Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| import os | |
| import shutil | |
| import asyncio | |
| import librosa | |
| import soundfile as sf | |
| import numpy as np | |
| import zipfile | |
| import tempfile | |
| import matplotlib | |
| matplotlib.use('Agg') | |
| # --------------------------- | |
| # UI helpers | |
| # --------------------------- | |
| def update_output_visibility(choice): | |
| if "2 Stems" in choice: | |
| return ( | |
| gr.update(visible=True), | |
| gr.update(visible=False), | |
| gr.update(visible=False), | |
| gr.update(visible=True, label="Instrumental (No Vocals)") | |
| ) | |
| else: # 4 stems | |
| return ( | |
| gr.update(visible=True), | |
| gr.update(visible=True), | |
| gr.update(visible=True), | |
| gr.update(visible=True, label="Other") | |
| ) | |
| # --------------------------- | |
| # Stem separation (Demucs) | |
| # --------------------------- | |
| async def separate_stems(audio_file_path, stem_choice, progress=gr.Progress(track_tqdm=True)): | |
| # outputs: [vocals_out, drums_out, bass_out, other_out, status_log] | |
| if audio_file_path is None: | |
| raise gr.Error("No audio file uploaded!") | |
| log_history = "Starting separation...\n" | |
| yield (gr.update(), gr.update(), gr.update(), gr.update(), log_history) | |
| try: | |
| progress(0.05, desc="Preparing audio file...") | |
| log_history += "Preparing audio file...\n" | |
| yield (gr.update(), gr.update(), gr.update(), gr.update(), log_history) | |
| original_filename_base = os.path.basename(audio_file_path).rsplit('.', 1)[0] | |
| stable_input_path = f"stable_input_{original_filename_base}.wav" | |
| # Ensure a .wav input for Demucs (demucs can read many formats, but filepath stability helps) | |
| shutil.copy(audio_file_path, stable_input_path) | |
| model_arg = "--two-stems=vocals" if "2 Stems" in stem_choice else "" | |
| output_dir = "separated" | |
| if os.path.exists(output_dir): | |
| shutil.rmtree(output_dir) | |
| command = f"python -m demucs {model_arg} -o \"{output_dir}\" \"{stable_input_path}\"" | |
| log_history += f"Running Demucs command: {command}\n" | |
| log_history += "(This may take a minute or more depending on track length)\n" | |
| yield (gr.update(), gr.update(), gr.update(), gr.update(), log_history) | |
| progress(0.2, desc="Running Demucs...") | |
| process = await asyncio.create_subprocess_shell( | |
| command, stdout=asyncio.subprocess.PIPE, stderr=asyncio.subprocess.PIPE | |
| ) | |
| stdout, stderr = await process.communicate() | |
| if process.returncode != 0: | |
| raise gr.Error(f"Demucs failed. Error: {stderr.decode(errors='ignore')[:800]}") | |
| log_history += "Demucs finished. Locating stem files...\n" | |
| progress(0.8, desc="Locating stems...") | |
| yield (gr.update(), gr.update(), gr.update(), gr.update(), log_history) | |
| stable_filename_base = os.path.basename(stable_input_path).rsplit('.', 1)[0] | |
| subfolders = [f.name for f in os.scandir(output_dir) if f.is_dir()] | |
| if not subfolders: | |
| raise gr.Error("Demucs output folder structure not found!") | |
| model_folder_name = subfolders[0] | |
| stems_path = os.path.join(output_dir, model_folder_name, stable_filename_base) | |
| if not os.path.exists(stems_path): | |
| raise gr.Error(f"Demucs output directory was not found! Looked for: {stems_path}") | |
| vocals_path = os.path.join(stems_path, "vocals.wav") if os.path.exists(os.path.join(stems_path, "vocals.wav")) else None | |
| drums_path = os.path.join(stems_path, "drums.wav") if os.path.exists(os.path.join(stems_path, "drums.wav")) else None | |
| bass_path = os.path.join(stems_path, "bass.wav") if os.path.exists(os.path.join(stems_path, "bass.wav")) else None | |
| other_name = "no_vocals.wav" if "2 Stems" in stem_choice else "other.wav" | |
| other_path = os.path.join(stems_path, other_name) if os.path.exists(os.path.join(stems_path, other_name)) else None | |
| try: | |
| os.remove(stable_input_path) | |
| except Exception: | |
| pass | |
