Documentation
¶
Overview ¶
Package origin binds each reader's bounded input snapshot to its original revision. Parsers, normalizers and splitters never hash rendered text.
Index ¶
- func Parts(ctx context.Context, p *document.Provenance, docs []*document.Document) ([]*document.Document, error)
- func Read(ctx context.Context, r io.Reader, budget etl.SourceBudget, name string) ([]byte, *document.Provenance, error)
- func Span(p *document.Provenance, start, end int64) *document.Provenance
- type Stream
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func Parts ¶
func Parts(ctx context.Context, p *document.Provenance, docs []*document.Document) ([]*document.Document, error)
Parts numbers only documents that were actually emitted. Parser output is generally not an exact byte slice, so Split discards any inherited byte span.
func Span ¶
func Span(p *document.Provenance, start, end int64) *document.Provenance
Span identifies bytes only when the parser can prove both original offsets. An anonymous single document need not acquire a fabricated source identity.
Types ¶
Source Files
Click to show internal directories.
Click to hide internal directories.