# OpenQuant

> Rust implementations of financial machine-learning methods, with Python bindings.

Source revision: 6335a56fa5f20abfb89d876f9eb90792eac1460b
Working tree: clean

This index is generated from the documentation collection and Python reference snapshot.
The source revision identifies the site build, not a separately verified installed package. Match your interpreter and imported package paths to the intended checkout before testing changes.
Before resuming work, preserve local edits and identify the selected interpreter, imported extension, dataset, registry and output paths; inspect saved artifacts before rerunning.
Read exact signatures, return shapes and limitations in the linked API reference; this file does not copy them.
The Python distribution is pyopenquant; import it as openquant. Do not install the unrelated openquant PyPI project.
Page status describes documentation review, not numerical correctness. Authored does not mean human-reviewed.

## Start here

- [Overview](https://open-quant.github.io/openquant/): [authored] Rust implementations of the methods in Advances in Financial Machine Learning, with Python bindings.
- [Quickstart](https://open-quant.github.io/openquant/quickstart/): [authored] Build OpenQuant from source, run a synthetic pipeline and interpret its output.
- [Agent guide](https://open-quant.github.io/openquant/agents/): [authored] Entry points, canonical sources and verification steps for agents using or changing OpenQuant.
- [Module Reference Index](https://open-quant.github.io/openquant/modules/): [generated] Full OpenQuant module documentation index with AFML-aligned summaries.
- [Research gallery](https://open-quant.github.io/openquant/runbooks/): [generated] Every research runbook in notebooks/python: its pre-registered hypothesis, figures, result and promotion decision, taken from the executed notebooks.

## Task routes

- [Find an API by task](https://open-quant.github.io/openquant/module-reference/api-task-guide/): [authored] Choose the Python wrapper, native binding or Rust-only entry point for a concrete task before copying an API call.
- [Python Bindings Setup](https://open-quant.github.io/openquant/setup/python-bindings/): [authored] Install or rebuild Python bindings; verify the imported extension before testing.
- [Rust Core Workflow](https://open-quant.github.io/openquant/workflows/rust-core-workflow/): [authored] Follow a Rust component walkthrough and inspect alignment, numerical conventions and its synthetic-input limits.
- [Notebook Research Workflow](https://open-quant.github.io/openquant/workflows/notebook-research-workflow/): [authored] Run or reproduce notebooks with the correct tier, environment and evidence.
- [Select data available at a historical decision](https://open-quant.github.io/openquant/workflows/point-in-time-data/): [authored] Select revisions available at a research cutoff; distinguish synthetic selection checks and content/policy identity from provider-history evidence.
- [provenance](https://open-quant.github.io/openquant/modules/provenance/): [authored] Select supplied revisions before cleaning and bind exact typed values through stable-ID rejoining, without claiming provider authentication.
- [feature_assembly](https://open-quant.github.io/openquant/modules/feature-assembly/): [authored] Require complete declared source fields, retain ordered lineage and missing-input exclusions, and inspect coverage without assuming representativeness.
- [availability](https://open-quant.github.io/openquant/modules/availability/): [authored] Intersect original training indices with explicit source-availability and label-maturity boundaries; keep purging and learned preprocessing separate.
- [data](https://open-quant.github.io/openquant/modules/data/): [authored] Load data and establish cleaning, cache and provenance contracts before analysis.
- [evaluation](https://open-quant.github.io/openquant/modules/evaluation/): [authored] Define explicit trial identity, registry scope and shared-read limitations before comparing searches.
- [End-to-end research journey](https://open-quant.github.io/openquant/workflows/research-journey/): [authored] Follow fold-local fitting and read back the saved JSON event, fold and prediction handoffs.
- [Select a model and evaluate a chronological holdout](https://open-quant.github.io/openquant/workflows/model-selection-holdout/): [authored] Separate development-only parameter selection and fold-local preprocessing from refitting and final chronological holdout classification.
- [Align observations and calculate portfolio weights](https://open-quant.github.io/openquant/workflows/multi-asset-allocation/): [authored] Align long-format observations, matrix axes and named weights; preserve moment units and fitting boundaries without treating allocation as a realized backtest.
- [Examples Catalog](https://open-quant.github.io/openquant/examples/catalog/): [authored] Choose runnable entry points and audit saved grid artifacts without confusing flags with validation.
- [cross_validation](https://open-quant.github.io/openquant/modules/cross-validation/): [authored] Construct label-span-aware evaluation and preserve row alignment.
- [Local Build Setup](https://open-quant.github.io/openquant/setup/local-build/): [authored] Build documentation and distinguish preview, generation and aggregate checks.
- [Troubleshooting](https://open-quant.github.io/openquant/setup/troubleshooting/): [authored] Diagnose command, interpreter and extension failures.

