rename: app.py -> app_cascade.py, app_lfm.py -> app.py

LFM variant is now the default app.py for HF Space compatibility.
Original 3-model pipeline preserved as app_cascade.py.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

app.py

@@ -10,6 +10,7 @@ from pathlib import Path
 
 from tts import split_into_chunks, generate_audio_stream
 from inference import transcribe_audio, answer_story_question
+from inference_lfm import answer_question_audio, SAMPLE_RATE as LFM_SR
 from voice_clone import load_default_profile, list_saved_profiles
 
 # Create directories for sample audio files
@@ -139,6 +140,8 @@ with open(os.path.join(os.path.dirname(__file__), "static", "style.css"), encodi
     css_code = _f.read()
 
 def generate_dashboard_html(books_list):
+    if not books_list:
+        return "<div style='padding:40px;text-align:center;color:#FAF7F2;'>No stories found. Add .txt files to the stories/ folder.</div>"
     featured = books_list[0]
     others = books_list[1:]
     featured_js_title = featured['title'].replace("'", "\\'").replace('"', '\\"')
@@ -1016,7 +1019,7 @@ with gr.Blocks(title="MomsVoice", css=css_code) as demo:
         }"""
     )
 
-    # 9. Submit question — always auto-submits when recording stops
+    # 9. Submit question — LFM2.5-Audio end-to-end (single model: audio-in → audio+text-out)
     def handle_question_submit(question_txt, question_audio_path, paragraphs, slider_val, profile_id):
         q_txt = (question_txt or "").strip()
         has_audio = question_audio_path is not None and question_audio_path != ""
@@ -1029,56 +1032,45 @@ with gr.Blocks(title="MomsVoice", css=css_code) as demo:
             """
             return gr.HTML(value=answer_html, visible=True), gr.Audio(visible=False)
 
-        # Step 9: Transcribe audio if text is empty (on-demand ASR)
-        if q_txt:
-            q_text = q_txt
-        else:
-            try:
-                q_text = transcribe_audio(question_audio_path)
-                if not q_text:
-                    q_text = "(could not understand audio)"
-            except Exception as e:
-                q_text = f"(transcription failed: {e})"
-
-        # Step 10: Generate grounded answer from story context using Qwen
-        current_idx = int(slider_val) if slider_val else 0
+        # Build story context from paragraphs
+        story_context = "\n\n".join(paragraphs) if paragraphs else ""
+
+        # Use LFM2.5-Audio end-to-end: accepts audio/text → produces text+audio answer
+        import uuid
+        import soundfile as sf
+        import logging as _log
+
         try:
-            answer_text = answer_story_question(q_text, paragraphs, current_idx)
+            answer_text, waveform, sr = answer_question_audio(
+                question_audio_path=question_audio_path if has_audio else None,
+                question_text=q_txt if q_txt and not has_audio else None,
+                story_context=story_context,
+                max_new_tokens=200,
+            )
             if not answer_text:
                 answer_text = "Hmm, I'm not sure about that! Let's keep listening to find out."
         except Exception as e:
             answer_text = f"Oops, I couldn't think of an answer right now. Let's keep reading! ({type(e).__name__})"
-            import logging
-            logging.getLogger(__name__).exception("Q&A failed: %s", e)
+            waveform = None
+            sr = LFM_SR
+            _log.getLogger(__name__).exception("LFM Q&A failed: %s", e)
 