| log_history += "✅ Stem separation complete!\n" | |
| yield ( | |
| gr.update(value=vocals_path), | |
| gr.update(value=drums_path), | |
| gr.update(value=bass_path), | |
| gr.update(value=other_path), | |
| log_history | |
| ) | |
| except Exception as e: | |
| err = f"❌ ERROR: {e}" | |
| print(f"Separation error: {e}") | |
| yield (gr.update(), gr.update(), gr.update(), gr.update(), log_history + err) | |
| # --------------------------- | |
| # Slicing (BPM override + Quantized grid) | |
| # --------------------------- | |
| def _grid_beats_per_step(grid_label: str) -> float: | |
| # "1/16 (Sixteenth)" -> "1/16" | |
| tok = grid_label.split(" ")[0].strip() | |
| mapping = {"1/1": 4.0, "1/2": 2.0, "1/4": 1.0, "1/8": 0.5, "1/16": 0.25} | |
| return mapping.get(tok, 0.25) | |
| def _load_audio_any(stem_input): | |
| """ | |
| Accepts a filepath string OR (sr, numpy_array) and returns (sr, mono_float_array). | |
| We use filepath everywhere in this app to keep memory stable. | |
| """ | |
| if stem_input is None: | |
| return None, None | |
| if isinstance(stem_input, str): | |
| y, sr = librosa.load(stem_input, sr=None, mono=False) | |
| else: | |
| sr, y = stem_input | |
| y = librosa.util.buf_to_float(y) | |
| y_mono = librosa.to_mono(y) if y.ndim > 1 else y | |
| return sr, y_mono | |
| def slice_stem_real( | |
| stem_input, | |
| loop_choice, | |
| sensitivity, | |
| stem_name, | |
| progress_fn=None, | |
| bpm_override_val: int = 0, | |
| quantize: bool = True, | |
| grid_label: str = "1/16 (Sixteenth)", | |
| ): | |
| sr, y_mono = _load_audio_any(stem_input) | |
| if sr is None or y_mono is None: | |
| return None, None | |
| if progress_fn: progress_fn(0.1, desc="Detecting BPM...") | |
| tempo, _ = librosa.beat.beat_track(y=y_mono, sr=sr) | |
| det_bpm = int(np.round(tempo)) if tempo and tempo > 0 else 120 | |
| bpm = int(bpm_override_val) if bpm_override_val and bpm_override_val > 0 else det_bpm | |
| if bpm <= 0: | |
| bpm = 120 | |
| output_files = [] | |
| loops_dir = tempfile.mkdtemp() | |
| if "One-Shots" in loop_choice: | |
| if progress_fn: progress_fn(0.3, desc="Finding transients...") | |
| onset_frames = librosa.onset.onset_detect( | |
| y=y_mono, sr=sr, delta=sensitivity, wait=1, pre_avg=1, post_avg=1, post_max=1 | |
| ) | |
| onset_samples = librosa.frames_to_samples(onset_frames) | |
| # Quantize onsets to musical grid if enabled | |
| if quantize and len(onset_samples) > 0: | |
| beats_per_step = _grid_beats_per_step(grid_label) | |
| samples_per_beat = sr * (60.0 / bpm) | |
| step = max(1, int(round(samples_per_beat * beats_per_step))) | |
| q = np.clip(np.round(onset_samples / step) * step, 0, len(y_mono) - 1).astype(int) | |
| q = np.unique(q) # dedupe after snapping | |
| onset_samples = q | |
| if progress_fn: progress_fn(0.5, desc="Slicing one-shots...") | |
| if len(onset_samples) > 0: | |
| num_onsets = len(onset_samples) | |
| min_len = max(1, int(0.02 * sr)) # 20ms min slice guard | |
| for i, start_sample in enumerate(onset_samples): | |
| end_sample = onset_samples[i+1] if i+1 < num_onsets else len(y_mono) | |
| if end_sample - start_sample < min_len: | |
| continue | |
| slice_data = y_mono[start_sample:end_sample] | |
| filename = os.path.join(loops_dir, f"{stem_name}_one_shot_{i+1:03d}.wav") | |
| sf.write(filename, slice_data, sr, subtype='PCM_16') | |
| output_files.append(filename) | |
| if progress_fn and num_onsets > 1: | |
| progress_fn(0.5 + (i / (num_onsets - 1) * 0.5), desc=f"Exporting slice {i+1}/{num_onsets}...") | |
| else: | |
| # Grid-true by construction; honor BPM override | |
| bars = int(loop_choice.split(" ")[0]) | |
| seconds_per_beat = 60.0 / bpm | |
| seconds_per_bar = seconds_per_beat * 4 | |
| loop_duration_seconds = seconds_per_bar * bars | |
| loop_duration_samples = int(loop_duration_seconds * sr) | |
| if progress_fn: progress_fn(0.4, desc=f"Slicing into {bars}-bar loops @ {bpm} BPM...") | |
| num_loops = len(y_mono) // loop_duration_samples | |
| if num_loops == 0: | |
| return None, None | |
