Two unrelated newer-toolchain breakages bundled because they hit on the
same CI run and have the same shape (newer Rust got stricter about
patterns we used):
1. clippy 1.95 added the manual_midpoint lint which fires on every
instance of (a + b) / 2.0 with a help suggesting f64::midpoint.
CI runs with -D warnings so it became a hard error. Twelve sites
were affected — three real call sites in src/ohlcv.rs (median_price),
src/indicators/donchian.rs (DonchianOutput.middle),
src/indicators/ease_of_movement.rs (mid), and
src/indicators/super_trend.rs (hl2); plus eight test-helper
Candle::new constructions across accelerator_oscillator,
atr_trailing_stop, chaikin_volatility, chandelier_exit,
chande_kroll_stop, choppiness_index, super_trend, true_range.
All twelve switched to f64::midpoint (stable since Rust 1.85,
our workspace MSRV).
2. usize::is_multiple_of is still unstable (rust-lang/rust#128101) and
only stabilizes in Rust 1.87, but the MSRV CI job uses 1.85. The
two call sites in bollinger.rs and sma.rs (added with the R7
periodic-reseed tests) switched back to i % 2 == 0.
The first MSRV bump to 1.80 fixed the rayon-core floor but ran into a
second transitive-dep floor on the same CI run:
error: failed to parse manifest at clap_lex-1.1.0/Cargo.toml
Caused by: feature `edition2024` is required
The package requires the Cargo feature called `edition2024`, but
that feature is not stabilized in this version of Cargo (1.80.1).
clap_lex 1.1.0 is pulled in by criterion (dev-dep on the wickra crate)
via clap 4.6 -> clap_builder 4.6 -> clap_lex 1.1. edition2024 was
stabilized in Rust 1.85, so lift the workspace MSRV one more step. The
1.85 floor subsumes the rayon-core 1.80 requirement; bindings/node
stays at 1.88 (napi-build) which already covers everything below it.
Also fix the fuzz job: cargo-fuzz defaulted to building for
x86_64-unknown-linux-musl, which is not installed on the GitHub-hosted
ubuntu runner. Pass --target x86_64-unknown-linux-gnu explicitly on
every cargo fuzz run invocation so it builds for the actual host
target.
The previous MSRV pins were below what current transitive dependencies
need:
- rayon-core 1.13.0 (pulled in by the optional `parallel` feature on
wickra-core via rayon) requires rustc >= 1.80; under the old 1.75
workspace MSRV the CI MSRV job broke with "package `rayon-core
v1.13.0` cannot be built because it requires rustc 1.80 or newer".
- napi-build 2.3.2 (the build-script crate that napi-derive 2.x calls
into) requires rustc >= 1.88; under the old 1.77 node-binding MSRV
the CI MSRV-node job broke with "package `napi-build v2.3.2` cannot
be built because it requires rustc 1.88 or newer".
Pinning the deps backwards would have frozen us out of upstream
security/fix releases for both crates. Lifting the MSRV is the cleaner
path for a young 0.x library — downstream consumers on older toolchains
can stay on the already-published 0.2.0.
Updated:
- Cargo.toml workspace rust-version 1.75 -> 1.80
- bindings/node/Cargo.toml rust-version 1.77 -> 1.88
- .github/workflows/ci.yml MSRV matrix names + toolchain values + comment
- CHANGELOG.md [Unreleased] documents the bump
The previous `cargo install cargo-fuzz --locked` resolved against
cargo-fuzz's own Cargo.lock which pins to rustix 0.36.5. That rustix
version still annotates its source with internal #[rustc_*] attributes,
and the current nightly compiler rejects those attributes from
out-of-tree crates, so cargo-fuzz never finished compiling on CI:
error: attributes starting with `rustc` are reserved for use by the
`rustc` compiler
--> rustix-0.36.5/src/backend/linux_raw/io/errno.rs:28
error: could not compile `rustix` (lib) due to 4 previous errors
error: failed to compile `cargo-fuzz v0.13.1`
Switch the install to taiki-e/install-action, the same prebuilt-binary
provider we already use for cargo-llvm-cov. That skips the full
transitive-dependency compile and lets the fuzz-smoke job actually run.
The Z5 reorganisation moved every runnable example out of the per-crate
examples/ folders and into a dedicated wickra-examples crate at
examples/rust/, with the binaries living under src/bin/<name>.rs. The
old ci.yml Compile-examples step still pointed at the now-deleted
cargo example targets backtest (wickra) and live_binance (wickra-data),
which is why Rust windows-latest failed with 'no example target named
backtest in wickra package' and 'no example target named live_binance
in wickra-data package'.
Replace both calls with a single cargo build -p wickra-examples --bins.
That covers backtest, live_binance, fetch_btcusdt, multi_timeframe,
parallel_assets and streaming in one shot, and the wickra-examples
crate already enables the live-binance feature on its wickra-data dep
so no extra --features flag is needed.
The mathematical result of HistoricalVolatility on a perfectly geometric
price series is exactly zero — but the underlying 1.01_f64.powi(i) +
log-return + std-dev cascade accumulates platform-sensitive FP drift on
the order of 1e-7 on x86_64 Linux and macOS (the Windows result happened
to round closer to zero, which is why the test passed locally and on the
Windows CI runner but failed on Linux and macOS).
Bump the tolerance from 1e-9 to 1e-6. That stays four decimal places
below any realistic annualised volatility value while comfortably
absorbing the observed cross-platform drift.
Also extend the comment to document the rationale so the next person
who reads the test does not tighten it back down.
