guillaumecherel / cached

Write data flows that cache values to disk. Run independant computations in parallel and only recompute what needs to be.

cached

Cache values to disk.

The module Data.Cached lets you cache values to disk to avoid re-running (potentially long) computations between consecutive executions of your program. Cached values are recomputed only when needed, i.e. when other cached values on which they depend change. Independent computations are run in parallel. It offers convenient fonctions for caching to text files, but caching and uncaching using arbitrary IO actions is also possible.

The module was motivated by writing scientific data flows, simulation experiments or data science scripts. Those often involve long computations and create "flows" where the output of some computation are the inputs of others, until final results are produced (values, figures, statistical tests, etc.).

See the module Data.Cached documentation:

• On hackage: https://hackage.haskell.org/package/cached/docs/Data-Cached.html
• In the source: ./src/Data/Cached.hs

cached's Issues

Revision 1 of 0.1.0.0 for GHC 9.8

Hi, I revised cached-0.1.0.0's bounds as they stopped me building it with GHC 9.8: https://hackage.haskell.org/package/cached-0.1.0.0/revisions/
The package builds fine with GHC 9.8 after the revision.

I notice some tests fail but this isn't connected to the GHC version.

What is the status of cached? Should I use it?
I was looking for something that could cache values painlessly between executions of a haskell script.
Maybe what is missing in this package is generic (binary) serialization. It seems to prefer Read/Show which isn't a good method for serialization in general.