We migrated the NL Interpreter API to PHP8.
New feature: We added a new endpoint to the NL Interpreter API. The endpoint URL is GET /v1/words?lexical_unit_id=<id>
. It can be used to retrieve all the words assigned to a specific lexical unit.
Fixed: The NL Interpreter API was not processing some special characters in its parameters. This is now fixed.
The NL Interpreter API now allows developers to configure the semantic contraction of lexical units through a parameter called contraction_level
. This means that they can use the parameter to have the API contract texts like "I want to stop my subscription" to "Cancel subscription", and decide to keep both the original and the contracted version, or only the contracted version.
We added the lexical_unit_id
property for each lexical_unit returned by the /interpretations
endpoint.
The Automatic Spell Correction can now be disabled by setting the query parameter spell_correction
to false. If set, the query is processed applying only the Manual Spell Correction.
The GET /v1/interpretations
endpoint now returns the identified word and the lexical unit semantic weight for each interpreted lexical unit within the response keys identified_word
and semantic_category_weight
respectively.