This release carries the full post-audit work — 46 new indicators
(25 → 71), an eight-family taxonomy restructure, new bindings for
RollingVWAP, the WASM streaming-update parity, the pyo3/numpy CVE
fix, the SMA/Bollinger drift bound, the O(1) LinearRegression
refactor, the UlcerIndex deque and the PSAR is_ready/reset fixes
plus a refreshed example suite and wiki. The earlier 0.1.5 number
was never published; jumping straight to 0.2.0 is the cleaner signal
for the scope of the change.
Bumped:
- Cargo.toml workspace + wickra-core workspace-dep version
- bindings/python/pyproject.toml
- bindings/node/package.json + optionalDependencies (six platform pins)
- 6 x bindings/node/npm/<target>/package.json
- Cargo.lock regenerated
- CHANGELOG.md [0.2.0] header + compare-link
- docs/wiki/Home.md published-versions table
- docs/wiki/Quickstart-{Rust,Node,WASM}.md + Warmup-Periods.md
version-pinned narrative lines
Verified locally:
- cargo fmt/clippy/test (628 passed, 0 failed)
- cargo deny check (no suppression)
- bindings/node node --test (92/92)
- bindings/python pytest (118/118)
- import wickra reports 0.2.0 with 72 indicator classes
Replace the TBD placeholder under ## [0.1.5] with today's date so the
release header is complete. The compare-link at v0.1.4...v0.1.5 is
already in place and points at the right pair.
Every other manifest in the repo carries the author string
kingchenc <kingchencp@gmail.com> — workspace Cargo.toml,
bindings/node/package.json, all per-platform npm package.jsons,
release.yml. pyproject.toml had neither authors nor maintainers, so
the PyPI listing fell back to the implicit empty value. Align with
the rest of the repo.
Indicators-Overview.md listed RollingVwap as its own row in the Volume
table (count = 9, total = 72), but Home.md and README.md treat it as
a sub-variant of Vwap (count = 8, total = 71). Drop the standalone row
in Indicators-Overview and fold the rolling variant into the Vwap row.
Warmup-Periods.md keeps both constructors (Vwap::new() and
RollingVwap::new(n) have different warmup periods, which is the whole
purpose of that page) but adds a note explaining that the row count
exceeds the 71 canonical indicators by one.
The module doc-comment in bindings/node/src/lib.rs still pointed at
@wickra/wickra. The package is published as bare "wickra" — every
README, Quickstart and example already uses that name. Align the
inline doc with the published name.
The Quickstart-Node API-surface table claimed `warmupPeriod()` was
"not exposed on every multi-output class". After the B-series fixes
in `todo-detailed.md` and the R3/R8 pass on this branch, every Node
indicator class — single- and multi-output, scalar- and candle-input —
exposes `warmupPeriod()`. Updated the note accordingly.
The Quickstart-WASM page only mentioned `MACD` and `BollingerBands` as
multi-output indicators. After R3 (this branch) every candle-input
WASM class also exposes a structured `update`: `Stochastic`, `ADX`,
`Keltner`, `Donchian`, `Aroon` (plus `SuperTrend`, which has always
been there).
The "Multi-output indicators" section now lists every multi-output
shape in a table, and a short note at the end calls out the 0.1.5
parity: every candle-input indicator ships `update` / `batch` /
`reset` / `isReady` / `warmupPeriod`, so browser code doesn't need
to replay `batch` on each tick anymore.
Three indicator pages get a short follow-up paragraph that surfaces an
internal implementation detail the audit findings made user-visible:
- `Indicator-LinearRegression.md` gains a "Complexity" section explaining
the O(1) update (precomputed `Σx`, `Σxx`; incrementally slid `Σy`,
`Σxy` via the closed-form sliding identity), and the existing
"Reset" bullet mentions the additional running accumulators. The same
story applies to `LinRegSlope` and `LinRegAngle` (the page now links
to both rather than repeating the derivation three times).
- `Indicator-Sma.md` and `Indicator-BollingerBands.md` mention the
periodic reseed (`16 · period` updates) that caps floating-point
drift on long-running streams. Amortised cost is still O(1) and the
user-facing behaviour on benign inputs is unchanged.
No behavioural claim, no API claim, no example changes — just narrative
catching up with the implementation.
Versions bumped to 0.1.5 in every authoritative location:
- workspace `Cargo.toml` (`[workspace.package].version`, the
`wickra-core` path dependency pin).
- `bindings/python/pyproject.toml`.
- `bindings/node/package.json` (main + all six `optionalDependencies`
pins).
- All six per-platform `bindings/node/npm/<target>/package.json`
templates.
CHANGELOG: the accumulated `[Unreleased]` block is promoted to
`[0.1.5] - TBD` (date left for the user to set at tag time); the new
`[Unreleased]` header sits empty above it; the compare link table is
extended with `[0.1.5]: …compare/v0.1.4...v0.1.5` and the
`[Unreleased]` link is repointed to `…compare/v0.1.5...HEAD`.
Wiki refresh for 0.1.5 (R20 + Z2):
- `Home.md` version pin table updated; the Quickstart-Node hint replaces
the "spam filter holding back Windows" caveat with "0.1.5 is the
first release in which `npm install wickra` works end-to-end on
Windows" (npm Support released the name on 2026-05-22).
- `Quickstart-Node.md`'s Windows caveat is rewritten to explain the
history (`0.1.1`–`0.1.4` of `wickra-win32-x64-msvc` are burned) and
the resolution (0.1.5+ installs cleanly).
- `Quickstart-Rust.md` version mention bumped.
- `Warmup-Periods.md` note bumped + corrected: every Node and WASM
class — single- and multi-output — now exposes `warmupPeriod()` after
R3 (this branch), not only the single-output ones.
