Struct ExtractorRegistry 

pub struct ExtractorRegistry { /* private fields */ }

Registry of the available extractors. Markdown takes precedence over plain text for .md; everything claims by extension.

Implementations

impl ExtractorRegistry

pub fn new_with(extractors: Vec<Box<dyn DocumentExtractor>>) -> Self

Build a registry with a custom set of extractors (useful in tests).

pub fn select(&self, extension: &str) -> Option<&dyn DocumentExtractor>

The extractor claiming extension, if any.

pub fn extract_with_context( &self, path: &ValidatedPath, context: &ExtractContext, ) -> OrbokResult<ExtractOutput>

Extract using resource limits. Unknown types are a typed UnsupportedType failure — never a panic, never a silent skip.

Prefer this over [extract] for production code paths.

pub fn extract_safely( &self, path: &ValidatedPath, context: &ExtractContext, ) -> OrbokResult<ExtractOutput>

Extract with panic isolation (RFC-044 §11).

Wraps the extractor call in catch_unwind. A parser panic is caught and returned as ErrorCategory::ParserPanic instead of crashing the worker thread.

The user-facing layer must translate ParserPanic to a plain message like “This file could not be prepared.”

pub fn extract(&self, path: &ValidatedPath) -> OrbokResult<ExtractOutput>

Legacy entry point (no limits, no panic isolation).

Kept for compatibility during the migration period. New code should call [extract_safely] instead.

Trait Implementations

impl Default for ExtractorRegistry

fn default() -> Self

Returns the “default value” for a type.