mixbench.py

import os
import json
import csv
import datasets

_DESCRIPTION = """
MixBench is a benchmark for mixed-modality retrieval across text, image, and image+text corpora.
"""

_HOMEPAGE = "https://huggingface.co/datasets/andy0207/mixbench"

class MixBench(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIGS = [
        datasets.BuilderConfig(name=name, version=datasets.Version("1.0.0"), description=f"MixBench subset: {name}")
        for name in ["MSCOCO", "Google_WIT", "VisualNews", "OVEN"]
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            homepage=_HOMEPAGE,
            features=datasets.Features({
                "query_id": datasets.Value("string"),
                "corpus_id": datasets.Value("string"), 
                "text": datasets.Value("string"),
                "image": datasets.Value("string"),
                "score": datasets.Value("int32"),
            })
        )

    def _split_generators(self, dl_manager):
        # 确保这个方法存在且正确实现
        subset_dir = os.path.join(dl_manager.manual_dir or dl_manager._base_path, self.config.name)
        
        return [
            datasets.SplitGenerator(
                name="query",
                gen_kwargs={"path": os.path.join(subset_dir, "queries.jsonl"), "split": "query"},
            ),
            datasets.SplitGenerator(
                name="corpus", 
                gen_kwargs={"path": os.path.join(subset_dir, "corpus.jsonl"), "split": "corpus"},
            ),
            datasets.SplitGenerator(
                name="mixed_corpus",
                gen_kwargs={"path": os.path.join(subset_dir, "mixed_corpus.jsonl"), "split": "mixed_corpus"},
            ),
            datasets.SplitGenerator(
                name="qrel",
                gen_kwargs={"path": os.path.join(subset_dir, "qrels", "qrels.tsv"), "split": "qrel"},
            ),
        ]

    def _generate_examples(self, path, split):
        if split == "query":
            with open(path, encoding="utf-8") as f:
                for idx, line in enumerate(f):
                    item = json.loads(line)
                    yield idx, {
                        "query_id": item.get("query_id", ""),
                        "corpus_id": "",
                        "text": item.get("text", ""),
                        "image": item.get("image", ""),
                        "score": 0,
                    }
        elif split == "corpus" or split == "mixed_corpus":
            with open(path, encoding="utf-8") as f:
                for idx, line in enumerate(f):
                    item = json.loads(line)
                    yield idx, {
                        "query_id": "",
                        "corpus_id": item.get("corpus_id", ""),
                        "text": item.get("text", ""),
                        "image": item.get("image", ""),
                        "score": 0,
                    }
        elif split == "qrel":
            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"])
                    }