EWB's Classifier API

The Classification Service is a tool designed to categorise text content based on a predefined hierarchy of language models. This service leverages the clf-inference-intelcomp library, enabling users to efficiently classify text data. Below, you'll find detailed information about the service's endpoints and how to use them effectively.

Please refer to the official documentation for a comprehensive guide on utilising the library's capabilities.

Note on Model Loading: It's worth noting that the classification process may take longer if the language models have not been loaded into the cache. To optimise performance, we recommend caching the required models in advance.

Endpoints

Endpoint	Description
/classification/classify	Classifies text using models trained within a specified taxonomy.
/classification/list_avail_taxonomies	Retrieves a list of available taxonomies.
/classification/cache_models	Caches language models in memory to expedite inference times.
/classification/list_models	Lists the models associated with each taxonomy, with hierarchical indentation for clarity.
/classification/list_classes	Provides a comprehensive list of possible classification outputs by mapping class codes to names.

Examples of classification (one per each available taxonomy)

• Classification according to FOS

• Classification according to IPC

• Classification according to NACE2