The reference for the public API of every header in include/EmrysFilter/. Start using the library from the usage guide and the programs in examples/; this reference is where to look up a class or a function they use.

Start with the modules page, which groups the headers by what they provide. One idea runs through all of them: a distribution is stored as its parameters, which a filter changes step by step, and evaluated through its density, which borrows what the parameters cached, so that a covariance is factorized once and evaluated many times. Parameters and densities, on the page of the Distributions group, explains it, and it is the one section to read before any other. The design decisions behind them are in docs/design/decisions.md, the ordered work in docs/design/roadmap.md, and the design reviews of the areas still under review in docs/guide/.

Every algorithm here is an implementation of one published in the public literature. Each cites its source where it is documented, and the Bibliography collects the entries. Where a concept is foundational rather than an algorithm, such as a distribution or a divergence, the documentation links to its Wikipedia article.