`release.yml` `publish_dir` no longer silently swallows a
second-attempt platform-package publish failure with a `::warning::`
and `return 0`. A real failure (after the existing 30s retry) now
emits an `::error::` and fails the job. The original mask is exactly
what allowed the `wickra-win32-x64-msvc@0.1.1–0.1.4` spam-filter
rejections to land four times in a row without anyone noticing (audit
finding R20). Failing loud means the next regression of this shape is
caught at the release run, not by a Windows user trying to
`require('wickra')`.
This commit does NOT push, tag, or trigger a release — the user
publishes the 0.1.5 tag themselves once the manual npm-republish
smoke test confirms `wickra-win32-x64-msvc@0.1.5` accepts publish on
the freshly-released name.
PSAR — the wiki page still claimed "Node streaming. Not exposed in the
Node binding." That was true for an early release but is no longer:
the Node binding has exposed `psar.update(high, low, close)` since the
B1 fix in todo-detailed.md, and the WASM binding now exposes the same
streaming surface (audit finding R3, this branch). The page lists all
three streaming + batch shapes and adds a paragraph on the new
`is_ready` convention (audit finding R6 — flips on the first non-None
SAR, not on the seed candle).
HistoricalVolatility — the "Non-positive prices" edge-case note still
described the old behaviour ("that return is treated as 0"). The new
behaviour skips the bad tick entirely (audit finding R13): the
indicator's state is left untouched, the previous valid value is
returned, and the next real tick re-anchors against the previous
*valid* price. Updated the note to describe the new behaviour and
explain why (silently treating bad ticks as "no movement" underreports
realised volatility).
R10 — `cross-library-bench` previously ran on every push and every PR
to `main`, adding 5–10 minutes of build + bench time per CI run with
no automated consumer of the artefact. It moves to a dedicated
workflow (`.github/workflows/bench.yml`) that fires nightly at 03:00
UTC and on-demand via `workflow_dispatch` (with optional `size` /
`iterations` inputs). The job in `ci.yml` is removed and a pointer
comment is left in its place so future readers find the new home.
R11 — `pyproject.toml`'s `requires-python = ">=3.9"` already allowed
Python 3.13 installs, but the classifier list stopped at 3.12 and CI
tested only 3.9 / 3.11 / 3.12. The Python CI matrix gains `"3.13"`
and the matching `Programming Language :: Python :: 3.13` classifier
is added so PyPI listings and version-search tooling reflect the
actually-tested range.
R15 — `bindings/node/package.json` and all six per-platform subpackage
templates (`npm/{darwin-arm64,darwin-x64,linux-arm64-gnu,linux-x64-gnu,
win32-arm64-msvc,win32-x64-msvc}/package.json`) plus the WASM
`release.yml` enrich step switch the `license` field from the npm
convention `SEE LICENSE IN LICENSE` to the SPDX identifier
`PolyForm-Noncommercial-1.0.0`. npm's license search and downstream
tooling can now surface the actual license.
R18 — every `engines.node` field is bumped from `>= 16` to `>= 18`.
The package's test script (`node --test __tests__/`) uses the built-in
`node --test` runner that landed in Node 18; advertising support for
Node 16 / 17 was a guarantee we never verified (CI tests on Node 18 /
20 only) and would have produced a confusing error on those versions.
R19 — README's benchmark section gains an explicit hardware /
software-version block and reframes the absolute µs values as a
relative-speedup snapshot rather than a universal contract. Also tells
the reader how to reproduce locally (`pip install -e
bindings/python[bench]` + `python -m benchmarks.compare_libraries`) and
where the CI artefact lives.
Side effects:
- R16 / R17 (built wheel under `bindings/python/dist/` and
`examples/node/node_modules/`) — verified that both are already
covered by `.gitignore` (`*.whl`, `dist/`, `**/node_modules/`) and
were never tracked by git. The local directories have been cleared;
no `.gitignore` change needed and no committed file removed.
R13 — `HistoricalVolatility::update` previously substituted `0.0` for
the log-return whenever `prev <= 0` or `input <= 0`. The log-return is
undefined there, and silently treating bad ticks as "no movement"
underreports realised volatility on broken data feeds. The fix skips
non-positive prices entirely: `self.last` is returned, state is left
untouched, and the next real tick re-anchors against the previous
*valid* `prev_price`. This matches how every other indicator handles
invalid inputs (SMA / EMA / ROC / Bollinger).
A new test `skips_non_positive_prices` proves the invariant: after a
warmed-up indicator, two consecutive bad ticks (`-5.0` and `0.0`) must
return the baseline value, and a subsequent real positive tick must
produce the same output as a control indicator that simply never saw
the bad ticks.
R14 — `Tick::new` previously returned `Error::InvalidCandle` for
negative volume. A tick is not a candle; downstream tick-stream
pipelines should be able to match on a semantically-correct error. A
new `Error::InvalidTick { message }` variant is added; the existing
test is updated to assert against it. Python's `map_err` is extended
to forward the new variant as `PyValueError`; the Node and WASM
bindings format via `Error::to_string()` and pick the new variant up
automatically without source changes.
`Sma` and `BollingerBands` both maintained their running `sum` (and
`sum_sq` for Bollinger) with a single-subtract incremental update. That
is correct in exact arithmetic, but in f64 the sequence `sum -= old;
sum += new` on long streams with alternating large/small magnitudes
can accumulate catastrophic-cancellation error. Bollinger's existing
`.max(0.0)` clamp on the computed variance was a band-aid for the same
root cause — the drift had already driven the running variance below
zero.
