Update czech_parliament_plenary_hearings.py

czech_parliament_plenary_hearings.py

@@ -1,10 +1,8 @@
+from datasets import Dataset, GeneratorBasedBuilder, Features
 import os
 import tarfile
-
-import soundfile as sf
-
+import librosa
 import datasets
-
 _LICENSE = "https://creativecommons.org/licenses/by/4.0/"
 _HOMEPAGE = "https://lindat.mff.cuni.cz/repository/xmlui/handle/11234/1-3126"
 _DATASET_URL = "https://lindat.mff.cuni.cz/repository/xmlui/bitstream/handle/11234/1-3126/snemovna.tar.xz"
@@ -23,8 +21,9 @@ _CITATION = """\
  year = {2019} } """
 
 
-class MyAudioDataset(datasets.GeneratorBasedBuilder):
-    VERSION = datasets.Version("1.0.0")
+class CzechParliamentPlenaryHearings(GeneratorBasedBuilder):
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
 
     def _info(self):
         return datasets.DatasetInfo(
@@ -32,9 +31,8 @@ class MyAudioDataset(datasets.GeneratorBasedBuilder):
             features=datasets.Features(
                 {
                     "id": datasets.Value("string"),
-                    "audio": datasets.features.Audio(),
-                    "transcription": datasets.Value("string"),
-                    "audio_sampling_rate": datasets.Value("int32"),
+                    "audio": datasets.features.Audio(sampling_rate=16000),
+                    "transcription": datasets.Value("string")
                 }
             ),
             supervised_keys=None,
@@ -44,40 +42,37 @@ class MyAudioDataset(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        archive_path = dl_manager.download_and_extract(_DATASET_URL)
+        data_dir = dl_manager.download_and_extract(_DATASET_URL)
+        data_dir = os.path.join(data_dir, 'ASR_DATA')
+        splits = ("train", "dev", "test")
 
-        return [
-            datasets.SplitGenerator(
-                name=datasets.Split.TRAIN,
-                gen_kwargs={"directory": os.path.join(
-                    archive_path, "ASR_DATA", "train")},
-            ),
-            datasets.SplitGenerator(
-                name=datasets.Split.VALIDATION,
-                gen_kwargs={"directory": os.path.join(
-                    archive_path, "ASR_DATA", "dev")},
-            ),
-            datasets.SplitGenerator(
-                name=datasets.Split.TEST,
-                gen_kwargs={"directory": os.path.join(
-                    archive_path, "ASR_DATA", "test")},
-            ),
-        ]
+        split_names = {
+            "train": datasets.Split.TRAIN,
+            "dev": datasets.Split.VALIDATION,
+            "test": datasets.Split.TEST,
+        }
+        split_generators = []
+        for split in splits:
+            split_generators.append(
+                 name=split_names.get(split, split),
+                 gen_kwargs={'split': split, 'data_dir': data_dir}
+            )
+        return split_generators
 
-    def _generate_examples(self, directory):
-        for root, _, files in os.walk(directory):
-            for filename in files:
-                if filename.endswith(".wav"):
-                    audio_path = os.path.join(root, filename)
-                    transcription_path = os.path.join(root, filename + ".trn")
+    def _generate_examples(self, split, data_dir):
+        split_dir = os.path.join(data_dir, split)
+        for folder_name in os.listdir(split_dir):
+            folder_path = os.path.join(split_dir, folder_name)
+            if os.path.isdir(folder_path):
+                for audio_file in os.listdir(folder_path):
+                    if audio_file.endswith('.wav'):
+                        audio_path = os.path.join(folder_path, audio_file)
+                        transcription_path = os.path.join(folder_path, audio_file + '.trn')
+                        transcription = open(transcription_path).read().strip()
 
-                    # Load audio and transcription
-                    audio, sampling_rate = sf.read(audio_path)
-                    with open(transcription_path, "r", encoding="utf-8") as f:
-                        transcription = f.read().strip()
+                        audio, sr = librosa.load(audio_path, sr=16000)
 
-                    yield f"{audio_path}-{transcription}", {
-                        "audio": audio,
-                        "transcription": transcription,
-                        "audio_sampling_rate": sampling_rate,
-                    }
+                        yield f"{folder_name}/{audio_file}", {
+                            'audio': audio,
+                            'transcription': transcription,
+                        }