| for i in range(num_loops): | |
| start_sample = i * loop_duration_samples | |
| end_sample = start_sample + loop_duration_samples | |
| slice_data = y_mono[start_sample:end_sample] | |
| filename = os.path.join(loops_dir, f"{stem_name}_{bars}bar_loop_{i+1:03d}_{bpm}bpm.wav") | |
| sf.write(filename, slice_data, sr, subtype='PCM_16') | |
| output_files.append(filename) | |
| if progress_fn and num_loops > 1: | |
| progress_fn(0.4 + (i / (num_loops - 1) * 0.6), desc=f"Exporting loop {i+1}/{num_loops}...") | |
| if not output_files: | |
| return None, None | |
| return output_files, loops_dir | |
| # --------------------------- | |
| # Batch wrapper | |
| # --------------------------- | |
| async def slice_all_and_zip_real(vocals, drums, bass, other, loop_choice, sensitivity, | |
| bpm_over, quantize_on, grid_sel, | |
| progress=gr.Progress(track_tqdm=True)): | |
| # outputs: [status_log, download_zip_file] | |
| log_history = "Starting batch slice...\n" | |
| yield (log_history, gr.update(visible=False)) | |
| await asyncio.sleep(0.05) | |
| stems_to_process = {"vocals": vocals, "drums": drums, "bass": bass, "other": other} | |
| present = {k: v for k, v in stems_to_process.items() if v} | |
| if not present: | |
| raise gr.Error("No stems to process! Please separate stems first.") | |
| zip_path = "Loop_Architect_Pack.zip" | |
| all_temp_dirs = [] | |
| try: | |
| with zipfile.ZipFile(zip_path, 'w') as zf: | |
| processed, total = 0, len(present) | |
| for name, data in present.items(): | |
| log_history += f"--- Slicing {name} stem ---\n" | |
| yield (log_history, gr.update(visible=False)) | |
| def update_main_progress(p, desc=""): | |
| progress((processed + max(0.01, p)) / total, desc=f"{name}: {desc}") | |
| sliced_files, temp_dir = slice_stem_real( | |
| data, loop_choice, sensitivity, name, | |
| progress_fn=update_main_progress, | |
| bpm_override_val=int(bpm_over), | |
| quantize=bool(quantize_on), | |
| grid_label=grid_sel | |
| ) | |
| if temp_dir: | |
| all_temp_dirs.append(temp_dir) | |
| if sliced_files: | |
| log_history += f"Generated {len(sliced_files)} slices for {name}.\n" | |
| for loop_file in sliced_files: | |
| zf.write(loop_file, os.path.join(name, os.path.basename(loop_file))) | |
| else: | |
| log_history += f"No slices generated for {name}.\n" | |
| processed += 1 | |
| yield (log_history, gr.update(visible=False)) | |
| log_history += "Packaging complete! ✅ Pack ready for download.\n" | |
| yield (log_history, gr.update(value=zip_path, visible=True)) | |
| except Exception as e: | |
| err = f"❌ ERROR: {e}" | |
| print(f"Batch slice error: {e}") | |
| yield (log_history + err, gr.update(visible=False)) | |
| finally: | |
| for d in all_temp_dirs: | |
| if d and os.path.exists(d): | |
| shutil.rmtree(d) | |
| # --------------------------- | |
| # Gradio UI | |
| # --------------------------- | |
| with gr.Blocks(theme=gr.themes.Default(primary_hue="blue", secondary_hue="red")) as demo: | |
| gr.Markdown("# 🎵 Loop Architect — SALT PRØPHET Edition") | |
| with gr.Row(): | |
| with gr.Column(scale=1): | |
| gr.Markdown("### 1) Separate Stems") | |
| audio_input = gr.Audio(type="filepath", label="Upload a Track") | |
| stem_options = gr.Radio( | |
| ["4 Stems (Vocals, Drums, Bass, Other)", "2 Stems (Vocals + Instrumental)"], | |
| label="Separation Type", | |
| value="4 Stems (Vocals, Drums, Bass, Other)" | |
| ) | |
| submit_button = gr.Button("Separate Stems", variant="secondary") | |
| with gr.Accordion("2) Slicing Options", open=True): | |
| loop_options_radio = gr.Radio( | |
| ["One-Shots (All Transients)", "4 Bar Loops", "8 Bar Loops"], | |
| label="Slice Type", | |
| value="One-Shots (All Transients)" | |
| ) | |
| sensitivity_slider = gr.Slider( | |
| minimum=0.01, maximum=0.5, value=0.05, step=0.01, | |
| label="One-Shot Sensitivity", info="Lower values = more slices" | |
| ) | |
| bpm_override = gr.Slider( | |
| minimum=0, maximum=240, value=0, step=1, | |
| label="BPM Override (0 = auto-detect)" | |
| ) | |
