Caching for download enpoint

[j92]

We should add caching for downloaded files from AWS as we pay for each download.

Should we cache all files?

Sounds like a bad idea, as the cache will be overloaded.
We could decide per file size if the file should be cached.
For getting the filesize we have two options:

1. Store the size of the file somewhere on uploading.
2. Do a HEAD request to AWS.

Should users be able to bypass the cache?

For the main API I don't think that will be needed as we need caching primarily to avoid unnecessary spendings.

Adapter

1. InMemory Adapter
2. Bring your own (Redis) adapter?