The fix: every `16 · period` finite updates, reseed `sum` (and `sum_sq`
for Bollinger) from the live window. Amortised cost stays at O(1) —
`O(period)` work amortised over `O(period)` updates — and the reseed
strategy is named after the constant `RECOMPUTE_EVERY` so the
intention is clear at the call site.
Behaviour is unchanged on inputs that did not drift to begin with
(every existing test still passes, including `batch_equals_streaming`
and the SMA proptest). Two new stress tests
(`long_stream_drift_stays_bounded` in each module) feed a
magnitude-alternating stream for `5 · RECOMPUTE_EVERY · period`
updates and assert the reported value tracks a fresh from-scratch
computation over the live window to within tight tolerance — these
would have failed without the reseed on Bollinger's `sum_sq`.
The misleading `sma.rs` comment that claimed drift was already
bounded by recomputing the sum after each pop is rewritten to
describe the actual reseed strategy (audit finding L2-Rust).
Follow-up to b340ecd — the doc comment on
`warmup_period_matches_first_some_for_every_parameter_set` had an
unbalanced inline-code span ("`Some``") that tripped
`clippy::doc_invalid_doc_attributes` (caught by `clippy -D warnings`
but not by `cargo build` or `cargo test`). Rephrased the sentence so
every backtick is paired. No code change, no test change.
Audit finding R12 claimed `Coppock::warmup_period()` was off by one
because it returns `max(roc_long, roc_short) + wma`, while
`Roc::warmup_period() = period + 1`. After tracing the actual emission
sequence the existing formula is correct: when both ROCs reach `Some`
at 0-based index L (the slower of `roc_long_period` and
`roc_short_period`), the WMA receives its first input there and emits
its `wma_period`-th value at 0-based index `L + wma_period − 1`. The
`warmup_period()` is the 1-based count of inputs needed before the
first `Some`, i.e. `L + wma_period`. R12 was a misread by both Sonnet
audit agents and the Opus verifier — none of them traced the actual
emission timeline.
This commit:
- Expands the doc comment on `warmup_period` with the precise emission
argument and a worked example for `Coppock::new(6, 4, 3)` (the
existing test) so a future reader cannot mis-derive the formula.
- Adds `warmup_period_matches_first_some_for_every_parameter_set`,
which asserts `out[warmup - 1].is_some()` for five parameter
combinations — including the audit's smoking gun `(4, 2, 3)`. The
audit's proposed `max + 1 + wma` formula would have predicted index
7 (the 8th input) for that combination; the real first `Some` lands
at index 6 (the 7th input), exactly what the current formula
reports.
No behaviour change — the audit was wrong and the test makes the
contract regression-proof.
`LinearRegression::fit` and `LinRegSlope::update` previously iterated the
full `period`-window on every tick to recompute `Σy` and `Σxy` from
scratch — O(period) per update, in violation of the `Indicator` trait's
O(1) contract. `LinRegAngle` inherits the cost transitively because it
delegates to `LinRegSlope`.
This commit slides the OLS state in closed form. The constant terms
(`Σx`, `Σxx`, the denominator `n·Σxx − (Σx)²`) were already precomputed
in `new`. The new running state is:
- `sum_y: f64` — running sum of the values currently in the window.
- `sum_xy: f64` — running Σ(x · y) where `x` is the position of each
value inside the trailing window (`0` for the oldest, `n−1` for the
newest).
On every push, when the window is already full the front value `y₀` is
popped and the indices of every remaining value shift down by 1; the
identity
new_Σxy = old_Σxy − old_Σy + y₀
closes the slide in O(1). The new value is then pushed at position `k`
(the current length before the push), contributing `k · new_value` to
`sum_xy` and `new_value` to `sum_y`. The output is the same TA-Lib OLS
formula evaluated against the incremental accumulators.
Behaviour is unchanged: same per-tick values, same warmup, same NaN
semantics. Two new tests compare the O(1) result bar-by-bar against a
fresh O(n) refit on a noisy ramp (sliding-phase dominated), a step
function (large pop/push deltas), and constants (tests floating-point
drift) — agreement is within `1e-9`.
`LinRegAngle` benefits automatically through its `LinRegSlope` field.
The fuzz suite previously covered only `Rsi(14)` and `Ema(20)` — 2 of
71 indicators, no OHLCV coverage at all. Audit finding R9 asked for
ATR/ADX/Stochastic/PSAR as a minimum; this commit goes further and
brings every indicator under fuzz.
- `indicator_update` (rewritten): drives every scalar-input indicator
through one streaming pass + one batch call per iteration. Covers
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, BollingerBands. A `drive` helper
marked `#[inline(never)]` keeps each indicator on its own panic
backtrace frame.
- `indicator_update_candle` (new): chunks the fuzz `f64` stream into
`[open, high, low, close, volume]` tuples, builds candles via
`Candle::new` (skipping ones that fail OHLCV validation — that path
is fuzz-tested separately), then drives every candle-input indicator
through streaming + batch. Covers 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.
- `fuzz/Cargo.toml` registers the new target; `fuzz/README.md`
describes both expanded targets.
- A `fuzz-smoke` CI job runs each of the five targets for 30 s on
every push and pull-request — enough to catch a regression in the
harness without slowing CI to a crawl. Long fuzz campaigns belong
on dedicated infrastructure with persistent corpora.
`Psar::is_ready` previously returned `self.initialised`, which flips to
`true` *after* the seed candle — but the seed candle itself returns
`None`. The contract every other indicator honours is
`is_ready() == true` ↔ "the most recent update produced (or could
produce) a real value". Streaming consumers writing
`if ind.is_ready() { use(ind.update(c)?) }` would hit an unexpected
`None` on the first post-seed update.
Fix: add a `has_emitted: bool` field that flips on the first
`Some(sar)` return; `is_ready` now reads that. New test
`is_ready_only_after_first_some_value` pins the contract.
