Update app.py

app.py

@@ -20,12 +20,23 @@ from utils import (
     DialogueItem
 )
 from prompts import SYSTEM_PROMPT
-# Q&A
+
+# NEW: For Q&A
 from qa import transcribe_audio_deepgram, handle_qa_exchange
 
-MAX_QA_QUESTIONS = 5
+MAX_QA_QUESTIONS = 5  # up to 5 voice/text questions
 
 def parse_user_edited_transcript(edited_text: str, host_name: str, guest_name: str):
+    """
+    Looks for lines like:
+        **Angela**: Hello
+        **Dimitris**: Great topic...
+    We treat 'Angela' as the raw display_speaker, 'Hello' as text.
+    Then we map 'Angela' -> speaker='Jane' (if it matches host_name),
+    'Dimitris' -> speaker='John' (if it matches guest_name), etc.
+
+    Returns a list of DialogueItem.
+    """
     pattern = r"\*\*(.+?)\*\*:\s*(.+)"
     matches = re.findall(pattern, edited_text)
 
@@ -60,9 +71,14 @@ def parse_user_edited_transcript(edited_text: str, host_name: str, guest_name: s
     return items
 
 def regenerate_audio_from_dialogue(dialogue_items, custom_bg_music_path=None):
+    """
+    Re-generates multi-speaker audio from user-edited DialogueItems,
+    then mixes with background music or custom music.
+    Returns (audio_bytes, transcript_str).
+    """
     audio_segments = []
     transcript = ""
-    crossfade_duration = 50
+    crossfade_duration = 50  # ms
 
     for item in dialogue_items:
         audio_file = generate_audio_mp3(item.text, item.speaker)
@@ -106,6 +122,12 @@ def generate_podcast(
     sponsor_style,
     custom_bg_music_path
 ):
+    """
+    Creates a multi-speaker podcast from PDF, URL, YouTube, or a research topic.
+    Ensures female voice for host (Jane), male voice for guest (John).
+    Sponsor content is either separate or blended based on sponsor_style.
+    Returns (audio_bytes, transcript_str).
+    """
     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 Topic)."
@@ -143,6 +165,7 @@ def generate_podcast(
         except Exception as e:
             return None, f"Error researching topic: {str(e)}"
 
+    from utils import truncate_text
     text = truncate_text(text)
 
     extra_instructions = []
@@ -154,13 +177,10 @@ def generate_podcast(
     if user_specs.strip():
         extra_instructions.append(f"Additional User Instructions: {user_specs}")
 
-    # If user provided sponsor content, we pass it along; otherwise skip
-    sponsor_instructions_present = False
     if sponsor_content.strip():
         extra_instructions.append(
             f"Sponsor Content Provided (should be under ~30 seconds):\n{sponsor_content}"
         )
-        sponsor_instructions_present = True
 
     from prompts import SYSTEM_PROMPT
     combined_instructions = "\n\n".join(extra_instructions).strip()
@@ -177,7 +197,7 @@ def generate_podcast(
             f"{length_minutes} Mins",
             host_name=host_name or "Jane",
             guest_name=guest_name or "John",
-            sponsor_style=sponsor_style  # If sponsor is empty, no sponsor lines appended
+            sponsor_style=sponsor_style
         )
     except Exception as e:
         return None, f"Error generating script: {str(e)}"
@@ -230,7 +250,7 @@ def highlight_differences(original: str, edited: str) -> str:
 
 def main():
     st.set_page_config(
-        page_title="MyPod v2: AI-based Podcast Generator",
+        page_title="MyPod - AI-based Podcast Generator",
         layout="centered"
     )
 
