# TSGap documentation

TSGap is a Python library for composable time-series missingness simulation.
It separates missingness mechanisms (why values are missing) from temporal
patterns (how the missing values are arranged), making it possible to construct
reproducible benchmarking scenarios for imputation methods.

```{toctree}
:maxdepth: 2
:caption: User guide

installation
concepts
mechanisms
patterns
mathematical_details
benchmarking
api
```

```{toctree}
:maxdepth: 2
:caption: API reference

generated/tsgap
```