lm-evaluation/lm_eval/tasks/indicxnli/utils.py

import argparse

import yaml


# Different languages that are part of xnli.
# These correspond to dataset names (Subsets) on HuggingFace.
# A yaml file is generated by this script for each language.

LANGUAGES = {
    "as": {  # Assamese
        "QUESTION_WORD": "শুধু",
        "ENTAILMENT_LABEL": "হাঁ",
        "NEUTRAL_LABEL": "তাই",
        "CONTRADICTION_LABEL": "না",
    },
    "bn": {  # Bengali
        "QUESTION_WORD": "সঠিক",
        "ENTAILMENT_LABEL": "হ্যাঁ",
        "NEUTRAL_LABEL": "তাহলে",
        "CONTRADICTION_LABEL": "না",
    },
    "gu": {  # Gujarati
        "QUESTION_WORD": "સાચું",
        "ENTAILMENT_LABEL": "હા",
        "NEUTRAL_LABEL": "તેનાથી",
        "CONTRADICTION_LABEL": "ના",
    },
    "hi": {  # Hindi
        "QUESTION_WORD": "सही",
        "ENTAILMENT_LABEL": "हाँ",
        "NEUTRAL_LABEL": "इसलिए",
        "CONTRADICTION_LABEL": "नहीं",
    },
    "kn": {  # Kannada
        "QUESTION_WORD": "ಸರಿ",
        "ENTAILMENT_LABEL": "ಹೌದು",
        "NEUTRAL_LABEL": "ಹಾಗಾಗಿ",
        "CONTRADICTION_LABEL": "ಇಲ್ಲ",
    },
    "ml": {  # Malayalam
        "QUESTION_WORD": "ശരി",
        "ENTAILMENT_LABEL": "അതെ",
        "NEUTRAL_LABEL": "എന്നാൽ",
        "CONTRADICTION_LABEL": "ഇല്ല",
    },
    "mr": {  # Marathi
        "QUESTION_WORD": "बरोबर",
        "ENTAILMENT_LABEL": "होय",
        "NEUTRAL_LABEL": "त्यानंतर",
        "CONTRADICTION_LABEL": "नाही",
    },
    "or": {  # Odia
        "QUESTION_WORD": "ସଠିକ",
        "ENTAILMENT_LABEL": "ହାଁ",
        "NEUTRAL_LABEL": "ଏହା",
        "CONTRADICTION_LABEL": "ନୁହଁ",
    },
    "pa": {  # Punjabi
        "QUESTION_WORD": "ਸਹੀ",
        "ENTAILMENT_LABEL": "ਹਾਂ",
        "NEUTRAL_LABEL": "ਇਸ ਲਈ",
        "CONTRADICTION_LABEL": "ਨਹੀਂ",
    },
    "ta": {  # Tamil
        "QUESTION_WORD": "சரி",
        "ENTAILMENT_LABEL": "ஆம்",
        "NEUTRAL_LABEL": "எனவே",
        "CONTRADICTION_LABEL": "இல்லை",
    },
    "te": {  # Telugu
        "QUESTION_WORD": "సరైన",
        "ENTAILMENT_LABEL": "అవును",
        "NEUTRAL_LABEL": "కాబట్టి",
        "CONTRADICTION_LABEL": "కాదు",
    },
}



def gen_lang_yamls(output_dir: str, overwrite: bool) -> None:
    """
    Generate a yaml file for each language.

    :param output_dir: The directory to output the files to.
    :param overwrite: Whether to overwrite files if they already exist.
    """
    err = []
    for lang in LANGUAGES.keys():
        file_name = f"indicxnli_{lang}.yaml"
        try:
            QUESTION_WORD = LANGUAGES[lang]["QUESTION_WORD"]
            ENTAILMENT_LABEL = LANGUAGES[lang]["ENTAILMENT_LABEL"]
            NEUTRAL_LABEL = LANGUAGES[lang]["NEUTRAL_LABEL"]
            CONTRADICTION_LABEL = LANGUAGES[lang]["CONTRADICTION_LABEL"]
            with open(
                f"{output_dir}/{file_name}", "w" if overwrite else "x", encoding="utf8"
            ) as f:
                f.write("# Generated by utils.py\n")
                yaml.dump(
                    {
                        "include": "indicxnli_common_yaml",
                        "dataset_name": lang,
                        "task": f"indicxnli_{lang}",
                        "doc_to_text": "",
                        "doc_to_choice": f"{{{{["
                        f"""premise+\", {QUESTION_WORD}? {ENTAILMENT_LABEL}, \"+hypothesis,"""
                        f"""premise+\", {QUESTION_WORD}? {NEUTRAL_LABEL}, \"+hypothesis,"""
                        f"""premise+\", {QUESTION_WORD}? {CONTRADICTION_LABEL}, \"+hypothesis"""
                        f"]}}}}",
                    },
                    f,
                    allow_unicode=True,
                )
        except FileExistsError:
            err.append(file_name)

    if len(err) > 0:
        raise FileExistsError(
            "Files were not created because they already exist (use --overwrite flag):"
            f" {', '.join(err)}"
        )


def main() -> None:

    gen_lang_yamls(output_dir=".", overwrite=True)


if __name__ == "__main__":
    main()