While in the same file, `reset()` is corrected to restore the compute
fields (`prev_high`, `prev_low`, `sar`, `ep`) to `f64::NAN` sentinels
instead of `0.0` (Opus bonus finding). The fields are gated by
`initialised` today, so the `0.0` sentinel never leaked into output —
but a future refactor that read them pre-init would have silently
treated `0.0` as a real price. A `debug_assert!` at the read site makes
the invariant explicit and catches a re-introduction of the bug in
debug builds.
Bit-equivalence with the previous behaviour is preserved
(`reset_allows_clean_reuse` and `batch_equals_streaming` continue to
pass unchanged).
`UlcerIndex::update` previously scanned the full `period`-window every
tick via `prices.iter().fold(NEG_INFINITY, f64::max)`, breaking the
`Indicator` trait's O(1) contract. For long windows (e.g. period 50+ on
a live tick stream) this turned a constant-time update into an O(period)
one, and full-history batch replays into O(n · period).
The window of raw prices is replaced with a monotonically-decreasing
deque of `(index, price)` pairs. On every push, all back entries
`<= input` are popped (they can never be the trailing max again, since
they are dominated and at least as old). On every step, the front is
popped if its index is older than `count - period + 1`. The deque's
front is therefore always the trailing max in O(1). `count: u64` is the
1-based input counter that drives expiration; on `reset()` it returns
to zero alongside the deque and the drawdown state.
Behaviour is unchanged: same per-tick values, same warmup
(`2 * period - 1`), same non-finite-input semantics. A new test
`monotone_deque_matches_naive_max_on_adversarial_inputs` compares the
deque output bar-by-bar against an independent O(n) trailing-max scan on
inputs designed to hit every code path: strictly increasing (full tail
pops), strictly decreasing (head expirations only), constants (the
`<= input` pop rule keeps a single newest entry), and a sawtooth.
The doc comment on `warmup_period()` is also corrected (B-Opus-2): the
two windows overlap by one bar, so the formula is `2 * period - 1`, not
`2 * period`.
The rolling-window VWAP indicator (`wickra_core::RollingVwap`) was only
available in the Rust crate, even though the README's Volume-family
table already advertised "VWAP (cumulative + rolling)" as a cross-
language feature. Users on Python, Node or in the browser had to fall
back to the cumulative `VWAP` or re-implement the rolling variant
themselves.
This commit closes the gap end-to-end:
- Python: `wickra.RollingVWAP(period)` — same constructor / `update` /
`batch` / `reset` / `is_ready` / `warmup_period` surface as `VWAP`,
plus a `period` property and a typed `__repr__`. The `__init__.py`
re-exports it and `__all__` lists it; the `.pyi` stub matches.
- Node: `RollingVWAP(period)` — napi class with the same lifecycle,
exported from `index.js` and declared in `index.d.ts`.
- WASM: `RollingVWAP(period)` — wasm-bindgen class with the same
`Float64Array` I/O as `VWAP`.
Tests added:
- Python: `test_rolling_vwap_streaming_matches_batch` — exercises
`update == batch` plus the full lifecycle on the shared OHLC fixture.
- Node: `RollingVWAP` row in the `candleScalar` parity table — covered
by the generic streaming-vs-batch + lifecycle harness.
- WASM: dedicated `wasm-bindgen-test` mirrors the Python test.
The wiki page `Indicator-Vwap.md` drops the "Rust-only" caveat and
gains Python / Node / WASM examples.
Twelve WASM classes previously exposed only `batch()` (and not even
`reset()` for ten of them): ADX, WilliamsR, CCI, MFI, PSAR, Keltner,
Donchian, VWAP, AwesomeOscillator, Aroon, Stochastic, OBV. Browser
consumers wanting per-tick updates had to replay `batch()` on every new
candle — the opposite of the library's streaming-first promise.
Each class now exposes:
- `update(...)` — per-tick streaming update with the same column inputs
as `batch()`. Single-output indicators return `Option<f64>`. Multi-
output indicators (ADX, Keltner, Donchian, Aroon, Stochastic) return a
named JS object (`{ plusDi, minusDi, adx }`, `{ upper, middle, lower }`,
`{ up, down }`, `{ k, d }`) once warm, or `null` during warmup. This
matches the existing `SuperTrend` convention so JS code can treat all
multi-output WASM indicators uniformly.
- `reset()`, `isReady()`, `warmupPeriod()` — bring the lifecycle API to
full parity with Python and Node.
`WasmKama` also gains the previously missing `warmupPeriod()` (R8). A
single new `wasm-bindgen-test` exercises every newly wired class against
a deterministic 40-bar synthetic OHLCV stream, asserting that
streaming `update` matches `batch` value-by-value and that the lifecycle
contract behaves the same as the core indicator.
Bumps the Python binding from pyo3 0.22 / numpy 0.22 to 0.28 / 0.28,
which resolves RUSTSEC-2025-0020 — a buffer overflow in
`PyString::from_object` that affected every published Python wheel.
Migration:
- `into_pyarray_bound(py)` → `into_pyarray(py)` (numpy 0.23 dropped the
`_bound` transitional suffix; the method now returns `Bound<'py, _>`
directly).
- `downcast::<PyDict>` → `cast::<PyDict>` (pyo3 renamed the method on
`PyAnyMethods`).
- Every `#[pyclass]` declares `skip_from_py_object` to opt out of the
now-deprecated automatic `FromPyObject` derive for `Clone` types.
Indicators are stateful — silently extracting them by value-clone is
never the intended FFI semantics.
