add some comments and todos

peoples_speech.py

@@ -61,35 +61,8 @@ _LICENSE = [
     "cc-by-sa-3.0", "cc-by-sa-4.0"
 ]
 
-# TODO: Add link to the official dataset URLs here
-# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
-# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
-_URLS = {
-    "clean-cc-by": {
-        "audio_tar": "",
-        "manifest": "",
-    },
-    "dirty-cc-by": {
-        "audio_tar": "",
-        "manifest": "",
-    },
-    "clean-cc-by-sa": {
-        "audio_tar": "",
-        "manifest": "",
-    },
-    "dirty-cc-by-sa": {
-        "audio_tar": "",
-        "manifest": "",
-    },
-    "microset": {
-        "audio_tar": "",
-        "manifest": "",
-    },
-}
-
-# _BASE_URL = "https://huggingface.co/datasets/MLCommons/peoples_speech/resolve/main/"
-
 # relative path to data inside dataset's repo
+# TODO: change according to the scheme of generating urls to the audio archives
 _DATA_URL = "{config}/{config}_00000{archive_id}.tar"
 
 # relative path to metadata inside dataset's repo
@@ -101,6 +74,7 @@ class PeoplesSpeech(datasets.GeneratorBasedBuilder):
 
     VERSION = datasets.Version("1.1.0")
     BUILDER_CONFIGS = [
+        # TODO: add "subset" config
         datasets.BuilderConfig(name="clean", version=VERSION, description="Clean, CC-BY licensed subset."),
         datasets.BuilderConfig(name="dirty", version=VERSION, description="Dirty, CC-BY licensed subset."),
         datasets.BuilderConfig(name="clean_sa", version=VERSION, description="Clean, CC-BY-SA licensed subset."),
@@ -132,17 +106,30 @@ class PeoplesSpeech(datasets.GeneratorBasedBuilder):
         # TODO: this should be changed to the actual number of archives further
         urls = [_DATA_URL.format(config=self.config.name, archive_id=i) for i in range(5)]
         archive_paths = [dl_manager.download(url) for url in urls]
+
+        # In non-streaming mode, we extract the archives to have the data locally:
         local_extracted_archive_paths = [dl_manager.extract(path) for path in archive_paths] \
             if not dl_manager.is_streaming else [None] * len(archive_paths)
 
         manifest_url = _MANIFEST_URL.format(config=self.config.name)
         manifest_path = dl_manager.download_and_extract(manifest_url)
 
+        # To access the audio data from the TAR archives using the download manager,
+        # we have to use the dl_manager.iter_archive method
+        #
+        # This is because dl_manager.download_and_extract
+        # doesn't work to stream TAR archives in streaming mode.
+        # (we have to stream the files of a TAR archive one by one)
+        #
+        # The iter_archive method returns an iterable of (path_within_archive, file_obj) for every
+        # file in a TAR archive.
+
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
                 gen_kwargs={
                     "local_extracted_archive_paths": local_extracted_archive_paths,
+                    # use iter_archive here to access the files in the TAR archives:
                     "archives": [dl_manager.iter_archive(path) for path in archive_paths],
                     "manifest_path": manifest_path
                 },
@@ -165,9 +152,11 @@ class PeoplesSpeech(datasets.GeneratorBasedBuilder):
                         "duration_ms": duration
                     }
 
-        print("generating examples")
         for local_extracted_archive_path, archive in zip(local_extracted_archive_paths, archives):
+            # Here we iterate over all the files within the TAR archive:
             for audio_filename, audio_file in archive:
+                # if an audio file exists locally (i.e. in default, non-streaming mode) set the full path to it
+                # joining path to directory that the archive was extracted to and audio filename.
                 path = os.path.join(local_extracted_archive_path, audio_filename) if local_extracted_archive_path \
                     else audio_filename
                 yield audio_filename, {