litedoc-py 0.1.0

Deterministic document parser for AI agents and LLM output

pyld

Deterministic document parser for AI agents and LLM output. Python bindings for the Rust litedoc-core library.

Install

pip install litedoc-py

Usage

import pyld

# Parse a document
doc = pyld.parse("""
# Hello World

This is a **paragraph** with *formatting*.

::list
- First item
- Second item
::
""")

# Iterate over blocks
for block in doc.blocks:
    match block:
        case pyld.Heading(level=level, content=content):
            print(f"H{level}: {content}")
        case pyld.Paragraph(content=content):
            print(f"Paragraph: {content}")
        case pyld.List(items=items):
            print(f"List with {len(items)} items")

# Parse with error recovery
result = pyld.parse_with_recovery("""
::list
- Unclosed list
""")

print(f"Blocks: {len(result.document.blocks)}")
print(f"Errors: {len(result.errors)}")
print(f"OK: {result.ok}")

API

Functions

• parse(input, profile=None) - Parse a string, raises ValueError on error
• parse_with_recovery(input, profile=None) - Parse with error recovery, always returns a result

Classes

• Parser(profile=None) - Reusable parser instance
• Document - Parsed document with blocks, metadata, profile
• ParseResult - Result with document, errors, ok

Profiles

• Profile.Litedoc - Full LiteDoc syntax (default)
• Profile.Md - CommonMark + GFM
• Profile.MdStrict - Strict CommonMark

Block Types

Heading, Paragraph, List, CodeBlock, Callout, Quote, Figure, Table, Footnotes, MathBlock, ThematicBreak, HtmlBlock, RawBlock

Inline Types

Text, Emphasis, Strong, Strikethrough, CodeSpan, Link, AutoLink, FootnoteRef, HardBreak, SoftBreak

License

Apache-2.0