Update functions.py

functions.py

@@ -145,9 +145,10 @@ def get_yt_audio(url):
     '''Get YT video from given URL link'''
     yt = YouTube(url)
 
+    title = yt.title
+
     # Get the first available audio stream and download it
     audio_stream =  yt.streams.filter(progressive=True, file_extension='mp4').order_by('resolution').desc().first().download()
-    title = audio_stream.split('\\')[-1].split('.')[0]
 
     return audio_stream, title
 
@@ -160,6 +161,101 @@ def load_whisper_api(audio):
 
     return transcript
 
+def inference(link, upload, _asr_model):
+    '''Convert Youtube video or Audio upload to text'''
+    
+    try:
+        
+        if validators.url(link):
+
+            st.info("`Downloading YT audio...`")
+            
+            audio_file, title = get_yt_audio(link)
+
+            if 'audio' not in st.session_state:
+                st.session_state['audio'] = audio_file
+    
+            #Get size of audio file
+            audio_size = round(os.path.getsize(audio_file)/(1024*1024),1)
+    
+            #Check if file is > 24mb, if not then use Whisper API
+            if audio_size <= 25:
+
+                st.info("`Transcribing YT audio...`")
+                
+                #Use whisper API
+                results = load_whisper_api(audio_file)['text']
+    
+            else:
+    
+                st.warning('File size larger than 24mb, applying chunking and transcription',icon="⚠️")
+    
+                song = AudioSegment.from_file(audio_file, format='mp4')
+    
+                # PyDub handles time in milliseconds
+                twenty_minutes = 20 * 60 * 1000
+                
+                chunks = song[::twenty_minutes]
+                
+                transcriptions = []
+                
+                for i, chunk in enumerate(chunks):
+                    chunk.export(f'output/chunk_{i}.mp4', format='mp4')
+                    transcriptions.append(load_whisper_api(f'output/chunk_{i}.mp4')['text'])
+    
+                results = ','.join(transcriptions)
+
+            st.info("`YT Video transcription process complete...`")
+
+            return results, title
+
+        elif _upload:
+
+            #Get size of audio file
+            audio_size = round(os.path.getsize(_upload)/(1024*1024),1)
+    
+            #Check if file is > 24mb, if not then use Whisper API
+            if audio_size <= 25:
+
+                st.info("`Transcribing uploaded audio...`")
+                
+                #Use whisper API
+                results = load_whisper_api(_upload)['text']
+    
+            else:
+    
+                st.write('File size larger than 24mb, applying chunking and transcription')
+    
+                song = AudioSegment.from_file(_upload)
+    
+                # PyDub handles time in milliseconds
+                twenty_minutes = 20 * 60 * 1000
+                
+                chunks = song[::twenty_minutes]
+                
+                transcriptions = []
+
+                st.info("`Transcribing uploaded audio...`")
+                
+                for i, chunk in enumerate(chunks):
+                    chunk.export(f'output/chunk_{i}.mp3', format='mp3')
+                    transcriptions.append(load_whisper_api('output/chunk_{i}.mp3')['text'])
+    
+                results = ','.join(transcriptions)
+
+            st.info("`Uploaded audio transcription process complete...`")
+
+            return results, "Transcribed Earnings Audio"
+                
+    except Exception as e:
+
+        st.error(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')
+      
+        return results['text'], title
+
 @st.cache_data
 def process_corpus(corpus, title, embedding_model, chunk_size=1000, overlap=50):
 
@@ -299,91 +395,6 @@ def get_spacy():
     nlp = en_core_web_lg.load()
     return nlp
     
-
-def inference(link, upload, _asr_model):
-    '''Convert Youtube video or Audio upload to text'''
-    
-    try:
-        
-        if validators.url(link):
-        
-            audio_file, title = get_yt_audio(link)
-            # title = yt.title
-            
-            if 'audio' not in st.session_state:
-                st.session_state['audio'] = audio_file
-    
-            #Get size of audio file
-            audio_size = round(os.path.getsize(audio_file)/(1024*1024),1)
-    
-            #Check if file is > 24mb, if not then use Whisper API
-            if audio_size <= 25:
-    
-                #Use whisper API
-                results = load_whisper_api(audio_file)['text']
-    
-            else:
-    
-                st.warning('File size larger than 24mb, applying chunking and transcription',icon="⚠️")
-    
-                song = AudioSegment.from_file(audio_file, format='mp3')
-    
-                # PyDub handles time in milliseconds
-                twenty_minutes = 20 * 60 * 1000
-                
-                chunks = song[::twenty_minutes]
-                
-                transcriptions = []
-                
-                for i, chunk in enumerate(chunks):
-                    chunk.export(f'output/chunk_{i}.mp3', format='mp3')
-                    transcriptions.append(load_whisper_api(f'output/chunk_{i}.mp3')['text'])
-    
-                results = ','.join(transcriptions)
-
-            return results, title
-
-        elif _upload:
-
-            #Get size of audio file
-            audio_size = round(os.path.getsize(_upload)/(1024*1024),1)
-    
-            #Check if file is > 24mb, if not then use Whisper API
-            if audio_size <= 25:
-    
-                #Use whisper API
-                results = load_whisper_api(_upload)['text']
-    
-            else:
-    
-                st.write('File size larger than 24mb, applying chunking and transcription')
-    
-                song = AudioSegment.from_file(_upload)
-    
-                # PyDub handles time in milliseconds
-                twenty_minutes = 20 * 60 * 1000
-                
-                chunks = song[::twenty_minutes]
-                
-                transcriptions = []
-                
-                for i, chunk in enumerate(chunks):
-                    chunk.export(f'output/chunk_{i}.mp3', format='mp3')
-                    transcriptions.append(load_whisper_api('output/chunk_{i}.mp3')['text'])
-    
-                results = ','.join(transcriptions)
-
-            return results, "Transcribed Earnings Audio"
-                
-    except Exception as e:
-
-        st.error(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')
-      
-        return results['text'], title
-    
       
 @st.cache_data
 def sentiment_pipe(earnings_text):