question_api/utils.py

import os
import re
import json
import logging

logger = logging.getLogger(__name__)


def parse_restructured_text(text_block):
    """
    Parse Gemini's restructured text into structured MCQs and Theory questions.
    
    Args:
        text_block: Raw text output from Gemini
    
    Returns:
        tuple: (mcqs_list, theories_list)
    """
    mcqs = []
    theories = []

    # Split into chunks by double newlines
    chunks = [c.strip() for c in text_block.split("\n\n") if c.strip()]

    for chunk in chunks:
        # Try to parse as MCQ
        if "MCQ" in chunk or "Stem:" in chunk:
            mcq = parse_mcq_chunk(chunk)
            if mcq:
                mcqs.append(mcq)
        
        # Try to parse as Theory question
        elif "Theory" in chunk or ("Question:" in chunk and "Answer:" in chunk):
            theory = parse_theory_chunk(chunk)
            if theory:
                theories.append(theory)

    logger.info(f"Parsed {len(mcqs)} MCQs and {len(theories)} theory questions")
    return mcqs, theories


def parse_mcq_chunk(chunk):
    """
    Parse a single MCQ chunk into structured format.
    
    Expected format:
        MCQ [number]
        Stem: [question]
        Key: [correct answer]
        Distractors:
        - [distractor 1]
        - [distractor 2]
        - [distractor 3]
    
    Returns:
        dict or None
    """
    try:
        # Extract stem
        stem_match = re.search(r"Stem:\s*(.+?)(?=\nKey:|\nDistractors:|$)", chunk, re.DOTALL)
        if not stem_match:
            logger.warning(f"No stem found in MCQ chunk: {chunk[:100]}")
            return None
        stem = stem_match.group(1).strip()

        # Extract key (correct answer)
        key_match = re.search(r"Key:\s*(.+?)(?=\nDistractors:|\n-|$)", chunk, re.DOTALL)
        if not key_match:
            logger.warning(f"No key found in MCQ chunk: {chunk[:100]}")
            return None
        key = key_match.group(1).strip()

        # Extract distractors
        distractor_pattern = r"-\s*(.+?)(?=\n-|\n\n|$)"
        distractors = re.findall(distractor_pattern, chunk, re.DOTALL)
        distractors = [d.strip() for d in distractors if d.strip()]

        if len(distractors) < 3:
            logger.warning(f"Only {len(distractors)} distractors found, need 3")
            # Pad with generic distractors if needed
            while len(distractors) < 3:
                distractors.append("None of the above")

        return {
            "stem": stem,
            "key": key,
            "distractors": distractors[:3]  # Ensure exactly 3
        }

    except Exception as e:
        logger.error(f"Error parsing MCQ chunk: {e}")
        return None


def parse_theory_chunk(chunk):
    """
    Parse a single Theory question chunk into structured format.
    
    Expected format:
        Theory [number]
        Question: [question text]
        Answer: [answer text]
    
    Returns:
        dict or None
    """
    try:
        # Extract question
        question_match = re.search(r"Question:\s*(.+?)(?=\nAnswer:|$)", chunk, re.DOTALL)
        if not question_match:
            logger.warning(f"No question found in theory chunk: {chunk[:100]}")
            return None
        question = question_match.group(1).strip()

        # Extract answer
        answer_match = re.search(r"Answer:\s*(.+)$", chunk, re.DOTALL)
        if not answer_match:
            logger.warning(f"No answer found in theory chunk: {chunk[:100]}")
            return None
        answer = answer_match.group(1).strip()

        return {
            "question": question,
            "answer": answer
        }

    except Exception as e:
        logger.error(f"Error parsing theory chunk: {e}")
        return None


def load_studykit_dataset():
    """
    Load the StudyKit dataset JSON file.
    
    Returns:
        list: List of study kit entries
    
    Raises:
        FileNotFoundError: If dataset file doesn't exist
    """
    # Determine path relative to this file
    here = os.path.dirname(os.path.abspath(__file__))
    
    # Try multiple possible paths
    possible_paths = [
        os.path.join(here, "..", "..", "dataset", "studykit_questions_dataset.json"),
        os.path.join(here, "..", "dataset", "studykit_questions_dataset.json"),
        os.path.join(here, "dataset", "studykit_questions_dataset.json"),
    ]
    
    for dataset_path in possible_paths:
        dataset_path = os.path.normpath(dataset_path)
        if os.path.exists(dataset_path):
            try:
                with open(dataset_path, "r", encoding="utf-8") as f:
                    data = json.load(f)
                logger.info(f"Loaded dataset from: {dataset_path}")
                return data
            except json.JSONDecodeError as e:
                logger.error(f"Invalid JSON in dataset: {e}")
                return []
    
    raise FileNotFoundError(
        f"Cannot find studykit_questions_dataset.json in any of: {possible_paths}"
    )


def fetch_internet_content(topic):
    """
    Placeholder for fetching content from the internet.
    You should implement this based on your requirements.
    
    Args:
        topic: Topic to search for
    
    Returns:
        str: Content text or empty string
    """
    logger.warning(f"fetch_internet_content called for topic: {topic}")
    logger.warning("This function is not implemented. Returning empty content.")
    # TODO: Implement web scraping or API calls to fetch content
    return ""