Update app.py

app.py

@@ -1,3 +1,5 @@
+# app.py
+
 import streamlit as st
 import time
 import re
@@ -5,28 +7,41 @@ import os
 import tempfile
 import pypdf
 from pydub import AudioSegment
+
 from utils import (
-    generate_script, 
-    generate_audio_mp3, 
-    truncate_text, 
-    extract_text_from_url, 
-    transcribe_youtube_video, 
+    generate_script,
+    generate_audio_mp3,
+    truncate_text,
+    extract_text_from_url,
+    transcribe_youtube_video,
     research_topic
 )
 from prompts import SYSTEM_PROMPT
 
+
 def parse_user_edited_transcript(edited_text: str):
+    """
+    Looks for lines like:
+        **Jane**: Hello
+        **John**: Sure, I'd love to talk about that.
+    Returns a list of (speaker, text).
+    """
     pattern = r"\*\*(Jane|John)\*\*:\s*(.+)"
     matches = re.findall(pattern, edited_text)
     if not matches:
-        # If the user completely changed the format, just treat it as "Jane"
+        # If user changed the format drastically, treat entire text as Jane
         return [("Jane", edited_text)]
     return matches
 
+
 def regenerate_audio_from_dialogue(dialogue_items):
+    """
+    Re-generates multi-speaker audio from user-edited text.
+    Returns final_mp3_path, updated_transcript.
+    """
     audio_segments = []
     transcript = ""
-    crossfade_duration = 50  # ms
+    crossfade_duration = 50  # in ms
 
     for speaker, line_text in dialogue_items:
         audio_file = generate_audio_mp3(line_text, speaker)
@@ -38,6 +53,7 @@ def regenerate_audio_from_dialogue(dialogue_items):
     if not audio_segments:
         return None, "No audio segments were generated."
 
+    # Combine with crossfade
     combined = audio_segments[0]
     for seg in audio_segments[1:]:
         combined = combined.append(seg, crossfade=crossfade_duration)
@@ -48,7 +64,17 @@ def regenerate_audio_from_dialogue(dialogue_items):
 
     return final_mp3_path, transcript
 
+
 def generate_podcast(file, url, video_url, research_topic_input, tone, length):
+    """
+    Creates a multi-speaker podcast from:
+      - PDF
+      - URL
+      - YouTube video
+      - or a research topic input.
+    Returns (podcast_file_path, transcript_str).
+    """
+    # Ensure only one input source
     sources = [bool(file), bool(url), bool(video_url), bool(research_topic_input)]
     if sum(sources) > 1:
         return None, "Provide only one input (PDF, URL, YouTube, or Research topic)."
@@ -57,6 +83,7 @@ def generate_podcast(file, url, video_url, research_topic_input, tone, length):
 
     text = ""
     if file:
+        # Handle PDF
         try:
             if not file.name.lower().endswith('.pdf'):
                 return None, "Please upload a PDF file."
@@ -65,6 +92,7 @@ def generate_podcast(file, url, video_url, research_topic_input, tone, length):
         except Exception as e:
             return None, f"Error reading PDF: {str(e)}"
     elif url:
+        # Handle URL
         try:
             text = extract_text_from_url(url)
             if not text:
@@ -72,6 +100,7 @@ def generate_podcast(file, url, video_url, research_topic_input, tone, length):
         except Exception as e:
             return None, f"Error extracting text from URL: {str(e)}"
     elif video_url:
+        # Handle YouTube
         try:
             text = transcribe_youtube_video(video_url)
             if not text:
@@ -79,6 +108,7 @@ def generate_podcast(file, url, video_url, research_topic_input, tone, length):
         except Exception as e:
             return None, f"Error transcribing YouTube video: {str(e)}"
     elif research_topic_input:
+        # Handle research topic
         try:
             text = research_topic(research_topic_input)
             if not text:
@@ -86,12 +116,14 @@ def generate_podcast(file, url, video_url, research_topic_input, tone, length):
         except Exception as e:
             return None, f"Error researching topic: {str(e)}"
 
