print debug

src/main.py

@@ -38,7 +38,7 @@ def get_youtube_video_id(url, ignore_playlist=True):
     http://www.youtube.com/embed/SA2iWivDJiE
     http://www.youtube.com/v/SA2iWivDJiE?version=3&hl=en_US
     """
-    query = urlparse(url)
+    query = urlparse(url, allow_fragments=True)
     if query.hostname == 'youtu.be':
         if query.path[1:] == 'watch':
             return query.query[2:]
@@ -63,6 +63,7 @@ def get_youtube_video_id(url, ignore_playlist=True):
 
 
 def yt_download(link):
+    print("[~] Downloading YouTube audio...")
     ydl_opts = {
         'format': 'bestaudio',
         'outtmpl': '%(title)s',
@@ -77,6 +78,7 @@ def yt_download(link):
         result = ydl.extract_info(link, download=True)
         download_path = ydl.prepare_filename(result, outtmpl='%(title)s.mp3')
 
+    print(f"[+] YouTube audio downloaded: {download_path}")
     return download_path
 
 
@@ -88,6 +90,7 @@ def raise_exception(error_msg, is_webui):
 
 
 def get_rvc_model(voice_model, is_webui):
+    print(f"[~] Getting RVC model: {voice_model}")
     rvc_model_filename, rvc_index_filename = None, None
     model_dir = os.path.join(rvc_models_dir, voice_model)
     for file in os.listdir(model_dir):
@@ -101,10 +104,12 @@ def get_rvc_model(voice_model, is_webui):
         error_msg = f'No model file exists in {model_dir}.'
         raise_exception(error_msg, is_webui)
 
+    print(f"[+] RVC model found: {rvc_model_filename}")
     return os.path.join(model_dir, rvc_model_filename), os.path.join(model_dir, rvc_index_filename) if rvc_index_filename else ''
 
 
 def get_audio_paths(song_dir):
+    print(f"[~] Getting audio paths from: {song_dir}")
     orig_song_path = None
     instrumentals_path = None
     main_vocals_dereverb_path = None
@@ -121,10 +126,12 @@ def get_audio_paths(song_dir):
         elif file.endswith('_Vocals_Backup.wav'):
             backup_vocals_path = os.path.join(song_dir, file)
 
+    print(f"[+] Audio paths found: {orig_song_path}, {instrumentals_path}, {main_vocals_dereverb_path}, {backup_vocals_path}")
     return orig_song_path, instrumentals_path, main_vocals_dereverb_path, backup_vocals_path
 
 
 def convert_to_stereo(audio_path):
+    print(f"[~] Converting to stereo: {audio_path}")
     wave, sr = librosa.load(audio_path, mono=False, sr=44100)
 
     # check if mono
@@ -132,12 +139,15 @@ def convert_to_stereo(audio_path):
         stereo_path = f'{os.path.splitext(audio_path)[0]}_stereo.wav'
         command = shlex.split(f'ffmpeg -y -loglevel error -i "{audio_path}" -ac 2 -f wav "{stereo_path}"')
         subprocess.run(command)
+        print(f"[+] Converted to stereo: {stereo_path}")
         return stereo_path
     else:
+        print("[+] Audio already in stereo")
         return audio_path
 
 
 def pitch_shift(audio_path, pitch_change):
+    print(f"[~] Pitch shifting: {audio_path} by {pitch_change}")
     output_path = f'{os.path.splitext(audio_path)[0]}_p{pitch_change}.wav'
     if not os.path.exists(output_path):
         y, sr = sf.read(audio_path)
@@ -146,16 +156,20 @@ def pitch_shift(audio_path, pitch_change):
         y_shifted = tfm.build_array(input_array=y, sample_rate_in=sr)
         sf.write(output_path, y_shifted, sr)
 
+    print(f"[+] Pitch shifted audio saved: {output_path}")
     return output_path
 
 
 def get_hash(filepath):
+    print(f"[~] Generating hash for: {filepath}")
     with open(filepath, 'rb') as f:
         file_hash = hashlib.blake2b()
         while chunk := f.read(8192):
             file_hash.update(chunk)
 
