Caching of response data is implemented via CacheControl, a library that extends requests. To enable a flexible usage of the caching mechanism, the use of cached data is steered through a cache policy, which can be specified either for the whole scraper or for a specific request.

The following policies are supported:

  • http will perform response caching and validation according to HTTP semantic, i.e. in the way that a browser would do it. This requires the server to set accurate cache control headers - which many applications are too stupid to do.
  • none will disable caching entirely and always revert to the server for up-to-date information.
  • force will always use the cached data and not check with the server for updated pages. This is useful in debug mode, but dangerous when used in production.

Per-request cache settings

While caching will usually be configured on a scraper-wide basis, it can also be set for individual (GET) requests by passing a cache argument set to one of the policy names:

import scrapekit
scraper = scrapekit.Scraper('example')

# No caching:
scraper.get('', cache='none')

# Cache according to HTTP semantics:
scraper.get('', cache='http')

# Force re-use of data, even if it is stale:
scraper.get('', cache='force')