rkstgr commited on
Commit
848ef96
1 Parent(s): f36edf6

Add urls of track files

Browse files
Files changed (1) hide show
  1. mtg_jamendo.py +12 -12
mtg_jamendo.py CHANGED
@@ -41,21 +41,20 @@ _HOMEPAGE = "https://github.com/MTG/mtg-jamendo-dataset"
41
  _LICENSE = "Apache License 2.0"
42
 
43
  # The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
44
- _URLS = {
45
  "train": {
46
- "df": "https://huggingface.co/datasets/rkstgr/mtg-jamendo/resolve/main/data/train.tsv",
47
- **{
48
  i: f"https://huggingface.co/datasets/rkstgr/mtg-jamendo/resolve/main/data/train/{i}.tar" for i in range(200)
49
- }
50
  },
51
  "val": {
52
- "df": "https://huggingface.co/datasets/rkstgr/mtg-jamendo/resolve/main/data/valid.tsv",
53
- **{
54
  i: f"https://huggingface.co/datasets/rkstgr/mtg-jamendo/resolve/main/data/val/{i}.tar" for i in range(22)
55
- }
56
  }
57
  }
58
 
 
 
 
 
 
59
 
60
  class MtgJamendo(datasets.GeneratorBasedBuilder):
61
  """
@@ -89,7 +88,8 @@ class MtgJamendo(datasets.GeneratorBasedBuilder):
89
  )
90
 
91
  def _split_generators(self, dl_manager):
92
- archive_path = dl_manager.download(_URLS)
 
93
 
94
  # (Optional) In non-streaming mode, we can extract the archive locally to have actual local audio files:
95
  local_extracted_archive = dl_manager.extract(archive_path) if not dl_manager.is_streaming else {}
@@ -98,10 +98,10 @@ class MtgJamendo(datasets.GeneratorBasedBuilder):
98
  return {x["id"]: x for x in xs}
99
 
100
  train_tracks = to_dict(
101
- pd.read_csv(local_extracted_archive["train"]["df"], sep="\t").to_dict("records")
102
  )
103
  valid_tracks = to_dict(
104
- pd.read_csv(local_extracted_archive["val"]["df"], sep="\t").to_dict("records")
105
  )
106
 
107
  train_splits = [
@@ -147,11 +147,11 @@ class MtgJamendo(datasets.GeneratorBasedBuilder):
147
  "instruments": eval(tracks[_id]["instruments"]),
148
  "moods": eval(tracks[_id]["moods"]),
149
  "chunk_nr": tracks[_id]["chunk_nr"],
150
- "audio": audio_path,
151
  }
152
  yield _id, data
153
 
154
 
155
  if __name__ == '__main__':
156
  mtg = MtgJamendo()
157
- mtg.download_and_prepare()
 
41
  _LICENSE = "Apache License 2.0"
42
 
43
  # The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
44
+ _URLS_DATA = {
45
  "train": {
 
 
46
  i: f"https://huggingface.co/datasets/rkstgr/mtg-jamendo/resolve/main/data/train/{i}.tar" for i in range(200)
 
47
  },
48
  "val": {
 
 
49
  i: f"https://huggingface.co/datasets/rkstgr/mtg-jamendo/resolve/main/data/val/{i}.tar" for i in range(22)
 
50
  }
51
  }
52
 
53
+ _URLS_TRACKS = {
54
+ "train": "https://huggingface.co/datasets/rkstgr/mtg-jamendo/raw/main/train.tsv",
55
+ "valid": "https://huggingface.co/datasets/rkstgr/mtg-jamendo/raw/main/valid.tsv"
56
+ }
57
+
58
 
59
  class MtgJamendo(datasets.GeneratorBasedBuilder):
60
  """
 
88
  )
89
 
90
  def _split_generators(self, dl_manager):
91
+ archive_path = dl_manager.download(_URLS_DATA)
92
+ tracks_path = dl_manager.download(_URLS_TRACKS)
93
 
94
  # (Optional) In non-streaming mode, we can extract the archive locally to have actual local audio files:
95
  local_extracted_archive = dl_manager.extract(archive_path) if not dl_manager.is_streaming else {}
 
98
  return {x["id"]: x for x in xs}
99
 
100
  train_tracks = to_dict(
101
+ pd.read_csv(tracks_path["train"], sep="\t").to_dict("records")
102
  )
103
  valid_tracks = to_dict(
104
+ pd.read_csv(tracks_path["valid"], sep="\t").to_dict("records")
105
  )
106
 
107
  train_splits = [
 
147
  "instruments": eval(tracks[_id]["instruments"]),
148
  "moods": eval(tracks[_id]["moods"]),
149
  "chunk_nr": tracks[_id]["chunk_nr"],
150
+ "audio": audio_path
151
  }
152
  yield _id, data
153
 
154
 
155
  if __name__ == '__main__':
156
  mtg = MtgJamendo()
157
+ mtg.download_and_prepare(ignore_verifications=True)