## API reference

- [Python API](https://open-quant.github.io/openquant/api/python/): Generated signatures and docstrings. Prefer public wrappers to compiled internals.
- [Rust API](https://open-quant.github.io/openquant/api/rust/openquant/): Generated rustdoc for the numerical core.
- [openquant.adapters](https://open-quant.github.io/openquant/api/python/adapters/): [generated] python module.
- [openquant.availability](https://open-quant.github.io/openquant/api/python/availability/): [generated] python module.
- [openquant.backtest_stats](https://open-quant.github.io/openquant/api/python/backtest_stats/): [generated] compiled module.
- [openquant.backtesting_engine](https://open-quant.github.io/openquant/api/python/backtesting_engine/): [generated] python module.
- [openquant.bars](https://open-quant.github.io/openquant/api/python/bars/): [generated] python module.
- [openquant.bet_sizing](https://open-quant.github.io/openquant/api/python/bet_sizing/): [generated] compiled module.
- [openquant.cla](https://open-quant.github.io/openquant/api/python/cla/): [generated] compiled module.
- [openquant.codependence](https://open-quant.github.io/openquant/api/python/codependence/): [generated] compiled module.
- [openquant.cross_validation](https://open-quant.github.io/openquant/api/python/cross_validation/): [generated] python module.
- [openquant.data](https://open-quant.github.io/openquant/api/python/data/): [generated] python module.
- [openquant.dynamic_allocation](https://open-quant.github.io/openquant/api/python/dynamic_allocation/): [generated] compiled module.
- [openquant.ef3m](https://open-quant.github.io/openquant/api/python/ef3m/): [generated] compiled module.
- [openquant.ensemble](https://open-quant.github.io/openquant/api/python/ensemble/): [generated] compiled module.
- [openquant.evaluation](https://open-quant.github.io/openquant/api/python/evaluation/): [generated] python module.
- [openquant.fast_ewma](https://open-quant.github.io/openquant/api/python/fast_ewma/): [generated] compiled module.
- [openquant.feature_assembly](https://open-quant.github.io/openquant/api/python/feature_assembly/): [generated] python module.
- [openquant.feature_diagnostics](https://open-quant.github.io/openquant/api/python/feature_diagnostics/): [generated] python module.
- [openquant.feature_importance](https://open-quant.github.io/openquant/api/python/feature_importance/): [generated] python module.
- [openquant.filters](https://open-quant.github.io/openquant/api/python/filters/): [generated] compiled module.
- [openquant.fracdiff](https://open-quant.github.io/openquant/api/python/fracdiff/): [generated] compiled module.
- [openquant.hcaa](https://open-quant.github.io/openquant/api/python/hcaa/): [generated] compiled module.
- [openquant.hrp](https://open-quant.github.io/openquant/api/python/hrp/): [generated] compiled module.
- [openquant.hyperparameter_tuning](https://open-quant.github.io/openquant/api/python/hyperparameter_tuning/): [generated] python module.
- [openquant.labeling](https://open-quant.github.io/openquant/api/python/labeling/): [generated] compiled module.
- [openquant.microstructural](https://open-quant.github.io/openquant/api/python/microstructural/): [generated] compiled module.
- [openquant.onc](https://open-quant.github.io/openquant/api/python/onc/): [generated] compiled module.
- [openquant.pipeline](https://open-quant.github.io/openquant/api/python/pipeline/): [generated] python module.
- [openquant.portfolio](https://open-quant.github.io/openquant/api/python/portfolio/): [generated] compiled module.
- [openquant.provenance](https://open-quant.github.io/openquant/api/python/provenance/): [generated] python module.
- [openquant.research](https://open-quant.github.io/openquant/api/python/research/): [generated] python module.
- [openquant.risk](https://open-quant.github.io/openquant/api/python/risk/): [generated] compiled module.
- [openquant.sample_weights](https://open-quant.github.io/openquant/api/python/sample_weights/): [generated] compiled module.
- [openquant.sampling](https://open-quant.github.io/openquant/api/python/sampling/): [generated] compiled module.
- [openquant.sb_bagging](https://open-quant.github.io/openquant/api/python/sb_bagging/): [generated] compiled module.
- [openquant.strategy_risk](https://open-quant.github.io/openquant/api/python/strategy_risk/): [generated] compiled module.
- [openquant.streaming_hpc](https://open-quant.github.io/openquant/api/python/streaming_hpc/): [generated] compiled module.
- [openquant.structural_breaks](https://open-quant.github.io/openquant/api/python/structural_breaks/): [generated] compiled module.
- [openquant.synthetic_bt](https://open-quant.github.io/openquant/api/python/synthetic_bt/): [generated] compiled module.
- [openquant.viz](https://open-quant.github.io/openquant/api/python/viz/): [generated] python module.
- [openquant.volatility](https://open-quant.github.io/openquant/api/python/volatility/): [generated] compiled module.

