995f119010
## Summary Adds a new **Chart Patterns** indicator family (counter 351 → 359, families 21 → 22), the first half of the A4 roadmap item (the harmonic patterns follow in a second PR). All eight detectors are built on a shared, non-repainting swing-pivot tracker — the internal, **uncounted** `indicators::pattern_swing` module (declared `pub(crate) mod`, re-exported nowhere). Each consumes candles and emits the uniform pattern sign convention already used by the candlestick family — `+1.0` bullish / `-1.0` bearish / `0.0` otherwise, never `None`. They are parameter-free, baking the swing threshold (5%) and level tolerance (3%) in as documented constants, mirroring how candlestick patterns bake in their geometric thresholds. ## Detectors | Indicator | Signal | |-----------|--------| | `DoubleTopBottom` | twin-peak / twin-trough reversal | | `TripleTopBottom` | three matching extremes (stronger reversal) | | `HeadAndShoulders` | central head + matching shoulders + flat neckline (and inverse) | | `Triangle` | ascending (+1) / descending (-1) / symmetrical | | `Wedge` | rising wedge (-1) / falling wedge (+1) | | `FlagPennant` | shallow consolidation against a pole → continuation | | `RectangleRange` | flat support/resistance mean-reversion | | `CupAndHandle` | rounded base + shallow handle (and inverse) | ## Touchpoints Core modules + `FAMILIES` group and assert, crate root re-exports, Python/Node/WASM bindings via the candle-pattern macros (Node `index.d.ts`/`index.js` regenerated), the candle fuzz target, Python reference + `CANDLE_SCALAR` registry tests and the Node candle-scalar factory, README catalogue counter + banner cache-buster + family table row + family-count word, `docs/README.md` counter, and the changelog. ## Verification - `cargo test -p wickra-core --lib` — 2915 passed - `cargo test -p wickra-core --doc` — 335 passed - `cargo clippy --workspace --all-targets --all-features -- -D warnings` — clean - Node `npm run build && npm test` — 436 passed - Python `maturin develop --release` + `pytest` — 732 passed Every detector branch is unit-tested; multi-condition predicates were flattened to single-line precomputed booleans to keep patch coverage at 100%.