Fixed bug in attribute_type_id detection of attribute_type_id namespaces

v3.5.9

Enhanced functionality to support SRI Testing

• implemented get_inverse_predicate() and is_symmetric() methods
• moved case node/edge validation predicate methods from SRI_Testing into reasoner-validator

input edge Biolink validation bug fix

code still expected 'subject' and 'object' rather than 'subject_id' and 'object_id' in test data edges. Now recognizes either tag value

trapi_response_validator.py enhancement

• ars_uuid_result_test_runner.py renamed to trapi_response_validator.py and generalized to accept a local TRAPI Response or possibly, an explicit endpoint URL and the TRAPI Request JSON with which to query, whose TRAPI Response is then validated.

Use of a local file path to a TRAPI schema implemented, with code validation using unit tests

v3.5.4

Fixed query graph validation that was deleting node identifiers in the input query graphs

v3.5.3

Fixed unit tests for code tweaked in 3.5.1 patch

v3.5.2

error.trapi.validation message pruning for UI readability

error.trapi.validation 'reason' message text shortened by pruning to use only the start and end of the JSON Schema generated ValidationError message, to a maximum of 160 character

Rehabilitate BMT dependency and usage

• BMT and related dependencies rehabilitated and simplified? Some BMT functionality removed along the way (hence Minor rather than patch release); removed direct KGX dependency

Validate 'null' TRAPI attributes; RetrievalSource infores; other tweaks

• detect all forms of null TRAPI attributes (i.e. strings like "n/a", "none" and "null")
• fixed some technical debt in other pieces of code along the way:
  • Split out the test for an empty message body from the empty response and missing key validation error
  • Query Graphs are allowed to have abstract qualifiers, so the test should not generate an error message
  • 1.4.0-beta4 is the latest TRAPI release
• remove explicit BMT package dependency (rather, expect to to be pulled in by KGX)
• Added deeper validation of RetrievalSource (i.e. checking the resource_id Infores identifiers) plus associated unit tests