-        # Synthesize answer in cloned voice via TTS
-        answer_audio_path = None
-        try:
-            from tts import split_into_chunks as _split, generate_audio_stream as _gen_stream
-            import soundfile as sf
-            import numpy as np
-            import uuid
-
-            chunks = _split(answer_text)
-            audio_segments = []
-            sample_rate = 16000
-            for sr, wav, idx, total, err in _gen_stream(chunks, voice_profile_id=profile_id):
-                if wav is not None:
-                    audio_segments.append(wav)
-                    sample_rate = sr
-            if audio_segments:
-                full_audio = np.concatenate(audio_segments)
-                answer_audio_path = f"sample_sounds/qa_answer_{uuid.uuid4().hex[:8]}.wav"
-                sf.write(answer_audio_path, full_audio, sample_rate)
-        except Exception:
-            # Fall back to no audio if TTS fails
-            pass
+        # Save audio if LFM produced it
+        answer_audio_path_out = None
+        if waveform is not None:
+            answer_audio_path_out = f"sample_sounds/qa_answer_{uuid.uuid4().hex[:8]}.wav"
+            sf.write(answer_audio_path_out, waveform, sr)
+
+        # Display question text (transcribe if needed for display only)
+        display_question = q_txt
+        if not display_question and has_audio:
+            display_question = "(audio question)"
 
         safe_answer = html.escape(answer_text)
-        safe_question = html.escape(q_text)
+        safe_question = html.escape(display_question)
         answer_html = f"""
             <div style="margin-top: 12px; padding: 16px; background: rgba(240,253,244,0.1); border: 1px solid rgba(187,247,208,0.3); border-radius: 14px;">
-                <div style="font-size: 10px; font-weight: 700; text-transform: uppercase; color: #4ade80; margin-bottom: 6px;">Answer in Narrator's Voice</div>
+                <div style="font-size: 10px; font-weight: 700; text-transform: uppercase; color: #4ade80; margin-bottom: 6px;">Answer (LFM2.5-Audio)</div>
                 <div style="font-family: 'Playfair Display', Georgia, serif; font-style: italic; color: #FAF7F2; font-size: 13px; line-height: 1.6;">
                     &ldquo;{safe_answer}&rdquo;
                 </div>
@@ -1087,8 +1079,8 @@ with gr.Blocks(title="MomsVoice", css=css_code) as demo:
                 </div>
             </div>
         """
-        if answer_audio_path:
-            return gr.HTML(value=answer_html, visible=True), gr.Audio(value=answer_audio_path, visible=True)
+        if answer_audio_path_out:
+            return gr.HTML(value=answer_html, visible=True), gr.Audio(value=answer_audio_path_out, visible=True)
         return gr.HTML(value=answer_html, visible=True), gr.Audio(visible=False)
 
     submit_question_btn.click(

app_lfm.py → app_cascade.py

@@ -10,7 +10,6 @@ from pathlib import Path
 
 from tts import split_into_chunks, generate_audio_stream
 from inference import transcribe_audio, answer_story_question
-from inference_lfm import answer_question_audio, SAMPLE_RATE as LFM_SR
 from voice_clone import load_default_profile, list_saved_profiles
 
 # Create directories for sample audio files
@@ -1017,7 +1016,7 @@ with gr.Blocks(title="MomsVoice", css=css_code) as demo:
         }"""
     )
 
-    # 9. Submit question — LFM2.5-Audio end-to-end (single model: audio-in → audio+text-out)
+    # 9. Submit question — always auto-submits when recording stops
     def handle_question_submit(question_txt, question_audio_path, paragraphs, slider_val, profile_id):
         q_txt = (question_txt or "").strip()
         has_audio = question_audio_path is not None and question_audio_path != ""
@@ -1030,45 +1029,56 @@ with gr.Blocks(title="MomsVoice", css=css_code) as demo:
             """
             return gr.HTML(value=answer_html, visible=True), gr.Audio(visible=False)
 
