Update app.py

app.py

@@ -1,8 +1,3 @@
-
-6. Changes in **`prompts.py`**: No Change
-
-7. Changes to **`APP.PY`**
-```python
 import streamlit as st
 import time
 import re
@@ -15,13 +10,14 @@ import difflib
 from utils import (
     generate_script,
     generate_audio_mp3,
-    truncate_text,
-    extract_text_from_url,
-    transcribe_youtube_video,
-    # research_topic, #Removed
+    # truncate_text, # Removed - no longer needed
+    # extract_text_from_url, # Removed - no longer primary input
+    # transcribe_youtube_video, # Removed - no longer primary input
+    # research_topic, # Removed - replaced by agent
     mix_with_bg_music,
     DialogueItem,
-    run_research_agent #Imported
+    run_research_agent, # Imported
+    generate_report
 )
 from prompts import SYSTEM_PROMPT
 
@@ -96,9 +92,6 @@ def regenerate_audio_from_dialogue(dialogue_items, custom_bg_music_path=None):
     return audio_bytes, transcript
 
 def generate_podcast(
-    # file, #Removed
-    # url, #Removed
-    # video_url, #Removed
     research_topic_input,
     tone,
     length_minutes,
@@ -110,47 +103,13 @@ def generate_podcast(
     sponsor_content,
     sponsor_style,
     custom_bg_music_path
-):#Removed sources and sources check
-    # 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)."
-    # if not any(sources):
-    #     return None, "Please provide at least one source."
+):
     if not research_topic_input:
       return None, "Please enter a topic to research for the podcast."
-    text = ""
-    # if file: #Removed
-    #     try:
-    #         if not file.name.lower().endswith('.pdf'):
-    #             return None, "Please upload a PDF file."
-    #         reader = pypdf.PdfReader(file)
-    #         text = " ".join(page.extract_text() for page in reader.pages if page.extract_text())
-    #     except Exception as e:
-    #         return None, f"Error reading PDF: {str(e)}"
-    # elif url: #Removed
-    #     try:
-    #         text = extract_text_from_url(url)
-    #         if not text:
-    #             return None, "Failed to extract text from URL."
-    #     except Exception as e:
-    #         return None, f"Error extracting text from URL: {str(e)}"
-    # elif video_url: #Removed
-    #     try:
-    #         text = transcribe_youtube_video(video_url)
-    #         if not text:
-    #             return None, "Failed to transcribe YouTube video."
-    #     except Exception as e:
-    #         return None, f"Error transcribing YouTube video: {str(e)}"
-    # elif research_topic_input:#Removed
-    #     try:
-    #         text = research_topic(research_topic_input)
-    #         if not text:
-    #             return None, f"Sorry, no information found on '{research_topic_input}'."
-    #     except Exception as e:
-    #         return None, f"Error researching topic: {str(e)}"
-    text = st.session_state["report_content"] #The text for the podcast is taken from the report content.
-    # from utils import truncate_text #Removed Truncate text
-    # text = truncate_text(text)
+    text = st.session_state.get("report_content", "") # Get report content
+    if not text:
+        return None, "Please generate a research report first, or enter a topic."
+
 
     extra_instructions = []
     if host_name or guest_name:
@@ -166,7 +125,6 @@ def generate_podcast(
             f"Sponsor Content Provided (should be under ~30 seconds):\n{sponsor_content}"
         )
 
-    from prompts import SYSTEM_PROMPT
     combined_instructions = "\n\n".join(extra_instructions).strip()
     full_prompt = SYSTEM_PROMPT
     if combined_instructions:
@@ -176,7 +134,6 @@ def generate_podcast(
     if st.session_state.get("language_selection") == "Hinglish":
         full_prompt += "\n\nPlease generate the script in Romanized Hindi.\n"
 
-    from utils import generate_script, generate_audio_mp3, mix_with_bg_music
     try:
         script = generate_script(
             full_prompt,
@@ -250,7 +207,7 @@ def highlight_differences(original: str, edited: str) -> str:
 
 def main():
     st.set_page_config(
-        page_title="MyPod v2: AI-Powered Podcast Magic",
+        page_title="MyPod v3: AI-Powered Podcast & Research",
         layout="centered"
     )
 
@@ -272,7 +229,7 @@ def main():
     with logo_col:
         st.image("logomypod.jpg", width=70)
     with title_col:
-        st.markdown("## MyPod v3: AI-Powered Podcast Magic")
+        st.markdown("## MyPod v3: AI-Powered Podcast & Research")
 
     st.markdown("""
     Welcome to **MyPod**, your go-to AI-powered podcast generator and research report tool! 🎉
@@ -331,7 +288,6 @@ def main():
                         st.text_area("Report Content", value=report_content, height=300)
 
                         # Generate PDF and offer download
-                        from utils import generate_report
                         with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmpfile:
                             pdf_path = tmpfile.name
                             generate_report(report_content, filename=pdf_path)  # Generate PDF
@@ -356,13 +312,7 @@ def main():
     elif mode == "Generate Podcast":
         st.markdown("### Generate Podcast")
 
