Release v1.6.5 no-cypher

v1.6.5 with the /cypher endpoint disabled

Release v1.6.5

This release includes a fix for a bug where removing the TRAPI version prefix from api endpoints caused some attributes to be missing from the open api spec.

Release v1.6.4

This release includes new versions of reasoner-transpiler and subsequently bmt, and fixes a bug related to invalid subclass_depth configurations.

Release v1.6.3

Changed /simple_spec so that it uses the meta kg, and not a cypher call, to generate the entire simple_spec when no nodes are specified. It also caches the result so that it doesn't have to generate it again.

Release v1.6.2

Removed /overlay endpoint and improved some descriptions for openapi

Release v1.6.1

Refactoring the consolidation of the /1.4/ and / sub-applications, cleaning up redundancies

Release v1.6.0

TRAPI 1.5 implementation

• removed /1.4/ part of path from the endpoints
• changes to support set_interpretation instead of is_set (with an error for a value of "MANY")
• some changes to attribute handling, now use the biolink model to attempt to determine attribute types and values but also look at the custom mappings
• use bmt to determine whether a property is a qualifier or not

Implementation of new parameters for the /query endpoint:

• profile (boolean) will cause the endpoint to return SpeedScope profiling results instead of an answer
• validate (boolean) will cause Plater to validate the result using the pydantic model

Implementation of new optional config/env vars:

• RESULTS_LIMIT can apply a limit on the number of results returned by /query queries
• SUBCLASS_DEPTH will specify the number of hops allowed for subclass inference

epc-testing

This includes a significant refactor of handling primary and aggregator knowledge sources.

Now we expect from the graph:

• primary and aggregator knowledge sources no longer have biolink: prefixes
• aggregator knowledge sources are ordered lists with the most upstream source first
• optionally includes multiple aggregator knowledge source attributes (ex. aggregator_knowledge_source, aggregator_knowledge_source_2, aggregator_knowledge_source_3)

There also shouldn't be null upstream ids for primary knowledge sources anymore.

Release v1.5.1

Contains updates related to the biolink model and biolink model toolkit:

• new transpiler version which uses the real biolink model toolkit instead of bmt-lite
• the env var BL_VERSION now actually enforces that version of the biolink model for the bmt instances in plater and the transpiler
• qualifiers or qualifier/value pairs that are not valid according to the biolink model throw 400s now

Release v1.5.0

Big performance upgrade for TRAPI /query endpoint