| quantize_toggle = gr.Checkbox( | |
| label="Quantize One-Shots to Grid", value=True | |
| ) | |
| grid_select = gr.Dropdown( | |
| choices=["1/1 (Whole)", "1/2 (Half)", "1/4 (Quarter)", "1/8 (Eighth)", "1/16 (Sixteenth)"], | |
| value="1/16 (Sixteenth)", | |
| label="Grid Resolution" | |
| ) | |
| gr.Markdown("### 3) Create Pack") | |
| slice_all_button = gr.Button("Slice All Stems & Create Pack", variant="primary") | |
| download_zip_file = gr.File(label="Download Your Loop Pack", visible=False) | |
| gr.Markdown("### Status") | |
| status_log = gr.Textbox(label="Status Log", lines=12, interactive=False) | |
| with gr.Column(scale=2): | |
| with gr.Accordion("Separated Stems", open=True): | |
| with gr.Row(): | |
| vocals_output = gr.Audio(type="filepath", label="Vocals", scale=4) | |
| slice_vocals_btn = gr.Button("Slice Vocals", scale=1) | |
| with gr.Row(): | |
| drums_output = gr.Audio(type="filepath", label="Drums", scale=4) | |
| slice_drums_btn = gr.Button("Slice Drums", scale=1) | |
| with gr.Row(): | |
| bass_output = gr.Audio(type="filepath", label="Bass", scale=4) | |
| slice_bass_btn = gr.Button("Slice Bass", scale=1) | |
| with gr.Row(): | |
| other_output = gr.Audio(type="filepath", label="Other / Instrumental", scale=4) | |
| slice_other_btn = gr.Button("Slice Other", scale=1) | |
| gr.Markdown("### Sliced Files") | |
| loops_files = gr.Files(label="Generated Loops / One-Shots") | |
| # wire events | |
| submit_button.click( | |
| fn=separate_stems, | |
| inputs=[audio_input, stem_options], | |
| outputs=[vocals_output, drums_output, bass_output, other_output, status_log] | |
| ) | |
| stem_options.change( | |
| fn=update_output_visibility, | |
| inputs=stem_options, | |
| outputs=[vocals_output, drums_output, bass_output, other_output] | |
| ) | |
| def slice_and_display(stem_path, loop_choice, sensitivity, stem_name, bpm_over, quantize_on, grid_sel): | |
| log_history = f"Slicing {stem_name}...\n" | |
| def _p(p, desc=""): | |
| gr.Progress(track_tqdm=True)(p, desc=desc) | |
| files, temp_dir = slice_stem_real( | |
| stem_path, loop_choice, sensitivity, stem_name, | |
| progress_fn=_p, | |
| bpm_override_val=int(bpm_over), | |
| quantize=bool(quantize_on), | |
| grid_label=grid_sel | |
| ) | |
| if temp_dir and os.path.exists(temp_dir): | |
| shutil.rmtree(temp_dir) | |
| if files: | |
| log_history += f"✅ Sliced {stem_name} into {len(files)} pieces." | |
| return gr.update(value=files), log_history | |
| else: | |
| log_history += f"⚠️ No slices generated for {stem_name}." | |
| return gr.update(value=None), log_history | |
| slice_vocals_btn.click( | |
| fn=slice_and_display, | |
| inputs=[vocals_output, loop_options_radio, sensitivity_slider, gr.Textbox("vocals", visible=False), | |
| bpm_override, quantize_toggle, grid_select], | |
| outputs=[loops_files, status_log] | |
| ) | |
| slice_drums_btn.click( | |
| fn=slice_and_display, | |
| inputs=[drums_output, loop_options_radio, sensitivity_slider, gr.Textbox("drums", visible=False), | |
| bpm_override, quantize_toggle, grid_select], | |
| outputs=[loops_files, status_log] | |
| ) | |
| slice_bass_btn.click( | |
| fn=slice_and_display, | |
| inputs=[bass_output, loop_options_radio, sensitivity_slider, gr.Textbox("bass", visible=False), | |
| bpm_override, quantize_toggle, grid_select], | |
| outputs=[loops_files, status_log] | |
| ) | |
| slice_other_btn.click( | |
| fn=slice_and_display, | |
| inputs=[other_output, loop_options_radio, sensitivity_slider, gr.Textbox("other", visible=False), | |
| bpm_override, quantize_toggle, grid_select], | |
| outputs=[loops_files, status_log] | |
| ) | |
| slice_all_button.click( | |
| fn=slice_all_and_zip_real, | |
| inputs=[vocals_output, drums_output, bass_output, other_output, | |
| loop_options_radio, sensitivity_slider, | |
| bpm_override, quantize_toggle, grid_select], | |
| outputs=[status_log, download_zip_file] | |
| ) | |
| # Spaces calls launch() automatically | |
| demo.queue(concurrency_count=1, max_size=8).launch() | |