Perform the Knowledge Base Search

Specify your knowledge base in the KnowledgeBaseSearch integration and run the integration to perform your search.

Prerequisites:

• Specify the Knowledge Base in the KnowledgeBaseSearch Integration

1. In your project, in the Integrations section, next to the KnowledgeBaseSearch integration, click Actions , then select Run.
   The Configure and run page is displayed.
2. In Request, select Body and specify your search.

   The type of search is based on which parameters you specify:

   • Text search: simple text search for an exact keyword match. You could use this for finding order numbers for example.

     • SemanticSearchQuery: Leave this parameter blank
     • TextSearchQuery: Specify what to search for
     • maxResults: Specify the maximum number of results you want returned
     • augmentResponseWithSurroundingTextChunksCount: Specify how many additional chunks to retrieve as context. This parameter is useful when using AI agents so that the AI agent can make sense of a broader context, not just the matched text.
     Example:

     {
       "SemanticSearchQuery": "",
       "maxResults": 1,
       "TextSearchQuery": "Mr. Smith",
       "augmentResponseWithSurroundingTextChunksCount": 1
     }

   • Semantic search: searches based on meaning, not exact words.
     • SemanticSearchQuery: Specify the value to search for.
     • TextSearchQuery: Leave this parameter blank
     • maxResults: Specify the maximum number of results you want returned
     • augmentResponseWithSurroundingTextChunksCount: Specify how many additional chunks to retrieve as context. This parameter is useful when using AI agents so that the AI agent can make sense of a broader context, not just the matched text.
     Example:

     {
       "SemanticSearchQuery": "Story about 3 sundays in a single week",
       "maxResults": 1,
       "TextSearchQuery": "",
       "augmentResponseWithSurroundingTextChunksCount": 1
     }

   • Hybrid search: searches both exact match and meaning.
     • SemanticSearchQuery: Specify the value to search for.
     • TextSearchQuery: Specify the value to search for.
     • maxResults: Specify the maximum number of results you want returned
     • augmentResponseWithSurroundingTextChunksCount: Specify how many additional chunks to retrieve as context. This parameter is useful when using AI agents so that the AI agent can make sense of a broader context, not just the matched text.
     Example:

     {
       "SemanticSearchQuery": "Story about 3 sundays in a single week",
       "maxResults": 1,
       "TextSearchQuery": "Mr. Smith",
       "augmentResponseWithSurroundingTextChunksCount": 1
     }

3. Click Run.

   The Activity Stream panel opens on the right to show the progress of the run and the Response provides the results based on your query.

   Tip:

   In the Response body metadata, document_link provides a link to the document in the OCI Object Storage bucket.