- Workspace clippy gains `unused_self = "allow"` on the python crate
only: Python's `__repr__` protocol forces `&self` even for parameter-
less indicators where the body does not read state.
- `map_err` arms collapsed into a single `PyValueError` arm
(clippy::match_same_arms).
`deny.toml` no longer suppresses RUSTSEC-2025-0020; `cargo deny check`
is green on advisories, bans, licenses and sources without exceptions.
* examples/README.md — replace the single-row WASM section with a build
block (one-time `wasm-pack` build), a serve note, and the full
five-row table (`index`, `backtest`, `live_trading`, `multi_timeframe`,
`parallel_assets`).
* examples/wasm/README.md — new dedicated index for the WASM demos
with the build and serve commands and a description of every file
including the module worker companion.
* CHANGELOG.md `[Unreleased]` gains three bullets: the Python and Node
`fetch_btcusdt` siblings; the four new WASM browser demos; and the
three new wiki pages from Z6 (TA-Lib-Migration, Cookbook, FAQ).
Close the final "parallel assets" cell of the cross-language matrix for
WASM.
* examples/wasm/parallel_assets.html — generates a synthetic
`(assets, bars)` panel deterministically (the LCG matches the Node
and Rust siblings so timings are directly comparable), runs the
serial baseline on the main thread, then dispatches the same workload
to a pool of module Workers and reports the speedup. The render is
three cards (serial / parallel / speedup) plus a sanity-check line
asserting per-asset agreement between the two paths.
* examples/wasm/parallel_worker.js — companion module worker that
loads its own copy of the WebAssembly module via `init()` and
processes whichever slice of the panel its parent dispatches.
Modern browsers ship module-worker support (`new Worker(url, {
type: "module" })`) which lets every worker do `import init, { SMA, RSI
} from "../../bindings/wasm/pkg/wickra_wasm.js"` without bundler
glue. The inline page module and the worker module both syntax-check
cleanly under `node --check`.
Close the "multi-timeframe" cell of the cross-language matrix for WASM.
* examples/wasm/multi_timeframe.html — fetches the bundled 1-minute
BTCUSDT CSV (or any 1-minute OHLCV CSV), rolls it up in-page to 5m,
15m, 1h, 4h and 1d buckets, and prints RSI(14), MACD(12,26,9)
histogram and ADX(14) per timeframe via the WebAssembly bindings.
Same inline-bucket aggregation as the Node sibling, same indicator
set as the Python and Rust siblings — the Rust version uses
`wickra-data::Resampler` directly which is currently a Rust-only API.
The render is a single table (one row per timeframe) so the cross-
language outputs sit side-by-side cleanly. The inline module script
syntax-checks cleanly under `node --check`.
Close the "live trading" cell of the cross-language matrix for WASM.
* examples/wasm/live_trading.html — opens a native browser `WebSocket`
to Binance's public kline stream, feeds every close through RSI(14),
MACD(12,26,9) and Bollinger(20, 2.0) via the WebAssembly bindings, and
flags BUY/SELL candidates when all three indicators agree. Mirrors
the Node and Python live-trading examples in indicator set, signal
logic and symbol/interval validation — the symbol is checked against
`^[A-Za-z0-9]+$` before being spliced into the stream URL, the
interval against the public-API allow-list.
The UI shows live close / RSI / MACD-histogram / Bollinger-band cards,
plus a scrolling log of the last 200 ticks with signal rows highlighted.
Browser-native `WebSocket` means no library dependency. Build the WASM
module once (`wasm-pack build bindings/wasm --target web --release
--features panic-hook`), serve the repository root and open
`examples/wasm/live_trading.html`.
The WASM example set had only `index.html` (the streaming canvas demo);
the "backtest" cell of the cross-language matrix was empty. Close it.
* examples/wasm/backtest.html — loads the wasm-pack `--target web`
bundle, fetches an OHLCV CSV from the same `examples/data/` directory
the other languages use (default: `btcusdt-1d.csv`), parses it
in-page and streams every candle through SMA, EMA, RSI, MACD,
Bollinger, ATR, ADX and OBV via the WebAssembly bindings. Renders a
summary table with mean / min / max / last per series — mirrors the
Rust, Python and Node backtest examples both in indicator set and in
output shape.
Build the WASM module once (`wasm-pack build bindings/wasm --target web
--release --features panic-hook`), then serve the repository root and
open `examples/wasm/backtest.html`. The inline module script
syntax-checks cleanly under `node --check` on the extracted body.
Node had no sibling for the Rust and Python `fetch_btcusdt`
data-generators — adding it closes the "fetch (data-gen)" cell for the
last remaining row of the cross-language example matrix where the
pattern makes sense.
* examples/node/fetch_btcusdt.js — uses Node 18+'s built-in global
`fetch` (no npm dependencies); same pagination logic as the Rust and
Python siblings (paginate backwards via `endTime`, drop the
in-progress bucket, sort and trim to the configured target). Applies
the same OHLC validity check the Rust `Candle::new` constructor
enforces so a malformed kline is skipped rather than written.
* JavaScript's `String(v)` already gives the shortest round-trip
representation and strips the `.0` suffix for whole-number floats, so
the CSV output is byte-for-byte identical to what the Rust and
Python fetchers produce on the same Binance snapshot. Verified by
running it and `git diff`-ing against the checked-in dataset: every
row older than the run is unchanged; only the most recent ~24 hours
drift because the market kept moving.
examples/README.md gains the new row.
Python had no sibling for the Rust `fetch_btcusdt` data-generator —
adding it closes the "fetch (data-gen)" cell for Python and lets users
without a Rust toolchain regenerate the bundled BTCUSDT datasets.
