pyramid_caching 0.2.0

TODO:

• DOCUMENTATION.

• Introspectables

• Content negociation: we should add content type in cache key.

ext.sqlalchemy: Increment new and deleted entities and table identities.

Simple dependency resolution for location-aware resources.

ext.royal: New extension for caching pyramid_royal resources.

Allow key identity generators to define resources that don’t need to be invalidated.

Breaking changes: cache_factory keyword arguments depends_on is a list of callables which receiving request as positional argument. Callables return the dependency to be identified by cache manager.

Add arguments predicates to cache_factory which permits to add predicates to cache key used for the view (useful for query strings).

Switch back to d2to1.

First draft of interfaces

Add a highly not efficient implementation of a key versioner

Add a model versioner which depends on key versioner and model identity inspector

Add an example application with dummy sqla model to run tests against

Add an sqla specific extention with hook on session events to get model modifications.

Add and remove cache_basic decorator to cache basic function with primitive typed argument. Let’s focus at view level. We’ll do basic decoraor later.

Add cache_factory to construct a decorator to pass to config.add_view method.

Add a cache manager which plug all components.

Add serializers to adapt objects to store on cache. Serializer objects follow a standard loads/dumps interface.

Add Redis client for caching and versioning

Implement master-version in redis version-store

The cache manager emits CacheHit and CacheMiss events. These events can be forwarded to a statistics aggregator by using the pyramid_metrics extension.

Sets the reponse ETag and custom X-View-Cache HTTP headers based on the versioned key of the resource.