A semantic search excels at understanding the user’s underlying intent when performing a search. It is robust against misspellings, synonyms and identifying the general topic of the query, whereas regular search just considers the exact wording.
TellusR comes with an integrated NLP-module that is capable of accommodating various languages. By default, TellusR comes with models that are well suited for most common search-related use cases. The models are pre-trained, so you don’t need to do any additional training efforts post-installation. The models are used generate embeddings for documents and for the query; the search finds top scoring documents where the score is determined by the vector-similarity of the query-vector to the document-vector(s).
If you wish to do fine-tuning, then this is technically supported, but you will need to inject your fine-tuned models into the nlp service that hosts them. Reach out to us if you need help setting this up.
If you want to enable the semantic search in TellusR, you will need to configure a semantic search index in the semantic indexing widget.
After creating a semantic index, all documents that are posted to TellusR will be automatically indexed according
to the fields that are selected for the index.
Assuming you have starting indexing your document base, or have added documents after creating the semantic index,
then the /tellusr/api/v1/{project}/query endpoint will return semantic hits if the semanticWeight is left untouched or if semanticWeight is non-zero.
Use semanticWeight=1 if you only want the search to return semantic hits.