* examples/python/fetch_btcusdt.py — uses only the standard library
(urllib.request + json + csv); same pagination strategy as the Rust
version (paginate backwards via `endTime`, drop the in-progress
bucket, sort and trim to the configured target). Applies the same
OHLC validity check the Rust `Candle::new` constructor enforces
(finite fields, high >= low/open/close, low <= open/close,
volume >= 0) so a malformed kline is skipped rather than written.
* Number formatting matches Rust's `f64` Display: shortest round-trip,
no trailing `.0` for whole-number floats. Verified by running the
script and `git diff`-ing against the checked-in dataset: every row
older than the run is byte-identical to Rust's output; the diff only
shows the most recent ~24 hours where Binance has produced fresh
candles since the original snapshot.
examples/README.md gains the new row.
The README and Streaming-vs-Batch benchmark tables were a stale snapshot
("5 000-bar series", numbers from an older machine). Re-run
`python -m benchmarks.compare_libraries` on the current hardware against
the same peer set (finta + talipp; TA-Lib and pandas-ta stay excluded on
Windows) and replace the tables with the fresh numbers.
The new run uses the script's current defaults: a 20 000-bar batch series
and a 5 000-bar seed + 15 000-bar live streaming workload — both more
representative of real backtests than the previous 5 000 / 2 000-bar
sizes. Wickra still wins every batch row outright (3.5× to 1 244× faster
than the nearest peer) and the streaming RSI is ~13.8× faster than
talipp's incremental implementation.
Three content gaps in the wiki: there was no migration story for users
porting from TA-Lib, no strategy cookbook, and no FAQ. Add all three as
self-contained pages and link them from Home.md's "Wiki contents".
* docs/wiki/TA-Lib-Migration.md — full one-to-one mapping table from
every common talib.X(...) call to the equivalent Wickra expression,
plus a "what Wickra has that TA-Lib does not" / "what TA-Lib has that
Wickra does not (yet)" delta.
* docs/wiki/Cookbook.md — seven concrete strategy recipes (RSI mean
reversion, MACD histogram crossover, Bollinger breakout, ADX-gated
trend, multi-timeframe confirmation, SuperTrend trailing stop,
Chain<EMA, RSI>) with Rust or Python snippets.
* docs/wiki/FAQ.md — common questions on warmup, NaN handling, thread
safety, installation, performance and comparing Wickra to TA-Lib /
pandas-ta / talipp / finta.
Also extend the [Unreleased] CHANGELOG entry that records the
examples/<lang>/ restructure with the wiki additions; Home.md gains
three new bullets under "Wiki contents".
Python's parallel_assets.py demoed GIL-release multi-core throughput;
Rust and Node both lacked a sibling that shows their own native
parallelism. Close the gap with two real, runnable examples.
* examples/rust/src/bin/parallel_assets.rs — synthesises an (assets,
bars) panel with a deterministic per-asset LCG, runs a serial baseline,
then `Sma::batch_parallel` / `Rsi::batch_parallel` via rayon, asserts
the two outputs are element-wise identical and prints the speedup.
Toggle indicator with `--indicator sma|rsi`.
* examples/node/parallel_assets.js — same shape, but the parallel run is
a `worker_threads` pool that re-loads the native binding in each
worker. Each worker computes the last non-null indicator value for its
slice; the main thread aggregates and verifies serial == parallel
per asset.
Both examples report timings and the serial-vs-parallel sanity check
passes. Defaults (200 × 5000) keep the example fast on dev hardware;
larger `--assets`/`--bars` is where the speedup numbers move (Node's
worker spawn cost dominates the smallest sizes, which is honest and
educational).
examples/README.md gains the two new rows.
Python's examples/python/multi_timeframe.py had no Rust or Node sibling.
Add both — the Rust version uses wickra-data's `Resampler` /
`resample_all` (the canonical path; no manual roll-up), the Node version
mirrors the Python one's inline aggregation because wickra-data's
resampler is currently Rust-only.
* examples/rust/src/bin/multi_timeframe.rs — reads the bundled 1m CSV via
`CandleReader`, resamples to 5m / 15m / 1h / 4h / 1d via `resample_all`,
prints last RSI(14), MACD(12,26,9) histogram and ADX(14) per timeframe.
* examples/node/multi_timeframe.js — same outputs from a hand-rolled
bucket aggregator; reuses the new examples/data/ default path.
* examples/README.md gains the new rows.
Run side by side: the Rust and Node summaries are bit-identical at every
timeframe (50000 / 10000 / 3334 / 834 / 209 / 35 bars; same RSI, MACD
histogram and ADX to two decimals) — confirming both the Rust resampler
and the inline Node aggregator produce the same OHLC buckets.
Python and Rust both lacked a standalone "streaming indicators" example
that mirrors examples/node/streaming.js — the quickstart docs cover the
pattern, but a runnable file makes the parity visible across all four
languages.
* examples/python/streaming.py — argparse-driven synthetic streaming demo
feeding SMA(20) / EMA(20) / RSI(14) / MACD(12,26,9), tagging BUY?/SELL?
candidates when RSI extremes and MACD-histogram direction agree.
* examples/rust/src/bin/streaming.rs — same demo as a wickra-examples
binary, reusing the seeded LCG so its first 40 rows are bit-identical
to the Python (and Node) sibling — a strong cross-language consistency
signal verified by running both side by side.
* examples/README.md gains a `streaming` row in the Rust and Python tables.
Finish the per-language `examples/<lang>/` restructure by relocating the
WASM browser demo from bindings/wasm/examples/ to examples/wasm/.
* `examples/wasm/index.html` is the moved file; its WASM module import
becomes `../../bindings/wasm/pkg/wickra_wasm.js` so the demo still loads
the wasm-pack output without copying it.
