{"categories":[{"label":"Artificial Intelligence","url":"https://skillfed.io/packages/category/scientific-engineering-artificial-intelligence/9"},{"label":"Application Frameworks","url":"https://skillfed.io/packages/category/software-development-libraries-application-frameworks/5"}],"enrichment":{"capability":"PyGrinder introduces missing values into datasets using multiple patterns (MCAR, MAR, MNAR, RDO, sequential, block) to simulate incomplete data for model evaluation and testing.","skillfed_tags":["time-series","data-simulation","missing-data"],"use_cases":["Evaluate time-series imputation models by creating datasets with known missingness patterns to test reconstruction accuracy.","Generate partially-observed datasets for training robust models that must handle incomplete real-world observations.","Benchmark data mining algorithms on incomplete data without needing to manually corrupt or collect partial datasets.","Simulate different missing-data mechanisms (MCAR vs. MNAR) to study how model performance varies with missingness type.","Create test suites for data preprocessing pipelines that must handle missing values in production."],"what_it_does":"PyGrinder is a data corruption toolkit designed to inject missing values into datasets using multiple well-defined patterns. It was originally part of PyPOTS (a time-series data mining framework) and separated to decouple missingness generation from learning algorithms. The package supports MCAR (Missing Completely At Random), MAR (Missing At Random), MNAR (Missing Not At Random with variants), RDO (Random Data Observation), and structured patterns like sequential and block missing values.\n\nThe typical use case is preparing synthetic incomplete datasets for evaluating how machine learning models handle partial observations or for testing data reconstruction algorithms. You pass a numpy array (often time-series data with shape like samples \u00d7 timesteps \u00d7 features) and a pattern function with parameters controlling the missing rate or structure, and receive back a dataset with missing values injected. The package also provides a utility to calculate the resulting missing rate.","worth_installing":"Yes. PyGrinder is actively maintained, has no known vulnerabilities, uses a permissive license, and solves a specific and common problem in machine learning: creating incomplete datasets for model evaluation. The low install friction and production-stable status make it a straightforward choice if you need to simulate missing data patterns. Install it if you work with time-series data or incomplete observations in model development or research."},"id":"pygrinder","links":{"html":"https://skillfed.io/packages/pygrinder","md":"https://skillfed.io/packages/pygrinder.md","pypi":"https://pypi.org/project/pygrinder/"},"maintenance":{"status":"active"},"meta":{"latest_release":"2025-02-03","license_spdx":null,"license_treatment":"permissive","name":"pygrinder","python_support":"supports_current","summary":"A Python toolkit for introducing missing values into datasets"},"popularity":{"monthly_downloads":122792,"position":11933,"tier":"top_15000"},"security":{"n_vulnerabilities":0},"version":"0.7"}