ReproWeave v0.1.0 — paper-to-replication evidence maps
ReproWeave v0.1.0 is the first public release of a local-first, zero-runtime-dependency toolkit for turning engineering and AI papers into auditable evidence maps and dependency-aware replication plans.
Highlights:
- Seven file-native artifact types with published JSON Schemas
- Offline BibTeX and CSL JSON import
- Eight-dimension evidence-anchored reconstructability rubric
- Typed evidence graph, gap backlog, and replication DAG planning
- Cross-reference audits and SHA-256 evidence seals
- Self-contained interactive HTML, CSV, Markdown, and JSON reports
- Fully synthetic 55-artifact EE/AI demonstration
- 75 tests across Windows and Ubuntu on Python 3.11 and 3.13
Interpretation boundary: scores describe documented reconstructability, not scientific quality, correctness, novelty, or integrity. The bundled demonstration is entirely fictional.
Wheel SHA-256: 465e7ed408e2ad878c04528464dab0d79016b1b0a27ed0ad15e135b1fe5ca078
Source distribution SHA-256: a425d8f020aa50328d701cc9d06a1ba0eb6b04cb4b7a276a1153ef51e19d2489