Files
wickra/fuzz
kingchenc 54194a4ff8 feat: Family 05 Bands & Channels - 11 new price-envelope indicators (#43)
* feat(bands-channels): add Family 05 with 11 indicators

Eleven price-envelope overlays organised into a new "Bands & Channels"
family, exposed across all four bindings (Rust core, Python, Node, WASM)
plus fuzz/test/bench/docs coverage:

- MaEnvelope - SMA centerline with fixed-percent envelope (the oldest
  band overlay still in regular use).
- AccelerationBands (Price Headley) - momentum-biased bands that widen
  with the bar's relative range (H - L) / (H + L).
- StarcBands (Stoller Average Range Channel) - SMA(close) +/- k*ATR;
  Keltner's SMA-centerline sibling.
- AtrBands - close-anchored envelope of width k*ATR; the standard
  volatility-targeting stop/target band.
- HurstChannel - SMA centerline wrapped by the rolling high-low range
  (Brian Millard / Hurst-cycle channel).
- LinRegChannel - rolling OLS endpoint +/- k * population stddev of the
  residuals; dispersion about the trend rather than the mean.
- StandardErrorBands - regression line +/- k * OLS standard error
  (denominator n - 2) for prediction-interval bands.
- DoubleBollinger (Kathy Lien) - two concentric BB envelopes
  (typically +/- 1 sigma and +/- 2 sigma) for the zone-partition setup.
- TtmSqueeze (John Carter) - BB-inside-KC squeeze flag paired with a
  detrended-close linear-regression momentum reading.
- FractalChaosBands - Bill Williams 5-bar fractal high/low envelope.
- VwapStdDevBands - cumulative VWAP with volume-weighted population
  standard deviation bands.

Each indicator ships:
- Core impl with the full Indicator trait, classic() where applicable,
  and unit tests (rejects_zero_period / multiplier, accessors, flat
  market, monotonic ordering, batch == streaming, reset, plus
  algebraically verifiable reference values).
- Python PyO3 binding with multi-column NumPy batch (PyArray2).
- Node napi binding with #[napi(object)] struct + interleaved flat
  batch.
- WASM wasm-bindgen binding via Object/Reflect for update +
  Float64Array for batch.
- Fuzz coverage in fuzz_targets/indicator_update{,_candle}.rs.
- Python streaming-vs-batch parametric test + reference test.
- Node streaming-vs-interleaved-batch test + reference test.
- Criterion microbench under crates/wickra/benches/indicators.rs.

README family table, README indicator-count line, and CHANGELOG
Unreleased entry updated: indicator total rises from 71 to 82 across
nine families. Wiki pages are updated in a separate commit in the
wickra.wiki repo.

* test(acceleration-bands): cover sum_hl==0 zero-price guard

Exercises line 104 (`0.0` branch of the `sum_hl == 0.0` guard) which
was the last patch-coverage miss on the family-05 PR. `Candle::new`
accepts a fully-zero bar so the branch is reachable in principle —
add a degenerate-candle unit test to hit it.
2026-05-25 18:37:12 +02:00
..
2026-05-22 16:44:19 +02:00

Fuzzing Wickra

cargo-fuzz harnesses for the parsing and stateful entry points of Wickra. Fuzzing requires a nightly Rust toolchain.

Setup

cargo install cargo-fuzz
rustup toolchain install nightly

Targets

Target What it exercises
csv_reader CandleReader over arbitrary bytes — headers, cells, BOM, binary noise.
binance_envelope RawWsEnvelope deserialization from arbitrary strings.
indicator_update Every scalar-input indicator (SMA / EMA / WMA / RSI / DEMA / TEMA / HMA / ROC / TRIX / SMMA / TRIMA / ZLEMA / KAMA / T3 / MOM / CMO / TSI / PMO / StochRSI / DPO / PPO / Coppock / StdDev / UlcerIndex / HistoricalVolatility / LinearRegression / LinRegSlope / LinRegAngle / VHF / ZScore / MACD / Bollinger) streamed + batched over arbitrary f64 sequences (NaN, ±inf, jumps).
indicator_update_candle Every candle-input indicator (ATR, NATR, TrueRange, ChaikinVolatility, Keltner, Donchian, PSAR, SuperTrend, ChandelierExit, ChandeKrollStop, ATRTrailingStop, ADX, Aroon, AroonOscillator, Vortex, MassIndex, ChoppinessIndex, CCI, WilliamsR, AwesomeOscillator, AcceleratorOscillator, UltimateOscillator, BalanceOfPower, OBV, MFI, VWAP, RollingVWAP, VWMA, ADL, VPT, CMF, ChaikinOscillator, ForceIndex, EaseOfMovement, TypicalPrice, MedianPrice, WeightedClose, Stochastic) streamed + batched over fuzz-derived OHLCV candles.
tick_aggregator TickAggregator over arbitrary (price, volume, timestamp) triples.

Run

# From the repository root:
cargo +nightly fuzz run csv_reader
cargo +nightly fuzz run binance_envelope
cargo +nightly fuzz run indicator_update
cargo +nightly fuzz run indicator_update_candle
cargo +nightly fuzz run tick_aggregator

Each run continues until a crash is found or it is interrupted. A short time-boxed smoke run is useful in CI:

cargo +nightly fuzz run csv_reader -- -max_total_time=60

The expectation for every target is that it never panics: malformed or adversarial input must surface as an Err, never a crash.