Delete CAMEO.py

CAMEO.py

@@ -1,118 +0,0 @@
-import os
-import json
-
-from datasets import (
-    Audio,
-    BuilderConfig,
-    DatasetInfo,
-    Features,
-    GeneratorBasedBuilder,
-    SplitGenerator,
-    Value,
-    Version,
-)
-
-from datasets.download.download_manager import DownloadManager
-
-
-_DATASETS = [
-    "CaFE", "CREMA-D", "EMNS", "Emozionalmente", "eNTERFACE",
-    "JL-Corpus", "MESD", "nEMO", "Oreau", "PAVOQUE",
-    "RAVDESS", "RESD", "SUBESCO",
-]
-
-_CONFIGS = _DATASETS + ["all"]
-
-_SPLITS = ["test"]
-
-
-class CAMEOConfig(BuilderConfig):
-    def __init__(self, name, **kwargs):
-        super(CAMEOConfig, self).__init__(
-            name=name,
-            version=Version("1.0.0", ""),
-            description="",
-            **kwargs,
-        )
-
-
-class CAMEODataset(GeneratorBasedBuilder):
-    DEFAULT_CONFIG_NAME = "all"
-    BUILDER_CONFIGS = [CAMEOConfig(name=config) for config in _CONFIGS]
-
-    def _info(self):
-        return DatasetInfo(
-            description="",
-            features=Features(
-                {
-                    "file_id": Value("string"),
-                    "audio": Audio(sampling_rate=16_000),
-                    "emotion": Value("string"),
-                    "transcription": Value("string"),
-                    "speaker_id": Value("string"),
-                    "gender": Value("string"),
-                    "age": Value("string"),
-                    "dataset": Value("string"),
-                    "language": Value("string"),
-                    "license": Value("string"),
-                }
-            ),
-            supervised_keys=None,
-            homepage="",
-            license="",
-            citation="",
-        )
-    
-    def _split_generators(self, dl_manager: DownloadManager) -> list[SplitGenerator]:
-        if self.config.name == "all":
-            audio_urls = {split: [f"data/{dataset}/test.tar.gz" for dataset in _DATASETS] for split in _SPLITS}
-            meta_urls = {split: [f"data/{dataset}/test.jsonl" for dataset in _DATASETS] for split in _SPLITS}
-
-        else:
-            audio_urls = {split: [f"data/{self.config.name}/test.tar.gz"] for split in _SPLITS}
-            meta_urls = {split: [f"data/{self.config.name}/test.jsonl"] for split in _SPLITS}
-
-        audio_paths = dl_manager.download(audio_urls)
-        audio_iters = {split: [dl_manager.iter_archive(path) for path in paths] for split, paths in audio_paths.items()}
-
-        local_extracted_archive = dl_manager.extract(audio_paths) if not dl_manager.is_streaming else {
-            split: [None] * len(paths) for split, paths in audio_paths.items()
-        }
-        path_to_clips = "CAMEO/data"
-
-        return [
-            SplitGenerator(
-                name=split,
-                gen_kwargs={
-                    "local_extracted_archive": local_extracted_archive.get(split),
-                    "audio_files": audio_iters.get(split),
-                    "metadata_path": dl_manager.download_and_extract(meta_urls.get(split)),
-                    "path_to_clips": os.path.join(path_to_clips, self.config.name, split),
-                },
-            ) for split in _SPLITS
-        ]
-    
-    def _generate_examples(self, local_extracted_archive, audio_files, metadata_path, path_to_clips):
-        data_fields = list(self._info().features.keys())
-        metadata = {}
-        id_ = 0
-
-        for metadata_pth, audio_fs, local_archive in zip(metadata_path, audio_files, local_extracted_archive):
-            with open(metadata_pth, "r", encoding="utf-8") as fp:
-                for row in fp:
-                    row = json.loads(row.strip())
-                    row["audio"] = os.path.join(path_to_clips, row["file_id"])
-
-                    for field in data_fields:
-                        if field not in row:
-                            row[field] = ""
-                            
-                    metadata[row["audio"]] = row
-                
-            for path, f in audio_fs:
-                result = dict(metadata[os.path.join(path_to_clips, path.split("/")[-1])])
-                path = os.path.join(local_archive, path) if local_archive else path
-                result["audio"] = {"path": path, "bytes": f.read()}
-                yield id_, result
-                id_ += 1
-