upd py

chest_falsetto.py

@@ -1,4 +1,5 @@
 import os
+import socket
 import random
 import datasets
 from datasets.tasks import AudioClassification
@@ -9,7 +10,9 @@ _NAMES = {
     'singing_method': ['falsetto', 'chest']
 }
 
-_HOMEPAGE = f"https://huggingface.co/datasets/ccmusic-database/{os.path.basename(__file__).split('.')[0]}"
+_NAME = os.path.basename(__file__).split('.')[0]
+
+_HOMEPAGE = f"https://huggingface.co/datasets/ccmusic-database/{_NAME}"
 
 _CITATION = """\
 @dataset{zhaorui_liu_2021_5676893,
@@ -31,8 +34,6 @@ the Mel-spectrogram, MFCC, and spectral characteristics of each audio segment ar
 for a total of 5120 CSV files.
 """
 
-_URL = f"{_HOMEPAGE}/resolve/main/data/client_data.zip"
-
 
 class chest_falsetto(datasets.GeneratorBasedBuilder):
     def _info(self):
@@ -40,7 +41,9 @@ class chest_falsetto(datasets.GeneratorBasedBuilder):
             features=datasets.Features(
                 {
                     "audio": datasets.Audio(sampling_rate=44_100),
-                    "image": datasets.Image(),
+                    "mel": datasets.Image(),
+                    "cqt": datasets.Image(),
+                    "chroma": datasets.Image(),
                     "label": datasets.features.ClassLabel(names=_NAMES['all']),
                     "gender": datasets.features.ClassLabel(names=_NAMES['gender']),
                     "singing_method": datasets.features.ClassLabel(names=_NAMES['singing_method']),
@@ -60,15 +63,54 @@ class chest_falsetto(datasets.GeneratorBasedBuilder):
             ],
         )
 
+    def _cdn_url(self, ip='127.0.0.1', port=80):
+        try:
+            # easy for local test
+            with socket.create_connection((ip, port), timeout=5):
+                return {
+                    'image': f'http://{ip}/{_NAME}/data/data.zip',
+                    'audio': f'http://{ip}/{_NAME}/data/raw_data.zip'
+                }
+
+        except (socket.timeout, socket.error):
+            return {
+                'image': f"{_HOMEPAGE}/resolve/main/data/data.zip",
+                'audio': f"{_HOMEPAGE}/resolve/main/data/raw_data.zip"
+            }
+
     def _split_generators(self, dl_manager):
-        data_files = dl_manager.download_and_extract(_URL)
-        files = dl_manager.iter_files([data_files])
-        dataset = []
+        audio_files = dl_manager.download_and_extract(self._cdn_url()['audio'])
+        wav_files = dl_manager.iter_files([audio_files])
+        data = {}
+
+        for wav_file in wav_files:
+            fname = os.path.basename(wav_file)
+            if fname.endswith(".wav"):
+                fname = fname[:-4]
+                data[fname] = {
+                    'audio': wav_file,
+                    'mel': '',
+                    'cqt': '',
+                    'chroma': ''
+                }
+
+        img_files = dl_manager.download_and_extract(self._cdn_url()['image'])
+        jpg_files = dl_manager.iter_files([img_files])
+        for jpg_file in jpg_files:
+            fname = os.path.basename(jpg_file)
+            if fname.endswith(".jpg"):
+                fname = fname[:-4]
+                dirname = os.path.dirname(jpg_file)
+                if 'mel' in dirname:
+                    data[fname]['mel'] = jpg_file
+
+                elif 'cqt' in dirname:
+                    data[fname]['cqt'] = jpg_file
 
-        for path in files:
-            if os.path.basename(path).endswith(".wav"):
-                dataset.append(path)
+                elif 'chroma' in dirname:
+                    data[fname]['chroma'] = jpg_file
 
+        dataset = list(data.values())
         random.shuffle(dataset)
         data_count = len(dataset)
         p80 = int(data_count * 0.8)
@@ -97,12 +139,14 @@ class chest_falsetto(datasets.GeneratorBasedBuilder):
 
     def _generate_examples(self, files):
         for i, path in enumerate(files):
-            file_name = os.path.basename(path)
+            file_name = os.path.basename(path['audio'])
             sex = file_name.split('_')[1]
             method = file_name.split('_')[2][:-4]
             yield i, {
-                "audio": path,
-                "image": path.replace('.wav', '.jpg').replace('/audio/', '/image/').replace('\\audio\\', '\\image\\'),
+                "audio": path['audio'],
+                "mel": path['mel'],
+                "cqt": path['cqt'],
+                "chroma": path['chroma'],
                 "label": f'{sex}_{method}',
                 "gender": 'male' if sex == 'm' else 'female',
                 "singing_method": method,

data/raw_data.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2c64f8880652961823f722255393ebe6bced54f2f88480a0001c12235d925118
-size 62675969
+oid sha256:8f14eb75b7270d022bab15bd164ac20cb84a607d1bbf6e74d78e81882d509e33
+size 39263817