Module

`parsing.ultra_fast`

Ultra-fast parser for simple documents.

Handles documents with only PARAGRAPH_LINE and BLANK_LINE tokens. Bypasses all block-level decision logic for maximum speed.

This covers 47.5% of CommonMark test cases and likely 60-80% of real-world documents (prose, comments, simple READMEs).

Functions

parse_ultra_simple 2 tuple[Block, ...] ▼

Parse document with only paragraphs and blank lines. Ultra-fast path: No block…

def parse_ultra_simple(tokens: list[Token], parse_inline_fn: Callable[[str, SourceLocation], tuple[Inline, ...]]) -> tuple[Block, ...]

Parse document with only paragraphs and blank lines.

Ultra-fast path: No block-level decisions, no container tracking, no indentation analysis. Just split on blank lines and parse inline.

Name	Type	Description
`tokens`	`list[Token]`	Token list (must contain only PARAGRAPH_LINE, BLANK_LINE, EOF)
`parse_inline_fn`	`Callable[[str, SourceLocation], tuple[Inline, ...]]`	Function to parse inline content

tuple[Block, ...]

can_use_ultra_fast 2 bool ▼

Check if document qualifies for ultra-fast path. O(n) check, but very cheap pe…

def can_use_ultra_fast(tokens: list[Token], *, tables_enabled: bool = False) -> bool

Check if document qualifies for ultra-fast path.

O(n) check, but very cheap per-token (just type comparison).

Name	Type	Description
`tokens`	`list[Token]`	Token list from lexer
`tables_enabled`	`bool`	Whether GFM table parsing is enabled. When True, any paragraph line containing a pipe is treated as a potential table and disqualifies the document (GFM tables need not have outer pipes, so the`startswith("\|")`heuristic alone is insufficient). Default:`False`

bool