stats.py

import pandas as pd
from datasets import load_dataset

for _type in ['nell', 'nell_filter']:
    data = load_dataset("relbert/fewshot_link_prediction", _type, split='test')
    df = data.to_pandas()

    print(f"\nEntity Types ({_type})")
    tail = df.groupby("tail_type")['relation'].count().to_dict()
    head = df.groupby("head_type")['relation'].count().to_dict()
    k = set(list(tail.keys()) + list(head.keys()))
    df_types = pd.DataFrame([{"entity_type": _k, "tail": tail[_k] if _k in tail else 0, "head": head[_k] if _k in head else 0} for _k in k])
    print(df_types.to_markdown(index=False))

    print(f"\nRelation Types ({_type})")
    print(df.groupby("relation")['relation'].count().to_markdown())

    print(f"\nVocab Size ({_type})")
    with open(f"data/{_type}.vocab.txt") as f:
        length = len([i for i in f.read().split('\n') if len(i) > 0])
    print(length)