+    # Generate the multi-speaker script
     try:
         text = truncate_text(text)
         script = generate_script(SYSTEM_PROMPT, text, tone, length)
     except Exception as e:
         return None, f"Error generating script: {str(e)}"
 
+    # Convert the script.dialogue to audio
     audio_segments = []
     transcript = ""
     crossfade_duration = 50  # ms
@@ -119,12 +151,26 @@ def generate_podcast(file, url, video_url, research_topic_input, tone, length):
     except Exception as e:
         return None, f"Error generating audio: {str(e)}"
 
+
 def main():
-    st.set_page_config(page_title="MyPod - AI-based Podcast Generator", layout="centered")
+    st.set_page_config(
+        page_title="MyPod - AI-based Podcast Generator",
+        layout="centered"
+    )
 
     st.title("🎙 MyPod - AI-based Podcast Generator")
 
-    # Standard input UI
+    st.markdown(
+        """
+        Welcome to **MyPod**, your AI-powered podcast generator!  
+        Provide a PDF, URL, YouTube link, or enter a research topic,  
+        choose a tone and target length, and click "Generate Podcast."
+        
+        **After** the audio is generated, you can **edit** the transcript 
+        and **re-generate** the audio with your edits if needed.
+        """
+    )
+
     col1, col2 = st.columns(2)
     with col1:
         file = st.file_uploader("Upload PDF (.pdf only)", type=["pdf"])
@@ -138,48 +184,43 @@ def main():
     generate_button = st.button("Generate Podcast")
 
     if generate_button:
-        # Show a custom progress bar & messages
+        # Show a pseudo progress bar for user engagement
         progress_bar = st.progress(0)
         progress_text = st.empty()
 
-        # We don't have real-time progress from TTS, so we'll do a pseudo countdown:
-        # Step 1: Start
+        # Steps to pretend some progress:
         progress_text.write("Alright, let's get started...")
         progress_bar.progress(10)
-        time.sleep(1.5)
+        time.sleep(1.0)
 
-        # Step 2: Some cheeky text
         progress_text.write("Working on the magic. Hang tight!")
         progress_bar.progress(40)
-        time.sleep(1.5)
+        time.sleep(1.0)
 
-        # Step 3: Almost there
         progress_text.write("Almost done. Adding a dash of awesomeness...")
         progress_bar.progress(70)
-        time.sleep(1.5)
+        time.sleep(1.0)
 
-        # Step 4: Actually generate
+        # Actual generation
         podcast_file, transcript = generate_podcast(
             file, url, video_url, research_topic_input, tone, length
         )
-        time.sleep(1.5)
+
+        time.sleep(1.0)
+        progress_bar.progress(100)
+        progress_text.write("Done!")
 
         if podcast_file is None:
-            # Reset progress to 0 if error
-            progress_bar.progress(100)
             st.error(transcript)
-            return
         else:
-            progress_bar.progress(100)
-            progress_text.write("Done!")
-
             st.success("Podcast generated successfully!")
+            # Show the audio
             audio_file = open(podcast_file, 'rb')
             audio_bytes = audio_file.read()
             audio_file.close()
             st.audio(audio_bytes, format='audio/mp3')
 
-            # Show transcript in an editable box
+            # Show the transcript in a text area for editing
             st.markdown("### Generated Transcript (Editable)")
             edited_text = st.text_area(
                 "Feel free to tweak lines, fix errors, or reword anything.",
@@ -187,8 +228,8 @@ def main():
                 height=300
             )
 
+            # Regenerate button
             if st.button("Regenerate Audio From Edited Text"):
-                # Another pseudo progress bar for regeneration
                 regen_bar = st.progress(0)
                 regen_text = st.empty()
 
@@ -200,7 +241,7 @@ def main():
                 regen_bar.progress(60)
                 time.sleep(1.0)
 
-                # Parse and regenerate
+                # Parse & regenerate
                 dialogue_items = parse_user_edited_transcript(edited_text)
                 new_audio_path, new_transcript = regenerate_audio_from_dialogue(dialogue_items)
 
@@ -227,5 +268,6 @@ def main():
             # Clean up the original file
             os.remove(podcast_file)
 
+
 if __name__ == "__main__":
     main()