v0.1.0
First release. Implements RFC-0001 at spec: 1.
Added
- The source definition model, with the JSON Schema generated from it and stamped with the version that produced it.
extendsresolution: scalars replace, mappings merge,fallbackprepends, every other list replaces.- The transform registry, typed, with a pipe whose types do not connect refused at validation time.
- The extractor engine:
css,json,regex,headerandconst. - Requests and pagination: three termination conditions,
fromalternatives, and form harvesting. - The interpreter, producing a novel, a table of contents with volumes, and bodies joined across pages.
- Hooks, bound by point name and refused if they import another host's implementation.
sourcelib try, reporting what every field produced and naming the spec field that failed.sourcelib explain, a structural digest of a page.sourcelib recordand offline replay.
Notes
- YAML is read as 1.2, not PyYAML's 1.1.
- IDNA2008 is required, not the standard library's codec.
- Fetching is an extra. A base install validates, resolves and transforms with no HTTP stack.