Semantria 5 builds on the Semantria 4 API. Most of the existing API functionality is unchanged, but there are a few backwards-incompatible changes to watch out for. Authentication handling is where most incompatible changes have occurred. Semantria has moved to a more standard authentication scheme that does not require any custom code to implement and adds flexibility for managing permissions and application life-cycle.
Semantria 4 has reached its end of life on April 30th 2022. After reading this guide, if you have any questions or need help implementing a migration, please contact [email protected]
- Ability to use languages and industry packs without creating a configuration by specifying a using parameter during document submission
- Taxonomy definitions may now include named entities, user entities, and auto categories.
- User groups and permissions allow more flexibility for managing your user base
- Management endpoints for assigning your custom machine learning models to configurations
- Interact with Semantria purely using REST calls, including using cURL from the command line
- All documents are now submitted in batches instead of having separate endpoints for batches and single documents. Submitting a single document is now a batch of one.
- The format of the configuration object has changed significantly.
- There is no longer a default configuration. You must explicitly specify a configuration id to modify a configuration or use it for document processing.
- Separate processing streams are now identified solely by job id.
There is less need for an SDK in Semantria 5 as all actions are easily accomplished with a straightforward set of REST calls. We do provide a Java SDK that includes a sample implementation of a client for processing documents and handling the returns.
You can now use a Swagger UI to interact with the API live here
Updated 16 days ago