ocwcourses.py

"""OCWCourses"""


import json
import os
import re
import zipfile

import datasets

logger = datasets.logging.get_logger(__name__)

_CITATION = """\
@misc{lewkowycz2022solving,
      title={Solving Quantitative Reasoning Problems with Language Models}, 
      author={Aitor Lewkowycz and Anders Andreassen and David Dohan and Ethan Dyer and Henryk Michalewski and Vinay Ramasesh and Ambrose Slone and Cem Anil and Imanol Schlag and Theo Gutman-Solo and Yuhuai Wu and Behnam Neyshabur and Guy Gur-Ari and Vedant Misra},
      year={2022},
      eprint={2206.14858},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
"""

_DESCRIPTION = """\
OCWCourses
"""
_URL = "https://openreview.net/forum?id=IFXTZERXdM7"
_URLS = {
    "test": "https://openreview.net/attachment?id=IFXTZERXdM7&name=supplementary_material",
}

def _read_from_zip(zip_path, text_filename='minerva_supplement/ocw_courses_problems.text'):
    with zipfile.ZipFile(zip_path, 'r') as myzip:
        with myzip.open(text_filename) as myfile:
            return myfile.read().decode('utf-8')


def _parse_latex(text):
    problems = re.findall(r'\\textbf{Problem:}(.*?)\\textbf{Solution:}', text, re.DOTALL)
    solutions = re.findall(r'\\textbf{Solution:}(.*?)(\\textbf{Problem:}|$)', text, re.DOTALL)
    
    parsed_list = []
    for problem, solution in zip(problems, solutions):
        parsed_list.append({
            'problem': problem.strip(),
            'solution': solution[0].strip()
        })
        
    return parsed_list

def _last_boxed_only_string(string):

    idx = string.rfind("\\boxed")
    if "\\boxed " in string:
        return "\\boxed " + string.split("\\boxed ")[-1].split("$")[0]
    if idx < 0:
        idx = string.rfind("\\fbox")
        if idx < 0:
            return None

    i = idx
    right_brace_idx = None
    num_left_braces_open = 0
    while i < len(string):
        if string[i] == "{":
            num_left_braces_open += 1
        if string[i] == "}":
            num_left_braces_open -= 1
            if num_left_braces_open == 0:
                right_brace_idx = i
                break
        i += 1

    if right_brace_idx is None:
        retval = None
    else:
        retval = string[idx : right_brace_idx + 1]

    return retval

def _remove_boxed(s):
    if "\\boxed " in s:
        left = "\\boxed "
        assert s[: len(left)] == left
        return s[len(left) :]

    left = "\\boxed{"

    assert s[: len(left)] == left
    assert s[-1] == "}"

    return s[len(left) : -1]


class OCWConfig(datasets.BuilderConfig):
    """BuilderConfig for OCW."""

    def __init__(self, **kwargs):
        """BuilderConfig for OCW.

        Args:
          **kwargs: keyword arguments forwarded to super.
        """
        super(OCWConfig, self).__init__(**kwargs)


class OCWCourses(datasets.GeneratorBasedBuilder):
    """OCWCourses"""

    BUILDER_CONFIGS = [
        OCWConfig(
            name="ocwcourses",
            version=datasets.Version("1.0.0", ""),
            description="OCWCourses",
        ),
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "problem": datasets.Value("string"),
                    "solution": datasets.Value("string"),
                    "answer": datasets.Value("string"),
                }
            ),
            # No default supervised_keys (as we have to pass both question
            # and context as input).
            supervised_keys=None,
            homepage="https://openreview.net/forum?id=IFXTZERXdM7",
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        downloaded_files = dl_manager.download_and_extract(_URLS)

        return [
            datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),
        ]

    def _generate_examples(self, filepath):
        """This function returns the examples in the raw (text) form."""
        logger.info("generating examples from = %s", filepath)

        with open(os.path.join(filepath, "minerva_supplement/ocw_courses_problems.tex")) as f:
            text = f.read()

        parsed_problems_solutions = _parse_latex(text)

        rows = [{**x, "answer": _remove_boxed(_last_boxed_only_string(x["solution"]))} for x in parsed_problems_solutions]

        for key, row in enumerate(rows):
            yield key, row