kingchenc
3ebcb3f758
Adds a `throughput` benchmark to every target and closes two small test-coverage documentation/QA gaps. One PR, no merge of binding code beyond the additive benchmarks and one C test. ## 1. Per-binding throughput benchmarks (all 9 targets) Each benchmark feeds a deterministic synthetic OHLCV series through three indicators chosen by **FFI call-signature archetype** (not algorithm — the same Rust core runs underneath all bindings): - `SMA(20)` — 1-in → 1-out (baseline boundary cost) - `ATR(14)` — multi-in → 1-out (input marshalling) - `MACD(12,26,9)` — 1-in → multi-out (output marshalling) Streaming is timed for all three; batch for the single-output SMA and ATR (median of 3 runs, after a warmup pass). New: Python (PyO3), WASM, C (CMake), C# (Stopwatch), Go, Java (FFM), R, and the Rust core baseline (`examples/rust/.../throughput.rs`, **no FFI** — the ceiling the bindings are measured against and the value their batch paths converge towards). Node already had `throughput.js`. **Not a speed claim:** there is no comparable streaming TA library for C, C#, Go, Java, R or WASM to compare against, so these are raw per-binding throughput numbers documenting each language's FFI overhead — see BENCHMARKS.md §3. The "Wickra is fast" claim still lives in §1/§2 (Rust core + the Python/Rust cross-library runs). ## 2. README `## Testing`: C# and C bullets The section listed every layer except C# and C, even though both have suites. Adds the two missing bullets. ## 3. C archetype ctest `examples/c/archetypes.c` drives one indicator per FFI archetype through the real C boundary (scalar + batch==streaming, multi-output, bars, profile, array input) plus reset, invalid-parameter and NULL-safety — the C counterpart of the Go/R/Java archetype suites. Runs on three OSes via the existing CMake/ctest. ## Notes - Benchmarks are not CI-gated (manual-run scripts, like the existing `throughput.js`); no `ci.yml`/`release.yml` changes. - Docs: BENCHMARKS.md §3, a `## Benchmark` section in every binding README, a CHANGELOG entry. - Verified locally by running: Rust, Python, C, C#, Go, Java (real numbers); the C archetype ctest with `-Wall -Wextra -Wpedantic -Werror`. WASM and R are API-correct and syntax-checked but need their own toolchains to run.
Wickra — C / C++ examples
The Wickra C ABI is a single shared/static library plus a generated header
(bindings/c/include/wickra.h). Any
C-capable language links against the same artifact; these examples show the
plain-C path.
Build the library
From the workspace root:
cargo build -p wickra-c --release
This produces, in target/release/:
| Platform | Shared library | Link target |
|---|---|---|
| Linux | libwickra.so |
-lwickra |
| macOS | libwickra.dylib |
-lwickra |
| Windows (MSVC) | wickra.dll |
wickra.dll.lib (import lib) |
A static library (libwickra.a / wickra.lib) is emitted alongside.
Build and run the smoke example
With CMake (portable, used by CI)
cmake -S examples/c -B examples/c/build -DWICKRA_LIB_DIR="$PWD/target/release"
cmake --build examples/c/build
ctest --test-dir examples/c/build --output-on-failure
Directly with a compiler
# Linux / macOS
cc examples/c/smoke.c -I bindings/c/include -L target/release -lwickra -lm -o smoke
LD_LIBRARY_PATH=target/release ./smoke # macOS: DYLD_LIBRARY_PATH
# Windows (MinGW gcc, linking the DLL directly)
gcc examples/c/smoke.c -I bindings/c/include target/release/wickra.dll -lm -o smoke.exe
Expected output:
OK: wickra C ABI smoke passed (SMA streaming + batch + reset + NULL-safety + free)
The examples
| Example | What it does |
|---|---|
smoke.c |
Links the boundary and asserts SMA streaming / batch / reset / NULL-safety values. |
streaming.c |
Feeds a synthetic price series through SMA / EMA / RSI / MACD tick by tick. |
backtest.c |
Runs an indicator basket over an OHLCV CSV (defaults to the bundled daily dataset). |
multi_timeframe.c |
Resamples the bundled 1-minute CSV to 5m / 15m / 1h / 4h / 1d and prints indicators per timeframe. |
parallel_assets.c |
Serial vs OpenMP fan-out over a synthetic panel (one handle per asset), with speedup. |
strategy_rsi_mean_reversion.c |
Hourly RSI(14) mean-reversion with a PnL / Sharpe / max-drawdown summary. |
strategy_macd_adx.c |
Hourly MACD crossover gated by ADX(14) > 20. |
strategy_bollinger_squeeze.c |
Daily Bollinger-squeeze breakout with an ATR(14) stop. |
fetch_btcusdt.c |
Downloads BTCUSDT klines from the Binance REST API into examples/data/ (shells out to curl). |
live_binance.c |
Polls the Binance REST klines endpoint via curl and streams closed candles through RSI(14). |
smoke.cpp |
C++ RAII via wickra::Handle from wickra.hpp. |
ctest builds and runs every example except fetch_btcusdt and live_binance,
which reach the network and are built only — run those two by hand. The C ABI
exposes only the indicators, not the wickra-data IO layer, so the examples read
CSV (wickra_csv.h) and resample themselves; the network ones
shell out to the system curl rather than adding an HTTP/TLS dependency.
Usage shape
Every indicator follows the same five-function pattern over an opaque handle:
#include "wickra.h"
struct Sma *sma = wickra_sma_new(14); /* NULL on invalid params */
double v = wickra_sma_update(sma, 42.0); /* NaN during warmup */
wickra_sma_reset(sma); /* back to fresh state */
wickra_sma_free(sma); /* exactly once per _new */
There is no RAII across the C boundary: every wickra_<ind>_new must be paired
with exactly one wickra_<ind>_free. All functions are NULL-safe (a NULL handle
yields NaN / a no-op, never a crash).