markitect-main/markitect/frontmatter.py

"""
Front matter parsing functionality for MarkiTect.

This module provides YAML front matter parsing for markdown files,
separating metadata from content.
"""

import re
import yaml
from typing import Dict, Tuple, Any


class FrontMatterParser:
    """Parser for YAML front matter in markdown files."""

    def __init__(self):
        """Initialize the front matter parser."""
        pass

    def parse(self, content: str) -> Tuple[Dict[str, Any], str]:
        """
        Parse front matter from markdown content.

        Args:
            content: Raw markdown content that may include YAML front matter

        Returns:
            Tuple of (front_matter_dict, markdown_content)
            - front_matter_dict: Parsed YAML as dictionary, empty dict if none
            - markdown_content: Markdown content with front matter removed
        """
        if not content.strip():
            return {}, content

        # Check if content starts with front matter delimiter
        if not content.strip().startswith('---'):
            return {}, content

        # Pattern to match YAML front matter
        # Must start with --- at beginning of string, end with --- on its own line
        pattern = r'^---\s*\n(.*?)\n---\s*\n(.*)$'
        match = re.match(pattern, content, re.DOTALL)

        if not match:
            # No valid front matter found
            return {}, content

        yaml_content = match.group(1)
        markdown_content = match.group(2)

        # Parse YAML content
        try:
            front_matter = yaml.safe_load(yaml_content)
            if front_matter is None:
                front_matter = {}
        except yaml.YAMLError:
            # Invalid YAML - return empty dict and preserve content
            front_matter = {}

        return front_matter, markdown_content