-    return file_hash.hexdigest()[:11]
+    hash_value = file_hash.hexdigest()[:11]
+    print(f"[+] Hash generated: {hash_value}")
+    return hash_value
 
 
 def display_progress(message, percent, is_webui, progress=None):
@@ -166,6 +180,7 @@ def display_progress(message, percent, is_webui, progress=None):
 
 
 def preprocess_song(song_input, mdx_model_params, song_id, is_webui, input_type, progress=None):
+    print("[~] Starting song preprocessing...")
     keep_orig = False
     if input_type == 'yt':
         display_progress('[~] Downloading song...', 0, is_webui, progress)
@@ -189,10 +204,12 @@ def preprocess_song(song_input, mdx_model_params, song_id, is_webui, input_type,
     display_progress('[~] Applying DeReverb to Vocals...', 0.3, is_webui, progress)
     _, main_vocals_dereverb_path = run_mdx(mdx_model_params, song_output_dir, os.path.join(mdxnet_models_dir, 'Reverb_HQ_By_FoxJoy.onnx'), main_vocals_path, invert_suffix='DeReverb', exclude_main=True, denoise=True)
 
+    print("[+] Song preprocessing completed")
     return orig_song_path, vocals_path, instrumentals_path, main_vocals_path, backup_vocals_path, main_vocals_dereverb_path
 
 
 def voice_change(voice_model, vocals_path, output_path, pitch_change, f0_method, index_rate, filter_radius, rms_mix_rate, protect, crepe_hop_length, is_webui):
+    print(f"[~] Starting voice change: {voice_model}")
     rvc_model_path, rvc_index_path = get_rvc_model(voice_model, is_webui)
     device = 'cuda:0'
     config = Config(device, True)
@@ -203,9 +220,11 @@ def voice_change(voice_model, vocals_path, output_path, pitch_change, f0_method,
     rvc_infer(rvc_index_path, index_rate, vocals_path, output_path, pitch_change, f0_method, cpt, version, net_g, filter_radius, tgt_sr, rms_mix_rate, protect, crepe_hop_length, vc, hubert_model)
     del hubert_model, cpt
     gc.collect()
+    print(f"[+] Voice change completed: {output_path}")
 
 
 def add_audio_effects(audio_path, reverb_rm_size, reverb_wet, reverb_dry, reverb_damping):
+    print(f"[~] Adding audio effects: {audio_path}")
     output_path = f'{os.path.splitext(audio_path)[0]}_mixed.wav'
 
     # Initialize audio effects plugins
@@ -225,14 +244,17 @@ def add_audio_effects(audio_path, reverb_rm_size, reverb_wet, reverb_dry, reverb
                 effected = board(chunk, f.samplerate, reset=False)
                 o.write(effected)
 
+    print(f"[+] Audio effects added: {output_path}")
     return output_path
 
 
 def combine_audio(audio_paths, output_path, main_gain, backup_gain, inst_gain, output_format):
+    print(f"[~] Combining audio: {audio_paths}")
     main_vocal_audio = AudioSegment.from_wav(audio_paths[0]) - 4 + main_gain
     backup_vocal_audio = AudioSegment.from_wav(audio_paths[1]) - 6 + backup_gain
     instrumental_audio = AudioSegment.from_wav(audio_paths[2]) - 7 + inst_gain
     main_vocal_audio.overlay(backup_vocal_audio).overlay(instrumental_audio).export(output_path, format=output_format)
+    print(f"[+] Audio combined: {output_path}")
 
 @spaces.GPU
 def song_cover_pipeline(song_input, voice_model, pitch_change, keep_files,
@@ -241,6 +263,7 @@ def song_cover_pipeline(song_input, voice_model, pitch_change, keep_files,
                         reverb_rm_size=0.15, reverb_wet=0.2, reverb_dry=0.8, reverb_damping=0.7, output_format='mp3',
                         progress=gr.Progress()):
     try:
+        print("[~] Starting AI Cover Generation Pipeline...")
         if not song_input or not voice_model:
             raise_exception('Ensure that the song input field and voice model field is filled.', is_webui)