## Guides and module explanations

- [Coverage Dashboard](https://open-quant.github.io/openquant/coverage/): [authored] What is documented, what is not, and the commands that produce those numbers.
- [By AFML Chapter](https://open-quant.github.io/openquant/module-reference/by-afml-chapter/): [reviewed] Map AFML concepts to concrete OpenQuant modules and workflows.
- [adapters](https://open-quant.github.io/openquant/modules/adapters/): [authored] Polars DataFrame adapters for signals, events, weights, backtest curves, and streaming buffers.
- [backtest_statistics](https://open-quant.github.io/openquant/modules/backtest-statistics/): [authored] Sharpe ratios with their uncertainty — probabilistic and deflated Sharpe, minimum track record length — plus drawdown, concentration and holding-period statistics.
- [backtesting_engine](https://open-quant.github.io/openquant/modules/backtesting-engine/): [authored] Walk-forward, purged cross-validation and combinatorial purged cross-validation splits, with the out-of-sample paths CPCV produces.
- [bet_sizing](https://open-quant.github.io/openquant/modules/bet-sizing/): [authored] From a model's confidence, or a price forecast, to a position size: probability-based sizing, averaging of live bets, discretisation, dynamic limit prices and concurrency-based sizing.
- [cla](https://open-quant.github.io/openquant/modules/cla/): [authored] Critical Line Algorithm portfolios: a bounded efficient-frontier walk, with a constrained minimum-variance solution for equal forecasts.
- [codependence](https://open-quant.github.io/openquant/modules/codependence/): [authored] Measures of dependence that are true metrics or that see non-linear relationships: correlation distances, distance correlation, mutual information and variation of information.
- [combinatorial_optimization](https://open-quant.github.io/openquant/modules/combinatorial-optimization/): [authored] Exhaustive search over small integer problems, and over single-instrument trading paths with a fixed cost per trade, as an exact baseline for heuristic solvers.
- [data_processing](https://open-quant.github.io/openquant/modules/data-processing/): [authored] Rust OHLCV preparation: stable deduplication, frequency-aware gap reports, and explicit missing and off-grid bars on regular calendars.
- [data_structures](https://open-quant.github.io/openquant/modules/data-structures/): [authored] Time, tick, volume, dollar, run and imbalance bars built from a stream of trades.
- [dynamic_allocation](https://open-quant.github.io/openquant/modules/dynamic-allocation/): [authored] Exact multi-asset portfolio trajectories on a discrete weight grid, scored by Sharpe ratio after square-root transaction costs.
- [ef3m](https://open-quant.github.io/openquant/modules/ef3m/): [authored] EF3M: fit a two-Gaussian mixture with randomized moment matching; inspect residual error and distinguish individual fits from parameter-wise modes.
- [ensemble_methods](https://open-quant.github.io/openquant/modules/ensemble-methods/): [authored] Diagnostics for bagged ensembles: how much variance averaging removes, given how correlated the estimators are.
- [etf_trick](https://open-quant.github.io/openquant/modules/etf-trick/): [authored] Turn a rebalanced basket of futures, or a single rolled contract, into one continuous series whose changes are achievable PnL.
- [feature_diagnostics](https://open-quant.github.io/openquant/modules/feature-diagnostics/): [authored] Python feature-importance reports on purged folds: MDA, SFI, a coefficient-based stand-in for MDI, PCA orthogonalisation, substitution effects and a pre-model feature screen.
- [feature_importance](https://open-quant.github.io/openquant/modules/feature-importance/): [authored] MDI, MDA and SFI feature importance, and a PCA cross-check, for models validated on purged folds.
- [filters](https://open-quant.github.io/openquant/modules/filters/): [authored] The symmetric CUSUM filter and a rolling z-score filter, for sampling events from a price series.
- [fingerprint](https://open-quant.github.io/openquant/modules/fingerprint/): [authored] Decompose what a fitted model has learned into linear, non-linear and pairwise-interaction effects per feature.
- [fracdiff](https://open-quant.github.io/openquant/modules/fracdiff/): [authored] Fractional differentiation: make a price series stationary while keeping as much of its memory as possible.
