refactor_loader_script

MOCKS.py

@@ -41,401 +41,164 @@ audio testset for evaluation and benchmarking Open-Vocabulary Keyword Spotting (
 
 
 _BASE_URL = "https://huggingface.co/datasets/voiceintelligenceresearch/MOCKS/tree/main"
-_DL_URLS = {
-    "de.MCV": {
-        "offline": "de/MCV/test/offline/data.tar.gz",
-        "online": "de/MCV/test/online/data.tar.gz",
-        "offline_transcription" : "de/MCV/test/data_offline_transcription.tsv",
-        "online_transcription" : "de/MCV/test/data_online_transcription.tsv",
-    },
-    "en.LS-clean": {
-        "offline": "en/LS-clean/test/offline/data.tar.gz",
-        "online": "en/LS-clean/test/online/data.tar.gz",
-        "offline_transcription" : "en/LS-clean/test/data_offline_transcription.tsv",
-        "online_transcription" : "en/LS-clean/test/data_online_transcription.tsv",
-    },
-    "en.LS-other": {
-        "offline": "en/LS-other/test/offline/data.tar.gz",
-        "online": "en/LS-other/test/online/data.tar.gz",
-        "offline_transcription" : "en/LS-other/test/data_offline_transcription.tsv",
-        "online_transcription" : "en/LS-other/test/data_online_transcription.tsv",
-    },
-    "en.MCV": {
-        "offline": "en/MCV/test/offline/data.tar.gz",
-        "online": "en/MCV/test/online/data.tar.gz",
-        "offline_transcription" : "en/MCV/test/data_offline_transcription.tsv",
-        "online_transcription" : "en/MCV/test/data_online_transcription.tsv",
-    },
-    "es.MCV": {
-        "offline": "es/MCV/test/offline/data.tar.gz",
-        "online": "es/MCV/test/online/data.tar.gz",
-        "offline_transcription" : "es/MCV/test/data_offline_transcription.tsv",
-        "online_transcription" : "es/MCV/test/data_online_transcription.tsv",
-    },
-    "fr.MCV": {
-        "offline": "fr/MCV/test/offline/data.tar.gz",
-        "online": "fr/MCV/test/online/data.tar.gz",
-        "offline_transcription": "fr/MCV/test/data_offline_transcription.tsv",
-        "online_transcription": "fr/MCV/test/data_online_transcription.tsv",
-    },
-    "it.MCV": {
-        "offline": "it/MCV/test/offline/data.tar.gz",
-        "online": "it/MCV/test/online/data.tar.gz",
-        "offline_transcription": "it/MCV/test/data_offline_transcription.tsv",
-        "online_transcription": "it/MCV/test/data_online_transcription.tsv",
-    },
-    "all": {
-        "de.MCV.offline": "de/MCV/test/offline/data.tar.gz",
-        "de.MCV.online": "de/MCV/test/online/data.tar.gz",
-        "en.LS-clean.offline": "en/LS-clean/test/offline/data.tar.gz",
-        "en.LS-clean.online": "en/LS-clean/test/online/data.tar.gz",
-        "en.LS-other.offline": "en/LS-other/test/offline/data.tar.gz",
-        "en.LS-other.online": "en/LS-other/test/online/data.tar.gz",
-        "en.MCV.offline": "en/MCV/test/offline/data.tar.gz",
-        "en.MCV.online": "en/MCV/test/online/data.tar.gz",
-        "es.MCV.offline": "es/MCV/test/offline/data.tar.gz",
-        "es.MCV.online": "es/MCV/test/online/data.tar.gz",
-        "fr.MCV.offline": "fr/MCV/test/offline/data.tar.gz",
-        "fr.MCV.online": "fr/MCV/test/online/data.tar.gz",
-        "it.MCV.offline": "it/MCV/test/offline/data.tar.gz",
-        "it.MCV.online": "it/MCV/test/online/data.tar.gz",
-        "de.MCV.offline_transcription": "de/MCV/test/data_offline_transcription.tsv",
-        "de.MCV.online_transcription": "de/MCV/test/data_online_transcription.tsv",
-        "en.LS-clean.offline_transcription": "en/LS-clean/test/data_offline_transcription.tsv",
-        "en.LS-clean.online_transcription": "en/LS-clean/test/data_online_transcription.tsv",
-        "en.LS-other.offline_transcription": "en/LS-other/test/data_offline_transcription.tsv",
-        "en.LS-other.online_transcription": "en/LS-other/test/data_online_transcription.tsv",
-        "en.MCV.offline_transcription": "en/MCV/test/data_offline_transcription.tsv",
-        "en.MCV.online_transcription": "en/MCV/test/data_online_transcription.tsv",
-        "es.MCV.offline_transcription": "es/MCV/test/data_offline_transcription.tsv",
-        "es.MCV.online_transcription": "es/MCV/test/data_online_transcription.tsv",
-        "fr.MCV.offline_transcription": "fr/MCV/test/data_offline_transcription.tsv",
-        "fr.MCV.online_transcription": "fr/MCV/test/data_online_transcription.tsv",
-        "it.MCV.offline_transcription": "it/MCV/test/data_offline_transcription.tsv",
-        "it.MCV.online_transcription": "it/MCV/test/data_online_transcription.tsv",
-    }
+_DL_URLS_TEMPLATE = {
+    "data": "%s/%s/test/%s/data.tar.gz",
+    "transcription" : "%s/%s/test/data_%s_transcription.tsv",
+    "positive" : "%s/%s/test/%s/all.pair.positive.tsv",
+    "similar" : "%s/%s/test/%s/all.pair.similar.tsv",
+    "different" : "%s/%s/test/%s/all.pair.different.tsv",
+    "positive_subset" : "%s/%s/test/%s/subset.pair.positive.tsv",
+    "similar_subset" : "%s/%s/test/%s/subset.pair.similar.tsv",
+    "different_subset" : "%s/%s/test/%s/subset.pair.different.tsv",
 }
 
