Skip to content

README.md

Latest commit

 

History

History
90 lines (68 loc) · 3.91 KB

README.md

File metadata and controls

90 lines (68 loc) · 3.91 KB
Raw

sc-compose — Repowise Code Health Analysis

This directory contains generated code health artifacts produced by repowise, a domain-aware codebase intelligence tool. The analysis uses a repo-specific config (.sc/repowise.yaml) to avoid false positives in prototype code, FFI bindings, and utility scripts.

The committed health, badge, and module reports are generated snapshots, not live measurements of the current checkout. health.md records the analyzed commit and date; regenerate the pipeline below after significant source changes before treating the metrics as current.

Analysis Pipeline

The full pipeline runs in three phases:

Phase 1 — Index (prerequisite)

repowise init --index-only -y

Builds the knowledge graph, AST index, and git history. Must be re-run after significant code changes before refreshing any derived data.

Phase 2 — Raw data export

Command Output What It Produces
repowise health --format json --module crates/sc-compose .sc/repowise/data/repowise-health-compose.json Per-file scores, KPIs, biomarker findings
repowise health --format json --refactoring-targets .sc/repowise/data/repowise-refactoring-targets.json Ranked refactoring targets with plans and ROI
repowise dead-code --format json .sc/repowise/data/repowise-dead-code.json Unreachable files, unused exports, zombie packages
repowise health --badge docs/repowise/badge.md Markdown health badge (for README embedding)
repowise risk --format json docs/repowise/risk.json HEAD vs recent commit risk delta

Phase 3 — Compile reports

Command Output What It Produces
python3 .sc/repowise/generate-report.py docs/repowise/health.md Comprehensive report: scores, worst files, biomarkers, refactoring targets, dead code, recommendations
python3 .sc/repowise/generate-architecture.py site/repowise/architecture.html Crate dependency diagram (reads .repowise/knowledge-graph.json and Cargo.toml, resolves version from git)
repowise init (full, with LLM) then repowise export --format html -o site/repowise/wiki/ site/repowise/wiki/ Per-file LLM-generated documentation (requires API key configured in repowise; optional)

Output Artifacts

File Phase Description
health.md 3 Comprehensive report with interpretation and recommendations
risk.json 2 Risk assessment at HEAD
badge.md 2 Markdown health badge
../site/repowise/architecture.html 3 Crate dependency diagram (web-facing)
../site/repowise/wiki/ 3 (LLM) Per-file documentation (69 pages; optional)

Regeneration (full pipeline)

cd <worktree-root>

# Phase 1: Re-index (after code changes)
repowise init --index-only -y

# Phase 2: Export raw data
repowise health --format json --module crates/sc-compose \
  > .sc/repowise/data/repowise-health-compose.json 2>&1
repowise health --format json --refactoring-targets \
  > .sc/repowise/data/repowise-refactoring-targets.json 2>&1
repowise dead-code --format json \
  > .sc/repowise/data/repowise-dead-code.json 2>&1
repowise health --badge > docs/repowise/badge.md 2>&1
repowise risk --format json > docs/repowise/risk.json 2>&1

# Phase 3: Compile reports
python3 .sc/repowise/generate-report.py
python3 .sc/repowise/generate-architecture.py

# Phase 3 (optional, requires LLM): Regenerate wiki
repowise init                    # full init with LLM page generation
repowise export --format html -o site/repowise/wiki/

The generate-report.py script resolves version, commit, and date from git at run time — no hardcoded metadata.

Config

See .sc/repowise.yaml for module scope, annotated paths, and diagram specs. The annotated section marks directories that are intentionally excluded from the reachability analysis (Python FFI bindings, prototypes, utility scripts).