Data Leakage Python Packages

unmasker

Detect hidden, residual, and machine-readable content in documents, images, and files.

sail-leakage

SAIL: detect state-action information leakage in offline reinforcement learning for sepsis — construction, reconstruction, temporal overlap, timing and persistence checks.

leakprobe

Find features that read data they were never supposed to see. Temporal leakage detection by metamorphic testing.

vigia-nids

Quality control for the datasets behind network intrusion detection models: leakage, duplicates, label noise, poisoning and drift.

targetleak

Find the data leakage that made your test score a lie - which column, why, and how to fix it.

flowstate-asof-core

High-throughput, Rust-accelerated temporal join engine for quantitative finance ML pipelines. Engineered via PyO3 to eliminate look-ahead bias and accelerate point-in-time feature generation on massive time-series market datasets

anthale

The official Python library for the anthale API

leakhound-ml

Find the data leakage inflating your ML score — duplicates, temporal, target, group, adversarial, and homology (protein/DNA) leaks — and measure the damage. One command, CI-ready.

heterosplit

Leakage-safe entity-disjoint splits and audits for heterogeneous link-prediction datasets (PyTorch Geometric-compatible).

safesplit

Leakage-safe train/test splits — group-aware, time-aware, sequence-aware (homology), de-duplicated. The companion that prevents what LeakHound detects.

gp3mlpy

Governance-first, leakage-resistant predictive modelling for Gazepoint research workflows in Python.

paneldx

Validate longitudinal dataset as what it claims to be. Entity-key verification, leakage detection, cumulative-counter traps and naive baselines, as a library, a CLI and a CI gate.

flowstate-asof

High-throughput, Rust-accelerated temporal join engine for quantitative finance ML pipelines. Engineered via PyO3 to eliminate look-ahead bias and accelerate point-in-time feature generation on massive time-series market datasets

leakguard-mcp

Squawk for backtests — static analyzer flagging lookahead bias & data leakage in ML/trading code before the backtest runs. Local stdio MCP.

dataleaks

Detect, explain, and prevent data leakage across machine learning datasets and workflows.

chemdatacheck

Quality checks for molecular datasets: integrity, duplicates, split leakage, ML readiness, and chemical space.

leakcheck-ml

Catch train/test data leakage before it ruins your model — a linter for ML datasets

pit-release-gate

Screen quant/ML feature pipelines for look-ahead bias from late-arriving data: measure each signal's leakage susceptibility and gate release until the cross-section is safe. Python, MIT.

cvflow

Audit YOLO/COCO datasets for broken images, bad annotations, duplicates, and split leakage.

pydextra

Data functions that explain themselves: 63 EDA/cleaning/feature-engineering/modeling helpers (+5 scikit-learn wrappers) that print a one-line Decision: and return replayable, leakage-safe params. pip install pydextra · import dextra as dx

medlint

Lightweight, evidence-first split-integrity audits for medical imaging datasets

wald-lint

Statistical-integrity linter for data-analysis notebooks. Code linters check style; wald checks whether the analysis is a lie.

alphaagent-quant

Multi-agent stock & crypto screening with leakage-aware (point-in-time) LLM evaluation and an honest, mechanical backtest.

safefeat

Leakage-safe, point-in-time feature engineering for ML — build features from event logs without future data leakage.