kingchenc
1aa7df1c97
`napi create-npm-dir` failed in CI with both invocations we tried:
positional arg form (`-t <triple> .`) was rejected as extraneous,
and the no-arg form crashed with "path must be a string, received
undefined". The fix used by every napi-rs reference project: commit
the four platform package.json templates directly under
bindings/node/npm/<target>/ instead of generating them at publish time.
- bindings/node/npm/{linux-x64-gnu,darwin-x64,darwin-arm64,win32-x64-msvc}/
each contain a static package.json with the correct os / cpu / libc
filters so npm only installs the right binary per platform.
- Main package.json gains optionalDependencies referencing all four
platform packages by version, so `npm install wickra` pulls the
matching binary on each user's machine.
- Release workflow drops the broken `create-npm-dir` loop. The
`napi artifacts` step now just copies the .node files from the
build artefacts into the existing npm/ directories before publish.
Bump every version to 0.1.2; cargo / pypi / wickra-wasm jobs are
idempotent so they accept the 0.1.1 they already published while still
emitting the new 0.1.2.
Wickra — Python bindings
Streaming-first technical indicators powered by a Rust core.
pip install wickra
Quick start
import numpy as np
import wickra as ta
# Batch — TA-Lib-style usage
prices = np.linspace(100, 200, 1000)
rsi = ta.RSI(14).batch(prices) # NumPy array; NaN during warmup
# Streaming — feed ticks one at a time
rsi = ta.RSI(14)
for price in live_prices:
v = rsi.update(price) # O(1) per tick
if v is not None and v > 70:
...
What's included
25 streaming-first indicators across four families. Every one passes a
batch == streaming equivalence test and reference-value tests:
- Trend — SMA, EMA, WMA, DEMA, TEMA, HMA, KAMA
- Momentum — RSI (Wilder), MACD, Stochastic, CCI, ROC, WilliamsR, ADX, MFI, TRIX, AwesomeOscillator, Aroon
- Volatility — BollingerBands, ATR, Keltner, Donchian, PSAR
- Volume — OBV, VWAP
Why streaming-first matters
Classic TA libraries are batch-only: every live tick triggers a full recomputation over the entire history. Wickra updates indicator state in O(1) per tick. On a 5K-bar history the streaming RSI gap is ~17× over the nearest peer with a streaming API and 100×+ over batch-only libraries.
Full project
See https://github.com/kingchenc/wickra for benchmarks, the Rust core, Node.js and WebAssembly bindings, examples, and CI.
License
Licensed under the PolyForm Noncommercial License 1.0.0. Personal, research, educational, and non-profit use are all permitted. Commercial sale requires a separate license — contact via the GitHub repo.