+_MOCKS_SETS = [
+    "en.LS-clean",
+    "en.LS-other",
+    "en.MCV",
+    "de.MCV",
+    "es.MCV",
+    "fr.MCV",
+    "it.MCV"]
+
+_MOCKS_SUFFIXES = [
+    "",
+    ".positive",
+    ".similar",
+    ".different",
+    ".subset",
+    ".positive_subset",
+    ".similar_subset",
+    ".different_subset"]
+
 
 class Mocks(datasets.GeneratorBasedBuilder):
     """Mocks Dataset."""
-    DEFAULT_CONFIG_NAME = "all"
-
-    BUILDER_CONFIGS = [
-        datasets.BuilderConfig(name="de.MCV", description="German Mozilla Common Voice."),
-        datasets.BuilderConfig(name="en.LS-clean", description="English LibriSpeech 'Clean'."),
-        datasets.BuilderConfig(name="en.LS-other", description="English LibriSpeech 'Other'."),
-        datasets.BuilderConfig(name="en.MCV", description="English Mozilla Common Voice."),
-        datasets.BuilderConfig(name="es.MCV", description="Spanish Mozilla Common Voice."),
-        datasets.BuilderConfig(name="fr.MCV", description="French Mozilla Common Voice."),
-        datasets.BuilderConfig(name="it.MCV", description="Italian Mozilla Common Voice."),
-        datasets.BuilderConfig(name="all", description="All test set."),
-    ]
+    DEFAULT_CONFIG_NAME = "en.LS-clean"
+
+    BUILDER_CONFIGS = [datasets.BuilderConfig(name=subset+suffix, description=subset+suffix)
+            for subset in _MOCKS_SETS for suffix in _MOCKS_SUFFIXES]
 
     def _info(self):
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
-            features=datasets.Features(
-                {
-                    "audio_id": datasets.Value("string"),
-                    "audio": datasets.Audio(sampling_rate=16_000),
-                    "transcription": datasets.Value("string"),
+            features=datasets.Features({
+                    "keyword_id": datasets.Value("string"),
+                    "keyword_transcription": datasets.Value("string"),
+                    "test_id": datasets.Value("string"),
+                    "test_transcription": datasets.Value("string"),
+                    "test_audio": datasets.Audio(sampling_rate=16000),
+                    "label": datasets.Value("bool"),
                 }
             ),
             homepage=_BASE_URL,
             citation=_CITATION
         )
 
