Update app.py

app.py

@@ -16,7 +16,11 @@ from PIL import Image
 import zipfile
 import datetime
 import librosa
-import librosa.display
+import joblib
+import warnings
+
+# Suppress warnings for cleaner logs
+warnings.filterwarnings("ignore")
 
 # === Helper Functions ===
 def audiosegment_to_array(audio):
@@ -116,7 +120,7 @@ def stem_split(audio_path):
         save_track(path, sources[i].cpu(), model.samplerate)
         stem_paths.append(path)
 
-    return stem_paths
+    return [gr.File(value=path) for path in stem_paths]
 
 # === Preset Loader with Fallback ===
 def load_presets():
@@ -150,101 +154,113 @@ preset_names = list(preset_choices.keys())
 
 # === Waveform + Spectrogram Generator ===
 def show_waveform(audio_file):
-    audio = AudioSegment.from_file(audio_file)
-    samples = np.array(audio.get_array_of_samples())
-    plt.figure(figsize=(10, 2))
-    plt.plot(samples[:10000], color="blue")
-    plt.axis("off")
-    buf = BytesIO()
-    plt.savefig(buf, format="png", bbox_inches="tight", dpi=100)
-    plt.close()
-    buf.seek(0)
-    return Image.open(buf)
-
-def show_spectrogram(audio_file):
-    y, sr = torchaudio.load(audio_file)
-    y_np = y.numpy().flatten()
-    S = librosa.feature.melspectrogram(y=y_np, sr=sr)
-    plt.figure(figsize=(10, 2))
-    librosa.display.specshow(librosa.power_to_db(S, ref=np.max), sr=sr, x_axis='time', y_axis='mel')
-    plt.colorbar(format='%+2.0f dB')
-    plt.title('Spectrogram')
-    plt.tight_layout()
-    buf = BytesIO()
-    plt.savefig(buf, format="png", bbox_inches="tight", dpi=100)
-    plt.close()
-    buf.seek(0)
-    return Image.open(buf)
+    try:
+        audio = AudioSegment.from_file(audio_file)
+        samples = np.array(audio.get_array_of_samples())
+        plt.figure(figsize=(10, 2))
+        plt.plot(samples[:10000], color="blue")
+        plt.axis("off")
+        buf = BytesIO()
+        plt.savefig(buf, format="png", bbox_inches="tight", dpi=100)
+        plt.close()
+        buf.seek(0)
+        return Image.open(buf)
+    except Exception as e:
+        return None
+
+def detect_genre(audio_path):
+    try:
+        y, sr = torchaudio.load(audio_path)
+        mfccs = librosa.feature.mfcc(y=y.numpy().flatten(), sr=sr, n_mfcc=13).mean(axis=1).reshape(1, -1)
+        # Dummy classifier – replace with real one later
+        return "Speech"
+    except Exception:
+        return "Unknown"
 
 # === Session Info Export ===
-def generate_session_log(audio_path, effects, isolate_vocals, export_format):
+def generate_session_log(audio_path, effects, isolate_vocals, export_format, genre):
     log = {
         "timestamp": str(datetime.datetime.now()),
         "filename": os.path.basename(audio_path),
         "effects_applied": effects,
         "isolate_vocals": isolate_vocals,
-        "export_format": export_format
+        "export_format": export_format,
+        "detected_genre": genre
     }
     return json.dumps(log, indent=2)
 
