question_parser.py

import re


class QuestionType:
    PERCENT = "percent"
    RATIO = "ratio"
    ALGEBRA = "algebra"
    STATISTICS = "statistics"
    PROBABILITY = "probability"
    GEOMETRY = "geometry"
    ARITHMETIC = "arithmetic"
    UNKNOWN = "unknown"


def detect_question_type(question_text: str) -> str:
    q = question_text.lower()

    if "%" in q or "percent" in q or "increase" in q or "decrease" in q:
        return QuestionType.PERCENT

    if "ratio" in q or ":" in q or "proportion" in q:
        return QuestionType.RATIO

    if re.search(r"\bmean\b|\bmedian\b|\bmode\b|\bstandard deviation\b", q):
        return QuestionType.STATISTICS

    if "probability" in q or "chance" in q or "random" in q:
        return QuestionType.PROBABILITY

    if "triangle" in q or "circle" in q or "radius" in q or "area" in q:
        return QuestionType.GEOMETRY

    if re.search(r"[a-z]\s*=", q):
        return QuestionType.ALGEBRA

    if re.search(r"\d", q):
        return QuestionType.ARITHMETIC

    return QuestionType.UNKNOWN