If we're caching things to files, we'll need to ensure that as we make changes to finch-tensor-lite, the cache updates with the changes. https://github.com/finch-tensor/finch-tensor-lite/blob/main/src/finchlite/util/cache.py has the implementation of the file cache. You'll need to add a timestamp file to the cache. On startup of finch, check all the files in finch codebase with a loop like this: https://shezi.de/posts/2022-09-11-mtimes.html, and figure out when it was last modified. If finch has been modified since we started caching, you should empty the cache.
If we're caching things to files, we'll need to ensure that as we make changes to finch-tensor-lite, the cache updates with the changes. https://github.com/finch-tensor/finch-tensor-lite/blob/main/src/finchlite/util/cache.py has the implementation of the file cache. You'll need to add a timestamp file to the cache. On startup of finch, check all the files in finch codebase with a loop like this: https://shezi.de/posts/2022-09-11-mtimes.html, and figure out when it was last modified. If finch has been modified since we started caching, you should empty the cache.