Think about cache

[leplatrem]

• Last-modified header of source url ?

[AntoineCezar]

using las-modified header will add some complexity.

• does url has already been fetched ?
• does the fetched version has last-modified ?
• HEAD before GET to know last-modified
• save file and save last-modified
• find the corresponding converted file if exists

1srt method: use a database 2nd method: use url hashes + metadata file

[leplatrem]

A simple key-value store seems fine ? Like a JSON or a pickledb, since pyramid has no file-based cache backend of its own.