add script

transcription_data/ami-ihm.py → ami-ihm.py

@@ -265,6 +265,10 @@ _AUDIO_ARCHIVE_URL = _BASE_DATA_URL + "audio/{subset}/{split}/{_id}.tar.gz"
 
 _ANNOTATIONS_ARCHIVE_URL = _BASE_DATA_URL + "annotations/{split}/text"
 
+_TRANSCRIPT_URL = "https://huggingface.co/datasets/distil-whisper/ami/resolve/main/transcription_data/greedy_search/"
+
+_TRANSCRIPT_URLS = _TRANSCRIPT_URL + "{config}/{split}.txt"
+
 logger = datasets.utils.logging.get_logger(__name__)
 
 
@@ -297,6 +301,7 @@ class AMI(datasets.GeneratorBasedBuilder):
                 "end_time": datasets.Value("float32"),
                 "microphone_id": datasets.Value("string"),
                 "speaker_id": datasets.Value("string"),
+                "whisper_transcript": datasets.Value("string"),
             }
         )
         return datasets.DatasetInfo(
@@ -329,6 +334,9 @@ class AMI(datasets.GeneratorBasedBuilder):
         annotations_urls = {split: _ANNOTATIONS_ARCHIVE_URL.format(split=split) for split in splits}
         annotations = dl_manager.download(annotations_urls)
 
+        transcription_urls = {split: _TRANSCRIPT_URLS.format(split=split) for split in splits}
+        transcript_archive_path = dl_manager.download(transcription_urls[self.config.name])
+
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
@@ -336,6 +344,7 @@ class AMI(datasets.GeneratorBasedBuilder):
                     "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_archives["train"]],
                     "local_extracted_archives_paths": local_extracted_archives_paths["train"],
                     "annotation": annotations["train"],
+                    "transcript_files": transcript_archive_path["train"],
                     "split": "train"
                 },
             ),
@@ -345,6 +354,7 @@ class AMI(datasets.GeneratorBasedBuilder):
                     "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_archives["dev"]],
                     "local_extracted_archives_paths": local_extracted_archives_paths["dev"],
                     "annotation": annotations["dev"],
+                    "transcript_files": transcript_archive_path["dev"],
                     "split": "dev"
                 },
             ),
@@ -354,12 +364,13 @@ class AMI(datasets.GeneratorBasedBuilder):
                     "audio_archives": [dl_manager.iter_archive(archive) for archive in audio_archives["eval"]],
                     "local_extracted_archives_paths": local_extracted_archives_paths["eval"],
                     "annotation": annotations["eval"],
+                    "transcript_files": transcript_archive_path["eval"],
                     "split": "eval"
                 },
             ),
         ]
 
-    def _generate_examples(self, audio_archives, local_extracted_archives_paths, annotation, split):
+    def _generate_examples(self, audio_archives, local_extracted_archives_paths, annotation, transcript_files, split):
         # open annotation file
         assert len(audio_archives) == len(local_extracted_archives_paths)
 
@@ -391,6 +402,14 @@ class AMI(datasets.GeneratorBasedBuilder):
                     "speaker_id": speaker_id,
                 }
 
+        whisper_transcripts = []
+
+        with open(transcript_files, encoding="utf-8") as f:
+            for row in f:
+                whisper_transcripts.append(row.rstrip("\n"))
+
+        idx = 0
+
         features = ["meeting_id", "audio_id", "text", "begin_time", "end_time", "microphone_id", "speaker_id"]
         for archive, local_archive_path in zip(audio_archives, local_extracted_archives_paths):
             for audio_path, audio_file in archive:
@@ -407,5 +426,7 @@ class AMI(datasets.GeneratorBasedBuilder):
                         "path": os.path.join(local_archive_path, audio_path) if local_archive_path else audio_path,
                         "bytes": audio_file.read(),
                     },
-                    **{feature: audio_meta[feature] for feature in features}
+                    **{feature: audio_meta[feature] for feature in features},
+                    "whisper_transcript": whisper_transcripts[idx],
                 }
+                idx += 1