-# === Main Processing Function ===
+# === Main Processing Function with Status Updates ===
 def process_audio(audio_file, selected_effects, isolate_vocals, preset_name, export_format):
-    audio = AudioSegment.from_file(audio_file)
-
-    effect_map = {
-        "Noise Reduction": apply_noise_reduction,
-        "Compress Dynamic Range": apply_compression,
-        "Add Reverb": apply_reverb,
-        "Pitch Shift": lambda x: apply_pitch_shift(x),
-        "Echo": apply_echo,
-        "Stereo Widening": apply_stereo_widen,
-        "Bass Boost": apply_bass_boost,
-        "Treble Boost": apply_treble_boost,
-        "Normalize": apply_normalize,
-    }
-
-    effects_to_apply = preset_choices.get(preset_name, selected_effects)
-    for effect_name in effects_to_apply:
-        if effect_name in effect_map:
-            audio = effect_map[effect_name](audio)
-
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as f:
-        if isolate_vocals:
-            temp_input = os.path.join(tempfile.gettempdir(), "input.wav")
-            audio.export(temp_input, format="wav")
-            vocal_path = apply_vocal_isolation(temp_input)
-            final_audio = AudioSegment.from_wav(vocal_path)
-        else:
-            final_audio = audio
-
-        output_path = f.name
-        final_audio.export(output_path, format=export_format.lower())
-
-        waveform_image = show_waveform(output_path)
-        spectrogram_image = show_spectrogram(output_path)
-        session_log = generate_session_log(audio_file, effects_to_apply, isolate_vocals, export_format)
-
-        return output_path, waveform_image, spectrogram_image, session_log
+    status = "🔊 Loading audio..."
+    try:
+        audio = AudioSegment.from_file(audio_file)
+        status = "🛠 Applying effects..."
+        
+        effect_map = {
+            "Noise Reduction": apply_noise_reduction,
+            "Compress Dynamic Range": apply_compression,
+            "Add Reverb": apply_reverb,
+            "Pitch Shift": lambda x: apply_pitch_shift(x),
+            "Echo": apply_echo,
+            "Stereo Widening": apply_stereo_widen,
+            "Bass Boost": apply_bass_boost,
+            "Treble Boost": apply_treble_boost,
+            "Normalize": apply_normalize,
+        }
+
+        effects_to_apply = preset_choices.get(preset_name, selected_effects)
+        for effect_name in effects_to_apply:
+            if effect_name in effect_map:
+                audio = effect_map[effect_name](audio)
+
+        status = "💾 Saving final audio..."
+        with tempfile.NamedTemporaryFile(delete=True, suffix=".wav") as f:
+            if isolate_vocals:
+                temp_input = os.path.join(tempfile.gettempdir(), "input.wav")
+                audio.export(temp_input, format="wav")
+                vocal_path = apply_vocal_isolation(temp_input)
+                final_audio = AudioSegment.from_wav(vocal_path)
+            else:
+                final_audio = audio
+
+            output_path = f.name
+            final_audio.export(output_path, format=export_format.lower())
+
+            waveform_image = show_waveform(output_path)
+            genre = detect_genre(output_path)
+            session_log = generate_session_log(audio_file, effects_to_apply, isolate_vocals, export_format, genre)
+
+            status = "🎉 Done!"
+            return output_path, waveform_image, session_log, genre, status
+
+    except Exception as e:
+        status = f"❌ Error: {str(e)}"
+        return None, None, status, "", status
 
 # === Batch Processing Function ===
 def batch_process_audio(files, selected_effects, isolate_vocals, preset_name, export_format):
-    output_dir = tempfile.mkdtemp()
-    results = []
-    session_logs = []
+    status = "🔊 Loading files..."
+    try:
+        output_dir = tempfile.mkdtemp()
+        results = []
+        session_logs = []
+
+        for file in files:
+            processed_path, _, log, _, _ = process_audio(file.name, selected_effects, isolate_vocals, preset_name, export_format)
+            results.append(processed_path)
+            session_logs.append(log)
 
-    for file in files:
-        processed_path, _, _, log = process_audio(file.name, selected_effects, isolate_vocals, preset_name, export_format)
-        results.append(processed_path)
-        session_logs.append(log)
+        zip_path = os.path.join(output_dir, "batch_output.zip")
+        with zipfile.ZipFile(zip_path, 'w') as zipf:
+            for i, res in enumerate(results):
+                filename = f"processed_{i}.{export_format.lower()}"
+                zipf.write(res, filename)
+                zipf.writestr(f"session_info_{i}.json", session_logs[i])
 
-    zip_path = os.path.join(output_dir, "batch_output.zip")
-    with zipfile.ZipFile(zip_path, 'w') as zipf:
-        for i, res in enumerate(results):
-            filename = f"processed_{i}.{export_format.lower()}"
-            zipf.write(res, filename)
-            zipf.writestr(f"session_info_{i}.json", session_logs[i])
+        return zip_path, "📦 ZIP created successfully!"
 
