Crate orbok_extract 

orbok-extract

Text extraction (RFC-005): pluggable extractors turn boundary- validated source files into normalized, location-tagged segments. Extraction output is derived data — cacheable, rebuildable, never authoritative.

RFC-044 hardening adds: resource limits (ExtractLimits), structured warnings (ExtractWarning), panic isolation (extract_safely), explicit location semantics (LocationKind), and removal of the orbok-db production dependency (chunker now produces ExtractedChunk; the pipeline layer maps to ChunkSpec).

Re-exports

pub use chunker::chunk;

pub use plugin::PluginManifest;

pub use plugin::PluginRegistry;

pub use registry::ExtractorRegistry;

pub use types::DocumentExtractor;

pub use types::ExtractContext;

pub use types::ExtractLimits;

pub use types::ExtractOutput;

pub use types::ExtractWarning;

pub use types::ExtractedChunk;

pub use types::ExtractedSegment;

pub use types::LocationKind;

pub use types::LocationQuality;

pub use types::SegmentKind;

Modules

chunker

Adaptive chunker (RFC-006 §7–§9; RFC-044 §14 boundary cleanup).

docx

DOCX text extractor (Microsoft Word 2007+; RFC-044 §16.4 resource limits).

html

HTML text extractor (RFC-005 §5; RFC-044 §16.3 resource limits).

normalize

Text normalization, version norm-v1 (RFC-005 §9).

pdf

PDF text extraction via lopdf (RFC-022 §6; RFC-044 §16.5 hardening).

plugin

Plugin extractor interface (RFC-028 §7).

registry

Extractor registry (RFC-005 §6; RFC-044 §11 panic isolation).

types

Extraction types (RFC-005 §6–§8; RFC-044 hardening).