- [hcaa](https://open-quant.github.io/openquant/modules/hcaa/): [authored] Hierarchical allocation down the cluster tree, cut into a chosen number of clusters, with a choice of risk measure: variance, standard deviation, expected shortfall, conditional drawdown, Sharpe ratio or equal splits.
- [hpc_parallel](https://open-quant.github.io/openquant/modules/hpc-parallel/): [authored] AFML's atoms and molecules: split a job into contiguous chunks of equal work, run a callback over each in serial or on threads, and get the outputs back in order.
- [hrp](https://open-quant.github.io/openquant/modules/hrp/): [authored] Hierarchical Risk Parity: portfolio weights from a clustering of the correlation matrix, with no matrix inversion.
- [hyperparameter_tuning](https://open-quant.github.io/openquant/modules/hyperparameter-tuning/): [authored] Grid and randomised hyperparameter search on purged k-fold splits, scored with sample weights.
- [labeling](https://open-quant.github.io/openquant/modules/labeling/): [authored] Triple-barrier labels and meta-labels: what happened after each event, measured in units of that moment's volatility.
- [microstructural_features](https://open-quant.github.io/openquant/modules/microstructural-features/): [authored] Spread, price-impact, order-flow and entropy features estimated from bars or from trades.
- [onc](https://open-quant.github.io/openquant/modules/onc/): [authored] Optimal Number of Clusters: partition a correlation matrix with k-means, choosing the number of clusters by silhouette quality.
- [pipeline](https://open-quant.github.io/openquant/modules/pipeline/): [authored] CUSUM signals, a single-instrument backtest, separate portfolio allocation, and ordering checks.
- [portfolio_optimization](https://open-quant.github.io/openquant/modules/portfolio-optimization/): [authored] Mean-variance allocation with weight bounds: inverse variance, minimum volatility, maximum Sharpe ratio, and minimum risk for a target return.
- [research](https://open-quant.github.io/openquant/modules/research/): [authored] Synthetic datasets, transaction-cost heuristics, research comparison grids, and promotion flags.
- [risk_metrics](https://open-quant.github.io/openquant/modules/risk-metrics/): [authored] Historical value at risk, expected shortfall, conditional drawdown at risk and portfolio variance.
- [sample_weights](https://open-quant.github.io/openquant/modules/sample-weights/): [authored] Training weights for overlapping labels: return attribution and time decay.
- [sampling](https://open-quant.github.io/openquant/modules/sampling/): [authored] Label concurrency, average uniqueness and the sequential bootstrap, for training sets whose labels overlap in time.
- [sb_bagging](https://open-quant.github.io/openquant/modules/sb-bagging/): [authored] A bagging ensemble that draws each estimator's sample with the sequential bootstrap, around a deliberately simple one-feature base learner.
- [strategy_risk](https://open-quant.github.io/openquant/modules/strategy-risk/): [authored] How precise and how frequent a strategy's bets must be to reach a target Sharpe ratio, and the probability that it falls short.
- [streaming_hpc](https://open-quant.github.io/openquant/modules/streaming-hpc/): [authored] VPIN and a venue-concentration HHI updated event by event in constant memory, with an alert when both cross their thresholds, and a synthetic flash crash to calibrate on.
- [structural_breaks](https://open-quant.github.io/openquant/modules/structural-breaks/): [authored] Tests for a change of regime in a price series: the supremum ADF test for explosive behaviour, a Chow-type Dickey-Fuller test, and the Chu-Stinchcombe-White CUSUM test.
- [synthetic_backtesting](https://open-quant.github.io/openquant/modules/synthetic-backtesting/): [authored] Choose profit-taking and stop-loss levels on simulated paths of a fitted mean-reverting process, instead of on the one historical path.
- [util::fast_ewma](https://open-quant.github.io/openquant/modules/util-fast-ewma/): [authored] Fast EWMA primitive shared across feature and volatility routines.
- [util::volatility](https://open-quant.github.io/openquant/modules/util-volatility/): [authored] Volatility estimators used across labeling and risk workflows.
- [viz](https://open-quant.github.io/openquant/modules/viz/): [authored] Visualization payload builders for feature importance, drawdown, regime, frontier, and cluster charts.