-    return zip_path
+    except Exception as e:
+        return None, f"❌ Batch processing failed: {str(e)}"
 
 # === Gradio Interface Setup ===
 effect_options = [
@@ -266,7 +282,6 @@ with gr.Blocks(title="AI Audio Studio", css="style.css") as demo:
     Upload, edit, and export audio with AI-powered tools.
     """)
 
-    # ----- Single File Studio Tab -----
     with gr.Tab("🎵 Single File Studio"):
         gr.Interface(
             fn=process_audio,
@@ -280,17 +295,17 @@ with gr.Blocks(title="AI Audio Studio", css="style.css") as demo:
             outputs=[
                 gr.Audio(label="Processed Audio", type="filepath"),
                 gr.Image(label="Waveform Preview"),
-                gr.Image(label="Spectrogram View"),
-                gr.Textbox(label="Session Log (JSON)", lines=5)
+                gr.Textbox(label="Session Log (JSON)", lines=5),
+                gr.Textbox(label="Detected Genre", lines=1),
+                gr.Textbox(label="Status", value="✅ Ready", lines=1)
             ],
             title="Edit One File at a Time",
-            description="Apply effects, preview waveform and spectrogram, and get full session log.",
+            description="Apply effects, preview waveform, and get full session log.",
             flagging_mode="never",
             submit_btn="Process Audio",
             clear_btn=None
         )
 
-    # ----- Batch Processing Tab -----
     with gr.Tab("🔊 Batch Processing"):
         gr.Interface(
             fn=batch_process_audio,
@@ -301,7 +316,10 @@ with gr.Blocks(title="AI Audio Studio", css="style.css") as demo:
                 gr.Dropdown(choices=preset_names, label="Select Preset", value=preset_names[0] if preset_names else None),
                 gr.Dropdown(choices=["MP3", "WAV"], label="Export Format", value="MP3")
             ],
-            outputs=gr.File(label="Download ZIP of All Processed Files"),
+            outputs=[
+                gr.File(label="Download ZIP of All Processed Files"),
+                gr.Textbox(label="Status", value="✅ Ready", lines=1)
+            ],
             title="Batch Audio Processor",
             description="Upload multiple files, apply effects in bulk, and download all results in a single ZIP.",
             flagging_mode="never",
@@ -309,14 +327,9 @@ with gr.Blocks(title="AI Audio Studio", css="style.css") as demo:
             clear_btn=None
         )
 
-    # ----- Remix Mode Tab -----
     with gr.Tab("🎛 Remix Mode (Split Stems)"):
-        def remix_mode(audio_file):
-            stems = stem_split(audio_file.name)
-            return [gr.File(value=stem) for stem in stems]
-
         gr.Interface(
-            fn=remix_mode,
+            fn=stem_split,
             inputs=gr.Audio(label="Upload Music Track", type="filepath"),
             outputs=[
                 gr.File(label="Vocals"),
@@ -330,25 +343,4 @@ with gr.Blocks(title="AI Audio Studio", css="style.css") as demo:
             clear_btn=None
         )
 
-    # ----- Session Info Tab -----
-    with gr.Tab("📝 Session Info"):
-        def get_session_info(audio_file, selected_effects, isolate_vocals, preset_name, export_format):
-            return generate_session_log(audio_file, selected_effects, isolate_vocals, export_format)
-
-        gr.Interface(
-            fn=get_session_info,
-            inputs=[
-                gr.Audio(label="Upload Audio", type="filepath"),
-                gr.CheckboxGroup(choices=effect_options, label="Apply Effects in Order"),
-                gr.Checkbox(label="Isolate Vocals After Effects"),
-                gr.Dropdown(choices=preset_names, label="Select Preset", value=preset_names[0] if preset_names else None),
-                gr.Dropdown(choices=["MP3", "WAV"], label="Export Format", value="MP3")
-            ],
-            outputs=gr.Textbox(label="Your Session Info (Copy or Save This)", lines=10),
-            title="Save Your Session Settings",
-            description="Get a full log of what was done to your track.",
-            flagging_mode="never",
-            clear_btn=None
-        )
-
 demo.launch()