-        # Build story context from paragraphs
-        story_context = "\n\n".join(paragraphs) if paragraphs else ""
-
-        # Use LFM2.5-Audio end-to-end: accepts audio/text → produces text+audio answer
-        import uuid
-        import soundfile as sf
-        import logging as _log
-
+        # Step 9: Transcribe audio if text is empty (on-demand ASR)
+        if q_txt:
+            q_text = q_txt
+        else:
+            try:
+                q_text = transcribe_audio(question_audio_path)
+                if not q_text:
+                    q_text = "(could not understand audio)"
+            except Exception as e:
+                q_text = f"(transcription failed: {e})"
+
+        # Step 10: Generate grounded answer from story context using Qwen
+        current_idx = int(slider_val) if slider_val else 0
         try:
-            answer_text, waveform, sr = answer_question_audio(
-                question_audio_path=question_audio_path if has_audio else None,
-                question_text=q_txt if q_txt and not has_audio else None,
-                story_context=story_context,
-                max_new_tokens=200,
-            )
+            answer_text = answer_story_question(q_text, paragraphs, current_idx)
             if not answer_text:
                 answer_text = "Hmm, I'm not sure about that! Let's keep listening to find out."
         except Exception as e:
             answer_text = f"Oops, I couldn't think of an answer right now. Let's keep reading! ({type(e).__name__})"
-            waveform = None
-            sr = LFM_SR
-            _log.getLogger(__name__).exception("LFM Q&A failed: %s", e)
-
-        # Save audio if LFM produced it
-        answer_audio_path_out = None
-        if waveform is not None:
-            answer_audio_path_out = f"sample_sounds/qa_answer_{uuid.uuid4().hex[:8]}.wav"
-            sf.write(answer_audio_path_out, waveform, sr)
+            import logging
+            logging.getLogger(__name__).exception("Q&A failed: %s", e)
 
-        # Display question text (transcribe if needed for display only)
-        display_question = q_txt
-        if not display_question and has_audio:
-            display_question = "(audio question)"
+        # Synthesize answer in cloned voice via TTS
+        answer_audio_path = None
+        try:
+            from tts import split_into_chunks as _split, generate_audio_stream as _gen_stream
+            import soundfile as sf
+            import numpy as np
+            import uuid
+
+            chunks = _split(answer_text)
+            audio_segments = []
+            sample_rate = 16000
+            for sr, wav, idx, total, err in _gen_stream(chunks, voice_profile_id=profile_id):
+                if wav is not None:
+                    audio_segments.append(wav)
+                    sample_rate = sr
+            if audio_segments:
+                full_audio = np.concatenate(audio_segments)
+                answer_audio_path = f"sample_sounds/qa_answer_{uuid.uuid4().hex[:8]}.wav"
+                sf.write(answer_audio_path, full_audio, sample_rate)
+        except Exception:
+            # Fall back to no audio if TTS fails
+            pass
 
         safe_answer = html.escape(answer_text)
-        safe_question = html.escape(display_question)
+        safe_question = html.escape(q_text)
         answer_html = f"""
             <div style="margin-top: 12px; padding: 16px; background: rgba(240,253,244,0.1); border: 1px solid rgba(187,247,208,0.3); border-radius: 14px;">
-                <div style="font-size: 10px; font-weight: 700; text-transform: uppercase; color: #4ade80; margin-bottom: 6px;">Answer (LFM2.5-Audio)</div>
+                <div style="font-size: 10px; font-weight: 700; text-transform: uppercase; color: #4ade80; margin-bottom: 6px;">Answer in Narrator's Voice</div>
                 <div style="font-family: 'Playfair Display', Georgia, serif; font-style: italic; color: #FAF7F2; font-size: 13px; line-height: 1.6;">
                     &ldquo;{safe_answer}&rdquo;
                 </div>
@@ -1077,8 +1087,8 @@ with gr.Blocks(title="MomsVoice", css=css_code) as demo:
                 </div>
             </div>
         """
-        if answer_audio_path_out:
-            return gr.HTML(value=answer_html, visible=True), gr.Audio(value=answer_audio_path_out, visible=True)
+        if answer_audio_path:
+            return gr.HTML(value=answer_html, visible=True), gr.Audio(value=answer_audio_path, visible=True)
         return gr.HTML(value=answer_html, visible=True), gr.Audio(visible=False)
 
     submit_question_btn.click(

test_lfm_answer.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bfeb07501ccf11c32d6b1d2614b8edabad10ac5f05fbf48b81563a66060432a
+size 188204

test_lfm_answer2.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f889b069437f2cf0d8466ee43174d8d5f0f169f337c4206bb9f53ca05e928358
+size 134444