- [Changelog](https://open-quant.github.io/openquant/project/changelog/): [generated] Notable changes to OpenQuant, rendered from CHANGELOG.md.
- [Governance](https://open-quant.github.io/openquant/project/governance/): [authored] The rules OpenQuant holds its research, documentation, benchmarks and releases to, and where to report a problem.
- [Performance](https://open-quant.github.io/openquant/project/performance/): [generated] Committed criterion and Python benchmark results, what each benchmark measures, and how CI gates regressions.
- [Runbook: bet sizing from predicted probabilities](https://open-quant.github.io/openquant/runbooks/bet-sizing-from-probabilities/): [authored] Does sizing meta-labeled bets by their predicted probability (AFML Snippets 10.1-10.3) beat a flat 0.5 filter, net of costs and deflated by every sizing choice tried? A pre-registered test of runbook 11's post hoc finding, on fresh SYNTHETIC planted-signal paths and no-signal controls.
- [Runbook: CPCV backtest with PSR and deflated Sharpe](https://open-quant.github.io/openquant/runbooks/cpcv-deflated-sharpe/): [authored] Best-of-74 strategy selection judged by the naive PSR, the deflated Sharpe ratio, a walk-forward path and CPCV path distributions, on SYNTHETIC data with a no-signal control and a planted signal.
- [Runbook: fracdiff features against returns in a classifier](https://open-quant.github.io/openquant/runbooks/fracdiff-features-classifier/): [authored] Do fractionally differentiated features beat plain returns and integer differences as inputs of a purged-CV classifier, net of costs and deflated by every feature set tried? Runbook 09's open question, on SYNTHETIC paths with a planted signal of known strength and no-signal controls.
- [Runbook: fracdiff, stationarity versus memory](https://open-quant.github.io/openquant/runbooks/fracdiff-stationarity-memory/): [authored] Sweeps the fractional differencing order d per series, picks the smallest d that passes ADF, and checks that rule against simulated series whose memory is known. Committed results are on SYNTHETIC data.
- [Runbook: HRP vs IVP and CLA out of sample](https://open-quant.github.io/openquant/runbooks/hrp-vs-ivp-cla-oos/): [authored] AFML §16.6's Monte Carlo, re-run with the openquant bindings: out-of-sample variance of HRP, inverse-variance and CLA minimum-variance portfolios on SYNTHETIC data.
- [Runbook: triple-barrier labeling and meta-labeling](https://open-quant.github.io/openquant/runbooks/meta-labeling-triple-barrier/): [authored] Does a meta-labeling model improve the precision, F1 and deflated Sharpe ratio of a primary model with a known, weak edge? CUSUM events, triple-barrier meta-labels, purged k-fold and a trial registry, on SYNTHETIC paths with a planted signal and a no-signal control.
- [Diagnose input errors](https://open-quant.github.io/openquant/setup/input-errors/): [authored] Identify validated input failures and assumptions that Python and Rust callers must check themselves.
- [Prerequisites](https://open-quant.github.io/openquant/setup/prerequisites/): [authored] Toolchain OpenQuant requires, how to install it on each platform, and why each version floor exists.
- [Python Core Workflow](https://open-quant.github.io/openquant/workflows/python-core-workflow/): [authored] A Python API tour of CSV ingestion and a separate synthetic pipeline, with Rust/Python conventions made explicit.
- [Notebook contract](https://open-quant.github.io/openquant/workflows/research-notebook-contract/): [authored] The sections, reproducibility footer and committed-output rules every notebook under notebooks/python/ follows, what `just notebooks-lint` checks, and how a runbook differs from an API tour.

## Repository sources

- [Source checkout](https://github.com/Open-Quant/openquant/tree/6335a56fa5f20abfb89d876f9eb90792eac1460b): Match the implementation and tests to this documentation revision.
- [Repository workflow](https://github.com/Open-Quant/openquant/blob/6335a56fa5f20abfb89d876f9eb90792eac1460b/AI-DLC.md): Read repository instructions before editing; use applicable instructions from your checkout.
- [Contribution workflow](https://github.com/Open-Quant/openquant/blob/6335a56fa5f20abfb89d876f9eb90792eac1460b/CONTRIBUTING.md): Setup, required checks and generated-source ownership.
- [Documentation ownership](https://github.com/Open-Quant/openquant/blob/6335a56fa5f20abfb89d876f9eb90792eac1460b/docs/documentation-guide.md): Canonical sources, documentation review and evidence requirements.