@@ -238,7 +258,7 @@ def main():
     with logo_col:
         st.image("logomypod.jpg", width=60)
     with title_col:
-        st.markdown("## MyPod v2: AI powered Podcast Generator")
+        st.markdown("## MyPod - AI powered Podcast Generator")
 
     st.markdown(
         "Welcome to **MyPod**, your go-to AI-powered podcast generator! 🎉\n\n"
@@ -246,7 +266,7 @@ def main():
         "conversational podcast.\n"
         "Select a tone and a duration range. The script will be on-topic, concise, and respect your chosen length.\n\n"
         "### How to use:\n"
-        "1. **Provide one source:** PDF Files, Website URL, YouTube link or a Topic to Research.\n"
+        "1. **Provide one source:** PDF Files, Website URL, YouTube videos, or a Topic to Research.\n"
         "2. **Choose the tone and the target duration.**\n"
         "3. **Click 'Generate Podcast'** to produce your podcast. After the audio is generated, "
         "   you can edit the transcript and re-generate the audio with your edits if needed.\n\n"
@@ -271,14 +291,17 @@ def main():
     st.markdown("### Customize Your Podcast (New Features)")
 
     with st.expander("Set Host & Guest Names/Descriptions (Optional)"):
-        host_name = st.text_input("Female Host Name (leave blank for 'Jane')")
-        host_desc = st.text_input("Female Host Description (Optional)")
-        guest_name = st.text_input("Male Guest Name (leave blank for 'John')")
-        guest_desc = st.text_input("Male Guest Description (Optional)")
+        host_name = st.text_input("Host Name (leave blank for 'Jane')")
+        host_desc = st.text_input("Host Description (Optional)")
+        guest_name = st.text_input("Guest Name (leave blank for 'John')")
+        guest_desc = st.text_input("Guest Description (Optional)")
 
     user_specs = st.text_area("Any special instructions or prompts for the script? (Optional)", "")
     sponsor_content = st.text_area("Sponsored Content / Ad (Optional)", "")
-    sponsor_style = st.selectbox("Sponsor Integration Style", ["Separate Break", "Blended"])
+    sponsor_style = st.selectbox(
+        "Sponsor Integration Style",
+        ["Separate Break", "Blended"]
+    )
 
     custom_bg_music_file = st.file_uploader("Upload Custom Background Music (Optional)", type=["mp3", "wav"])
     custom_bg_music_path = None
@@ -293,6 +316,7 @@ def main():
         st.session_state["transcript"] = None
     if "transcript_original" not in st.session_state:
         st.session_state["transcript_original"] = None
+    # For Q&A
     if "qa_count" not in st.session_state:
         st.session_state["qa_count"] = 0
     if "conversation_history" not in st.session_state:
@@ -357,6 +381,7 @@ def main():
             st.session_state["audio_bytes"] = audio_bytes
             st.session_state["transcript"] = transcript
             st.session_state["transcript_original"] = transcript
+            # Reset Q&A
             st.session_state["qa_count"] = 0
             st.session_state["conversation_history"] = ""
 
@@ -381,6 +406,7 @@ def main():
                 st.session_state["transcript_original"],
                 edited_text
             )
+
             st.markdown("### **Edited Transcript Highlights**", unsafe_allow_html=True)
             st.markdown(highlighted_transcript, unsafe_allow_html=True)
 
@@ -428,9 +454,9 @@ def main():
                 st.markdown("### Updated Transcript")
                 st.markdown(new_transcript)
 
-        # -------------------------------------------
-        # Post-Podcast Q&A using st.audio_input():
-        # -------------------------------------------
+        # -----------------------
+        # POST-PODCAST Q&A Logic
+        # -----------------------
         st.markdown("## Post-Podcast Q&A")
         used_questions = st.session_state["qa_count"]
         remaining = MAX_QA_QUESTIONS - used_questions
@@ -439,7 +465,6 @@ def main():
             st.write(f"You can ask up to {remaining} more question(s).")
 
             typed_q = st.text_input("Type your follow-up question:")
-            # Replacing file_uploader with st.audio_input (Streamlit >= 1.41)
             audio_q = st.audio_input("Or record an audio question (WAV)")
 
             if st.button("Submit Q&A"):
@@ -448,8 +473,9 @@ def main():
                 else:
                     question_text = typed_q.strip()
                     if audio_q is not None:
-                        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
-                            tmp.write(audio_q.getvalue())
+                        suffix = ".wav"
+                        with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as tmp:
+                            tmp.write(audio_q.read())
                             local_audio_path = tmp.name
                         st.write("Transcribing your audio question...")
                         audio_transcript = transcribe_audio_deepgram(local_audio_path)
@@ -472,4 +498,4 @@ def main():
 
 
 if __name__ == "__main__":
-    main()
+    main()