testbed/embeddings-benchmark__mteb/scripts/data/bucc/create_data.py

from __future__ import annotations

import gzip
import json
import re

from huggingface_hub import upload_file

repo_name = "bucc-bitext-mining"
# create_repo(repo_name, organization="mteb", repo_type="dataset")

with open("bucc-data/zh-en/zh-en.training.zh", "r") as f:
    sentence1 = f.readlines()

with open("bucc-data/zh-en/zh-en.training.en", "r") as f:
    sentence2 = f.readlines()

with open("bucc-data/zh-en/zh-en.training.gold", "r") as f:
    gold = f.readlines()


def process_sentence(x):
    x = re.sub("\n", "", x).strip()
    return x.split("\t")[1]


def process_gold(x):
    id1, id2 = x.strip().split("\t")
    id1 = id1.split("-")[1]
    id2 = id2.split("-")[1]
    return int(id1), int(id2)


sentence1 = list(map(process_sentence, sentence1))
sentence2 = list(map(process_sentence, sentence2))
gold = list(map(process_gold, gold))

data = {"sentence1": sentence1, "sentence2": sentence2, "gold": gold}

with gzip.open("test.json.gz", "wt", encoding="UTF-8") as zipfile:
    json.dump(data, zipfile)

upload_file(
    path_or_fileobj="test.json.gz",
    path_in_repo="zh-en/test.json.gz",
    repo_id=f"mteb/{repo_name}",
    repo_type="dataset",
)