Add decriptive stat almost to all datasets

[Samoed]

Checklist

• Run tests locally to make sure nothing is broken using make test.
• Run the formatter to format the code using make lint.

I ran almost all tasks, except for those with issues mentioned in #1465 and some large datasets that couldn’t fit on my laptop. I’ll try running those on Kaggle. Additionally, I fixed a few bugs related to dataset loading.

I don’t have statistics for the larger tasks because they exceed the RAM capacity of my laptop, and Kaggle's disk space is insufficient. I’ll need some assistance to run them.

Tasks without stats:

• FEVER
• BrightRetrieval
• HotpotQA
• MSMARCO
• MSMARCOv2
• MIRACLRetrieval
• MrTidyRetrieval
• TopiOCQA
• MultiLongDocRetrieval
• NeuCLIR2022Retrieval
• NeuCLIR2023Retrieval
• MultiEURLEXMultilabelClassification
• SwissJudgementClassification
• BibleNLPBitextMining
• FloresBitextMining
• MIRACLReranking
• MindSmallReranking
• VoyageMMarcoReranking
• WebLINXCandidatesReranking

Script to run these tasks

import mteb
from tqdm import tqdm

for task_name in tqdm(
    [
        "HotpotQA", 
        "MSMARCO",
        "MSMARCOv2",
        "MIRACLRetrieval",
        "MrTidyRetrieval",
        "TopiOCQA",
        "MultiLongDocRetrieval",
        "NeuCLIR2022Retrieval",
        "NeuCLIR2023Retrieval",
        "MultiEURLEXMultilabelClassification",
        "SwissJudgementClassification",
        "BibleNLPBitextMining",
        "FloresBitextMining",
        "MIRACLReranking",
        "MindSmallReranking",
        "VoyageMMarcoReranking",
        "WebLINXCandidatesReranking",
    ]
):
    task = mteb.get_task(task_name)
    stat = task.calculate_metadata_metrics()

[KennethEnevoldsen]

@imenelydiaker just so that you know this exists

[KennethEnevoldsen]

Feel free to create an issue on this and merge in the current stats

[Samoed]

I will downgrade datasets version and run tasks that were broken