You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add has_node_name function to AnnotationStorage that can be more efficient
than get_node_id_name.
Changed
Changed API to use new types NodeAnnotationStorage and EdgeAnnotationStorage instead of AnnoStorageImpl<NodeID> or AnnoStorageImpl<NodeID>. (backward incompatible change in the Rust API)
get_node_id_from_name is now a function of the AnnotationStorage instead
of the Graph. This allows for more specific and efficient implementations
based on the type of annotation storage.
Improved performance of the Graph::apply_update function.
Use jemalloc memory allocator for webservice and CLI.
Removed
Remove all heap size estimation code. This also means that information about
heap consumption of a single corpus has been removed, like the fields of the graphannis::corpusstorage::LoadStatus enum.
Remove EvictionStrategy::MaximumBytes for DiskMap.
Fixed
Polling when importing a web corpus through the webservice could fail because
the background job list was not shared between the web server threads.