Add files using upload-large-folder tool

mixbench.py

@@ -3,95 +3,83 @@ import json
 import csv
 import datasets
 
-_DESCRIPTION = "MixBench: Mixed-Modality Retrieval Benchmark"
+_DESCRIPTION = """
+MixBench is a benchmark for mixed-modality retrieval across text, image, and image+text corpora.
+Each config corresponds to a (dataset, split) pair.
+"""
+
 _HOMEPAGE = "https://huggingface.co/datasets/mixed-modality-search/MixBench25"
 
+_SUBSETS = ["MSCOCO", "Google_WIT", "VisualNews", "OVEN"]
+_SPLITS = ["query", "corpus", "mixed_corpus", "qrel"]
+
 class MixBench(datasets.GeneratorBasedBuilder):
     BUILDER_CONFIGS = [
-        datasets.BuilderConfig(name=name, version=datasets.Version("1.0.0"), description=f"Subset = {name}")
-        for name in ["MSCOCO", "Google_WIT", "VisualNews", "OVEN"]
+        datasets.BuilderConfig(
+            name=f"{subset}_{split}",
+            version=datasets.Version("1.0.0"),
+            description=f"{subset} - {split}"
+        )
+        for subset in _SUBSETS
+        for split in _SPLITS
     ]
 
     def _info(self):
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
+            homepage=_HOMEPAGE,
             features=datasets.Features({
-                "id": datasets.Value("string"),
+                "query_id": datasets.Value("string"),
+                "corpus_id": datasets.Value("string"),
                 "text": datasets.Value("string"),
-            }),
-            homepage=_HOMEPAGE,
+                "image": datasets.Value("string"),
+                "score": datasets.Value("int32"),
+            })
         )
 
     def _split_generators(self, dl_manager):
-        name = self.config.name
-        base_dir = os.path.join(dl_manager.manual_dir or dl_manager.download_and_extract(""), name)
+        subset, split = self.config.name.split("_", 1)
+        subset_dir = os.path.join(dl_manager.manual_dir or dl_manager.download_and_extract(""), subset)
+
+        if split == "query":
+            file_path = os.path.join(subset_dir, "queries.jsonl")
+        elif split == "corpus":
+            file_path = os.path.join(subset_dir, "corpus.jsonl")
+        elif split == "mixed_corpus":
+            file_path = os.path.join(subset_dir, "mixed_corpus.jsonl")
+        elif split == "qrel":
+            file_path = os.path.join(subset_dir, "qrels", "qrels.tsv")
+        else:
+            raise ValueError(f"Unknown split: {split}")
 
         return [
-            datasets.SplitGenerator(name="query", gen_kwargs={"split": "query", "filepath": os.path.join(base_dir, "queries.jsonl")}),
-            datasets.SplitGenerator(name="corpus", gen_kwargs={"split": "corpus", "filepath": os.path.join(base_dir, "corpus.jsonl")}),
-            datasets.SplitGenerator(name="mixed_corpus", gen_kwargs={"split": "mixed_corpus", "filepath": os.path.join(base_dir, "mixed_corpus.jsonl")}),
-            datasets.SplitGenerator(name="qrel", gen_kwargs={"split": "qrel", "filepath": os.path.join(base_dir, "qrels", "qrels.tsv")}),
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"path": file_path, "split": split},
+            )
         ]
 
-    def _generate_examples(self, filepath, split):
-        if split == "query":
-            with open(filepath, encoding="utf-8") as f:
+    def _generate_examples(self, path, split):
+        if split in {"query", "corpus", "mixed_corpus"}:
+            with open(path, encoding="utf-8") as f:
                 for idx, line in enumerate(f):
                     item = json.loads(line)
                     yield idx, {
-                        "query_id": item["query_id"],
-                        "text": item["text"],
-                        "image": item.get("image", ""),
-                    }
-
-        elif split in {"corpus", "mixed_corpus"}:
-            with open(filepath, encoding="utf-8") as f:
-                for idx, line in enumerate(f):
-                    item = json.loads(line)
-                    yield idx, {
-                        "corpus_id": item["corpus_id"],
-                        "text": item["text"],
+                        "query_id": item.get("query_id", "") if split == "query" else "",
+                        "corpus_id": item.get("corpus_id", "") if split in {"corpus", "mixed_corpus"} else "",
+                        "text": item.get("text", ""),
                         "image": item.get("image", ""),
+                        "score": 0,
                     }
 
         elif split == "qrel":
-            with open(filepath, encoding="utf-8") as f:
+            with open(path, encoding="utf-8") as f:
                 reader = csv.DictReader(f, delimiter="\t")
                 for idx, row in enumerate(reader):
                     yield idx, {
                         "query_id": row["query_id"],
                         "corpus_id": row["corpus_id"],
+                        "text": "",
+                        "image": "",
                         "score": int(row["score"]),
                     }
-
-    def _info(self):
-        split = datasets.Split(self.config.data_files["split"]) if hasattr(self.config, "data_files") else None
-        if split == "query":
-            features = datasets.Features({
-                "query_id": datasets.Value("string"),
-                "text": datasets.Value("string"),
-                "image": datasets.Value("string"),
-            })
-        elif split in {"corpus", "mixed_corpus"}:
-            features = datasets.Features({
-                "corpus_id": datasets.Value("string"),
-                "text": datasets.Value("string"),
-                "image": datasets.Value("string"),
-            })
-        elif split == "qrel":
-            features = datasets.Features({
-                "query_id": datasets.Value("string"),
-                "corpus_id": datasets.Value("string"),
-                "score": datasets.Value("int32"),
-            })
-        else:
-            features = datasets.Features({
-                "id": datasets.Value("string"),
-                "text": datasets.Value("string"),
-            })
-
-        return datasets.DatasetInfo(
-            description=_DESCRIPTION,
-            features=features,
-            homepage=_HOMEPAGE,
-        )