+
     def _split_generators(self, dl_manager):
-        archive_path = dl_manager.download(_DL_URLS[self.config.name])
-
-        if self.config.name == "de.MCV":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-
-        elif self.config.name == "en.LS-clean":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-
-        elif self.config.name == "en.LS-other":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]        
-
-        elif self.config.name == "en.MCV":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-
-        elif self.config.name == "es.MCV":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-        
-        elif self.config.name == "fr.MCV":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-
-        elif self.config.name == "it.MCV":
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["offline"]),
-                        "transcription": archive_path["offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["online"]),
-                        "transcription": archive_path["online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-
-        elif self.config.name == "all": 
-            offline_split = [
-                datasets.SplitGenerator(
-                    name="de.MCV.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["de.MCV.offline"]),
-                        "transcription": archive_path["de.MCV.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="en.LS-clean.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["en.LS-clean.offline"]),
-                        "transcription": archive_path["en.LS-clean.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="en.LS-other.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["en.LS-other.offline"]),
-                        "transcription": archive_path["en.LS-other.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="en.MCV.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["en.MCV.offline"]),
-                        "transcription": archive_path["en.MCV.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="es.MCV.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["es.MCV.offline"]),
-                        "transcription": archive_path["es.MCV.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="fr.MCV.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["fr.MCV.offline"]),
-                        "transcription": archive_path["fr.MCV.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="it.MCV.offline",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["it.MCV.offline"]),
-                        "transcription": archive_path["it.MCV.offline_transcription"],
-                        "s_type": "offline"
-                    }
-                )
-            ]
-            online_split = [
-                datasets.SplitGenerator(
-                    name="de.MCV.online",
-                    gen_kwargs={
-                        "transcription": archive_path["de.MCV.offline_transconline"],
-                        "s_type": "online"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="en.LS-clean.online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["en.LS-clean.online"]),
-                        "transcription": archive_path["en.LS-clean.online_transcription"],
-                        "s_type": "online"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="en.LS-other.online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["en.LS-other.online"]),
-                        "transcription": archive_path["en.LS-other.online_transcription"],
-                        "s_type": "online"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="en.MCV.online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["en.MCV.online"]),
-                        "transcription": archive_path["en.MCV.online_transcription"],
-                        "s_type": "online"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="es.MCV.online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["es.MCV.online"]),
-                        "transcription": archive_path["es.MCV.online_transcription"],
-                        "s_type": "online"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="fr.MCV.online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["fr.MCV.online"]),
-                        "transcription": archive_path["fr.MCV.online_transcription"],
-                        "s_type": "online"
-                    }
-                ),
-                datasets.SplitGenerator(
-                    name="it.MCV.online",
-                    gen_kwargs={
-                        "audio_files": dl_manager.iter_archive(archive_path["it.MCV.online"]),
-                        "transcription": archive_path["it.MCV.online_transcription"],
-                        "s_type": "online"
-                    }
-                )
-            ]
-
-        return online_split + offline_split
-
-    def _generate_examples(self, audio_files, transcription, s_type):
-        """Lorem ipsum."""
-        metadata = {}
