'output/chunk_{i}.mp3', format='mp3')
+                    transcriptions.append(load_whisper_api('output/chunk_{i}.mp3')['text'])
     
                 results = ','.join(transcriptions)
 
@@ -365,8 +361,8 @@ def inference(link, upload, _asr_model):
                 transcriptions = []
                 
                 for i, chunk in enumerate(chunks):
-                    chunk.export(f'output/chunk_{i}.mp4', format='mp4')
-                    transcriptions.append(load_whisper_api('output/chunk_{i}.mp4')['text'])
+                    chunk.export(f'output/chunk_{i}.mp3', format='mp3')
+                    transcriptions.append(load_whisper_api('output/chunk_{i}.mp3')['text'])
     
                 results = ','.join(transcriptions)
 
@@ -374,8 +370,8 @@ def inference(link, upload, _asr_model):
                 
     except Exception as e:
 
-        st.write(f'''Whisper API Error: {e}, 
-                    Using Whisper module from GitHub, might take longer than expected''')
+        st.warning(f'''Whisper API Error: {e}, 
+                    Using Whisper module from GitHub, might take longer than expected''',icon="⚠️")
         
         results = _asr_model.transcribe(st.session_state['audio'], task='transcribe', language='en')