* bindings/wasm/README.md, Quickstart-WASM.md, examples/README.md and the
root README "Languages" + project-layout block all point at the new
path. The serve command in the docs now says "serve the repository root
and open examples/wasm/index.html".
`bindings/wasm/examples/` is empty after the move; the now-empty
directory is removed.
Continue the per-language `examples/<lang>/` restructure: move the three
Node example files (streaming.js, backtest.js, live_trading.js) out of
bindings/node/examples/ and into a top-level examples/node/ directory.
* `examples/node/package.json` is a `private` package that pulls the
native binding via `file:../../bindings/node` and lists `ws` as a
dev-dependency for the live-trading example. `require('..')` in each
file becomes `require('wickra')` — exactly what a downstream user would
write — and the file-header run instructions are updated to the new
two-step workflow (`npm install` in bindings/node, then in
examples/node).
* `backtest.js`'s default-CSV path becomes the much shorter
`__dirname/../data/btcusdt-1d.csv` from the new location.
* `bindings/node/package.json` drops the now-unused `ws` devDependency.
* `.gitignore` is broadened from `bindings/node/node_modules/` to
`**/node_modules/` so the new `examples/node/node_modules/` directory is
not tracked.
* The README "Languages" table, project-layout block and
`examples/README.md` Node section are updated for the new paths and run
commands.
Verified by running `node backtest.js` (3200 BTCUSDT daily bars, matching
output), `node streaming.js`, and `node --check live_trading.js` from the
new location.
The three Rust examples (backtest, fetch_btcusdt, live_binance) used to
live each in their own crate's examples/ dir, splitting the example set
across crates and burying it inside the source tree. Move them into a new
workspace member crate at `examples/rust/` (package `wickra-examples`,
`publish = false`) so all language examples sit under one top-level
`examples/<lang>/` tree.
* `examples/rust/Cargo.toml` declares the per-binary deps (wickra,
wickra-data with the `live-binance` feature always on, serde_json, tokio
for the macro and current-thread runtime).
* `examples/rust/src/bin/{backtest,fetch_btcusdt,live_binance}.rs` are the
three migrated binaries; their doc-comments and the fetch_btcusdt output
path are updated for the new location and run command
(`cargo run -p wickra-examples --bin <name>`).
* Workspace `Cargo.toml` lists the new member; the now-empty
`[dev-dependencies]` extras (`wickra`, `tokio` in wickra-data and
`serde_json` in wickra) that existed only for these examples are dropped.
* The `[[example]] live_binance` table is removed from wickra-data's
manifest since the file moved out.
* README "Languages" + project-layout, examples/README.md, Quickstart-Rust
and Data-Layer are pointed at the new paths and commands.
`cargo build -p wickra-examples` and `cargo run --release -p wickra-examples
--bin backtest -- examples/data/btcusdt-1d.csv` both succeed; the rest of
the workspace (core, data, wickra) builds, clippies (`--all-targets -D
warnings`) and tests (508 core + 28 data + 1 integration + 74+3+1
doctests) all stay green.
The seven BTCUSDT OHLCV datasets used to live under
crates/wickra/examples/data/, which buried them inside a Rust crate even
though the Node backtest example and the upcoming Rust/Node/WASM example
restructure need to reach them too. Move them to the workspace-level
examples/data/ so every language's examples can resolve the same path.
The bench (crates/wickra/benches/indicators.rs), the example_data
integration test, fetch_btcusdt.rs and the Node backtest example all take
the new ../../examples/data/ path; Data-Layer.md, examples/README.md and
the CHANGELOG entry are updated to match. No data file content changes.
The top-level examples/ directory held only python/, which made the
examples look Python-only even though Rust, Node and WASM all ship their
own. Add examples/README.md: a single index of every runnable example
across Rust, Python, Node and WASM, each with its run command, plus a note
on the bundled BTCUSDT datasets.
Point the README "Languages" table at the Node backtest example and link
the new index from both the table and the project-layout section.
Mirror examples/python/live_trading.py for the Node binding: connect to the
public Binance kline WebSocket, stream close prices through RSI / MACD /
Bollinger Bands, and print BUY/SELL candidate signals when all three agree.
The symbol and interval are validated before being spliced into the stream
URL, and non-kline frames (acks, heartbeats) are skipped.
Uses the standard `ws` package, added as a devDependency so it installs
with `npm install` for anyone running the examples but never reaches a
consumer of the published package.
The Node binding shipped only one example (a synthetic streaming demo),
while Python and Rust both have a CSV backtest. Add the Node counterpart of
examples/python/backtest.py and crates/wickra/examples/backtest.rs: it reads
an OHLCV CSV, streams every candle through a basket of indicators (SMA, EMA,
RSI, MACD, Bollinger Bands, ATR, ADX, OBV) via the O(1) update call, and
prints a per-series summary.
With no argument it runs against the bundled BTCUSDT daily dataset, so it is
runnable out of the box; pass a path to use any other OHLCV CSV.
The Python binding README still advertised "63 indicators across four
families" with the pre-restructure five-group taxonomy, missing the eight
indicators added since. Update it to "71 indicators across eight families"
with the catalogue grouped to match the main README.
The Node binding README referred to the package as @wickra/wickra in its
title, install command and import example; the published package is named
wickra (per bindings/node/package.json). Correct all three.
Add the still-unreleased Z1/Z2 work to the [Unreleased] section: the seven
real-BTCUSDT example datasets plus the fetch_btcusdt example, the
Timeframe::minutes/hours/days constructors, and the switch of the indicator
benchmarks from a synthetic series to the checked-in BTCUSDT dataset.
kingchenc