-        with open(transcription, encoding="utf-8") as f:
-            f = csv.reader(f, delimiter="\t")
-            for row in f:
-                audio_id = row[0].split("/")[-1]
-                keyword_transcription = row[1]
-                metadata[audio_id] = {"audio_id": audio_id, "transcription": keyword_transcription}
+        logger.info("split_generator")
+        name_split = self.config.name.split(".")
+        subset_lang = name_split[0]
+        subset_name = name_split[1]
+
+        if len(name_split) == 2:
+            pairs_types = ["positive", "similar", "different"]
+        elif name_split[2] == "subset":
+            pairs_types = ["positive_subset", "similar_subset", "different_subset"]
+        else:
+            pairs_types = [name_split[2]]
+
+        offline_archive_path = dl_manager.download({
+            k: v%(subset_lang, subset_name, "offline") 
+            for k, v in _DL_URLS_TEMPLATE.items()
+            })
+        online_archive_path = dl_manager.download({
+            k: v%(subset_lang, subset_name, "online") 
+            for k, v in _DL_URLS_TEMPLATE.items()
+            })
+
+        split_offline = [datasets.SplitGenerator(
+                name="offline",
+                gen_kwargs={
+                    "audio_files": dl_manager.iter_archive(offline_archive_path["data"]),
+                    "transcription_keyword": offline_archive_path["transcription"],
+                    "transcription_test": offline_archive_path["transcription"],
+                    "pairs": [offline_archive_path[pair_type] for pair_type in pairs_types],
+                }
+            )
+        ]
+
+        split_online = [datasets.SplitGenerator(
+                name="online",
+                gen_kwargs={
+                    "audio_files": dl_manager.iter_archive(online_archive_path["data"]),
+                    "transcription_keyword": offline_archive_path["transcription"],
+                    "transcription_test": online_archive_path["transcription"],
+                    "pairs": [online_archive_path[pair_type] for pair_type in pairs_types],
+                }
+            )
+        ]
+
+        return split_offline + split_online
+
+
+    def _read_transcription(self, transcription_path):
+        transcription_metadata = {}
+
+        with open(transcription_path, encoding="utf-8") as f:
+            reader = csv.reader(f, delimiter="\t")
+            next(reader, None)
+
+            for row in reader:
+                _, audio_id = os.path.split(row[0])
+                transcription = row[1]
+                transcription_metadata[audio_id] = {
+                    "audio_id": audio_id,
+                    "transcription": transcription}
+
+        return transcription_metadata
+
+
+    def _generate_examples(self, audio_files, transcription_keyword, transcription_test, pairs):
+        transcription_keyword_metadata = self._read_transcription(transcription_keyword)
+
+        transcription_test_metadata = self._read_transcription(transcription_test)
+
+        pair_metadata = {}
+        for pair in pairs:
+            with open(pair, encoding="utf-8") as f:
+                reader = csv.reader(f, delimiter="\t")
+                next(reader, None)
+
+                for row in reader:
+                    _, keyword_id = os.path.split(row[0])
+                    _, test_id = os.path.split(row[1])
+
+                    if keyword_id not in transcription_keyword_metadata:
+                        logger.error("No transcription and audio for keyword %s"%(keyword_id))
+                        continue
+                    if test_id not in transcription_test_metadata:
+                        logger.error("No transcription and audio for test case %s"%(test_id))
+                        continue
+
+                    if test_id not in pair_metadata:
+                        pair_metadata[test_id] = []
+
+                    pair_metadata[test_id].append([keyword_id, int(row[-1])])
 
         id_ = 0
-        for path, f in audio_files:
-            _, audio_name = os.path.split(path)
-            if audio_name in metadata:
-                audio = {"bytes": f.read()}
-                yield id_, {**metadata[audio_name], "audio": audio}
-                id_ +=1
+        for test_path, test_f in audio_files:
+            _, test_id = os.path.split(test_path)
+            if test_id in pair_metadata:
+                test_audio = {"bytes": test_f.read()}
+                for keyword_id, label in pair_metadata[test_id]:
+                    yield id_, {
+                        "keyword_id": keyword_id,
+                        "keyword_transcription": transcription_keyword_metadata[keyword_id]["transcription"],
+                        "test_id": test_id,
+                        "test_transcription": transcription_test_metadata[test_id]["transcription"],
+                        "test_audio": test_audio,
+                        "label": label}
+                    id_ += 1
+