-        # col1, col2 = st.columns(2) #Removed
-        # with col1:
-        #     file = st.file_uploader("Upload File (.pdf only)", type=["pdf"])
-        #     url = st.text_input("Or Enter Website URL")
-        #     video_url = st.text_input("Or Enter YouTube Link (Captioned videos)")
-        # with col2:
-        research_topic_input = st.text_input("Or Research a Topic for the podcast")
+        research_topic_input = st.text_input("Enter research topic for the podcast (or use a generated report):")
         tone = st.radio("Tone", ["Casual", "Formal", "Humorous", "Youthful"], index=0)
         length_minutes = st.slider("Podcast Length (in minutes)", 1, 60, 3)
 
@@ -373,7 +323,7 @@ def main():
         )
         st.session_state["language_selection"] = language
 
-        st.markdown("### Customize Your Podcast (New Features)")
+        st.markdown("### Customize Your Podcast (Optional)")
 
         with st.expander("Set Host & Guest Names/Descriptions (Optional)"):
             host_name = st.text_input("Female Host Name (leave blank for 'Jane')")
@@ -433,9 +383,6 @@ def main():
             time.sleep(1.0)
 
             audio_bytes, transcript = generate_podcast(
-                # file, #Removed
-                # url, #Removed
-                # video_url, #Removed
                 research_topic_input,
                 tone,
                 length_minutes,
@@ -465,7 +412,115 @@ def main():
                 st.session_state["qa_count"] = 0
                 st.session_state["conversation_history"] = ""
 
-        if st.session_state["audio_bytes"]:
+        if st.session_state.get("audio_bytes"):
             st.audio(st.session_state["audio_bytes"], format='audio/mp3')
             st.download_button(
-                label="Download Podcast (MP3)",
+                label="Download Podcast (MP3)",
+                data=st.session_state["audio_bytes"],
+                file_name="my_podcast.mp3",
+                mime="audio/mpeg"
+            )
+
+            st.markdown("### Generated Transcript (Editable)")
+            edited_text = st.text_area(
+                "Feel free to tweak lines, fix errors, or reword anything.",
+                value=st.session_state["transcript"],
+                height=300
+            )
+
+            if st.session_state.get("transcript_original"):
+                highlighted_transcript = highlight_differences(
+                    st.session_state["transcript_original"],
+                    edited_text
+                )
+
+                st.markdown("### **Edited Transcript Highlights**", unsafe_allow_html=True)
+                st.markdown(highlighted_transcript, unsafe_allow_html=True)
+
+            if st.button("Regenerate Audio From Edited Text"):
+                regen_bar = st.progress(0)
+                regen_text = st.empty()
+
+                regen_text.write("🔄 Regenerating your podcast with the edits...")
+                regen_bar.progress(25)
+                time.sleep(1.0)
+
+                regen_text.write("🔧 Adjusting the script based on your changes...")
+                regen_bar.progress(50)
+                time.sleep(1.0)
+
+                dialogue_items = parse_user_edited_transcript(
+                    edited_text,
+                    host_name or "Jane",
+                    guest_name or "John"
+                )
+                new_audio_bytes, new_transcript = regenerate_audio_from_dialogue(dialogue_items, custom_bg_music_path)
+
+                regen_bar.progress(75)
+                time.sleep(1.0)
+
+                if new_audio_bytes is None:
+                    regen_bar.progress(100)
+                    st.error(new_transcript)
+                else:
+                    regen_bar.progress(100)
+                    regen_text.write("✅ Regeneration complete!")
+                    st.success("Regenerated audio below:")
+
+                    st.session_state["audio_bytes"] = new_audio_bytes
+                    st.session_state["transcript"] = new_transcript
+                    st.session_state["transcript_original"] = new_transcript
+
+                    st.audio(new_audio_bytes, format='audio/mp3')
+                    st.download_button(
+                        label="Download Edited Podcast (MP3)",
+                        data=new_audio_bytes,
+                        file_name="my_podcast_edited.mp3",
+                        mime="audio/mpeg"
+                    )
+                    st.markdown("### Updated Transcript")
+                    st.markdown(new_transcript)
+
+            st.markdown("## Post-Podcast Q&A")
+            used_questions = st.session_state.get("qa_count", 0)
+            remaining = MAX_QA_QUESTIONS - used_questions
+
+            if remaining > 0:
+                st.write(f"You can ask up to {remaining} more question(s).")
+
+                typed_q = st.text_input("Type your follow-up question:")
+                audio_q = st.audio_input("Or record an audio question (WAV)")
+
+                if st.button("Submit Q&A"):
+                    if used_questions >= MAX_QA_QUESTIONS:
+                        st.warning("You have reached the Q&A limit.")
+                    else:
+                        question_text = typed_q.strip()
+                        if audio_q is not None:
+                            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)
+                            if audio_transcript:
+                                question_text = audio_transcript
+
+                        if not question_text:
+                            st.warning("No question found (text or audio).")
+                        else:
+                            st.write("Generating an answer...")
+                            ans_audio, ans_text = handle_qa_exchange(question_text)
+                            if ans_audio:
+                                st.audio(ans_audio, format='audio/mp3')
+                                st.markdown(f"**John**: {ans_text}")
+                                st.session_state["qa_count"] = used_questions + 1
+                            else:
+                                st.warning("No response could be generated.")
+            else:
+                st.write("You have used all 5 Q&A opportunities.")
+
+    st.markdown("<footer>©2025 MyPod. All rights reserved.</footer>", unsafe_allow_html=True)
+
+if __name__ == "__main__":
+    main()