Upload dave.py with huggingface_hub

dave.py

@@ -1,17 +1,11 @@
 import json
-import os
-from pathlib import Path
-from typing import Dict, List, Tuple
 
 import datasets
 
-_CITATION = ""  # Optional
-_DESCRIPTION = "A multimodal dataset including video/audio and rich annotations."
+_DESCRIPTION = "Dataset with video and audio references for epic and ego4d tasks."
 _HOMEPAGE = "https://huggingface.co/datasets/gorjanradevski/dave"
 _LICENSE = "MIT"
-_URL = "https://huggingface.co/datasets/gorjanradevski/dave/resolve/main/epic.json"
 
-# These fields contain file references
 _MEDIA_FIELDS = [
     "raw_video_path",
     "compressed_video_path",
@@ -21,9 +15,7 @@ _MEDIA_FIELDS = [
     "overlayed_audio_path",
 ]
 
-class Dave(datasets.GeneratorBasedBuilder):
-    VERSION = datasets.Version("1.0.0")
-
+class DaveDataset(datasets.GeneratorBasedBuilder):
     def _info(self):
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
@@ -31,7 +23,7 @@ class Dave(datasets.GeneratorBasedBuilder):
                 "raw_video_path": datasets.Value("string"),
                 "compressed_video_path": datasets.Value("string"),
                 "overlayed_event_index": datasets.Value("int32"),
-                "events": datasets.features.Sequence({
+                "events": datasets.Sequence({
                     "start": datasets.Value("string"),
                     "end": datasets.Value("string"),
                     "duration": datasets.Value("float64"),
@@ -46,64 +38,67 @@ class Dave(datasets.GeneratorBasedBuilder):
                 "video_id": datasets.Value("string"),
                 "participant_id": datasets.Value("string"),
                 "type": datasets.Value("string"),
-                "raw_choices_simple_audio_classification": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_overlayed_full_audio_classification": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_video_segment": datasets.features.Sequence(datasets.Value("string")),
-                "correct_temporal_order": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_temporal_video": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_multimodal": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_silent_video": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_audio": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_text_only": datasets.features.Sequence(datasets.Value("string")),
-                "raw_choices_pipeline_event_classification": datasets.features.Sequence(datasets.Value("string")),
+                "raw_choices_simple_audio_classification": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_overlayed_full_audio_classification": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_video_segment": datasets.Sequence(datasets.Value("string")),
+                "correct_temporal_order": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_temporal_video": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_multimodal": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_silent_video": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_audio": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_text_only": datasets.Sequence(datasets.Value("string")),
+                "raw_choices_pipeline_event_classification": datasets.Sequence(datasets.Value("string")),
             }),
-            supervised_keys=None,
             homepage=_HOMEPAGE,
             license=_LICENSE,
         )
 
     def _split_generators(self, dl_manager):
-        json_path = dl_manager.download(_URL)
+        base_url = "https://huggingface.co/datasets/gorjanradevski/dave/resolve/main/"
+        json_files = {
+            "epic": "epic.json",
+            "ego4d": "ego4d.json"
+        }
 
-        # Load JSON to collect file URLs
-        with open(json_path, "r", encoding="utf-8") as f:
-            data = json.load(f)
+        split_generators = []
+        for split_name, json_file in json_files.items():
+            json_path = dl_manager.download(base_url + json_file)
 
-        file_urls = set()
-        for item in data:
-            for field in _MEDIA_FIELDS:
-                if field in item:
-                    file_urls.add(item[field])
+            with open(json_path, "r", encoding="utf-8") as f:
+                data = json.load(f)
 
-        # Prefix all file paths with hf:// if they are not already
-        hf_file_urls = {
-            f: f.replace("files/", "hf://datasets/gorjanradevski/dave@main/files/")
-            for f in file_urls
-            if "P02" in f
-        }
+            file_paths = set()
+            for item in data:
+                for field in _MEDIA_FIELDS:
+                    if field in item:
+                        file_paths.add(item[field])
+
+            hf_file_urls = {
+                f: f.replace("files/", "hf://datasets/gorjanradevski/dave@main/files/")
+                for f in file_paths
+            }
 
-        # Download media files
-        downloaded_files = dl_manager.download(hf_file_urls)
+            downloaded_files = dl_manager.download(hf_file_urls)
 
-        return [
-            datasets.SplitGenerator(
-                name=datasets.Split.TRAIN,
-                gen_kwargs={
-                    "json_path": json_path,
-                    "downloaded_files": downloaded_files,
-                },
+            split_generators.append(
+                datasets.SplitGenerator(
+                    name=split_name,
+                    gen_kwargs={
+                        "json_path": json_path,
+                        "downloaded_files": downloaded_files,
+                    },
+                )
             )
-        ]
 
-    def _generate_examples(self, json_path, downloaded_files: Dict[str, str]):
+        return split_generators
+
+    def _generate_examples(self, json_path, downloaded_files):
         with open(json_path, "r", encoding="utf-8") as f:
             data = json.load(f)
 
         for idx, item in enumerate(data):
-            # Replace file references with local downloaded paths
             for field in _MEDIA_FIELDS:
                 if field in item and item[field] in downloaded_files:
                     item[field] = downloaded_files[item[field]]
 
             yield idx, item
-