Package ai.vectros.types
Class RagSearch
java.lang.Object
ai.vectros.types.RagSearch
-
Method Details
-
getMode
- Returns:
- Retrieval mode:
HYBRIDcombines keyword and semantic search,SEMANTICuses vector similarity only,TEXTuses keyword (BM25) matching only. Defaults toHYBRID.
-
getLimit
- Returns:
- Maximum number of results to retrieve before building the prompt. Defaults to 10; the maximum is 50.
-
getFilters
- Returns:
- Field filters to apply to your indexed content, as a map of field name to required value.
-
getUserId
- Returns:
- Restrict retrieval to content owned by this user id.
-
getOrgId
- Returns:
- Restrict retrieval to content owned by this organization id.
-
getClientId
- Returns:
- Restrict retrieval to content tagged with this client id.
-
getContentTypes
- Returns:
- Narrow retrieval to specific content types. When omitted, both documents and records are searched — records are valid grounding content too (for example, structured patient data or intake forms). Pass ["documents"] for documents only or ["records"] for records only. Each retrieved result carries a
sourceTypefield indicating which type it is.
-
getFolderId
- Returns:
- Restrict retrieval to content (documents or records) in this exact folder. Folders are a unified organizational primitive and may hold mixed content. Provide the UUID of a folder. To include a folder and all of its descendants instead, use
rootFolderId.
-
getRootFolderId
- Returns:
- Restrict retrieval to content under this subtree — the folder and all of its descendants. Provide the UUID of a top-level (root) folder. Applies uniformly to documents and records. Use
folderIdto scope to a single folder exactly.
-
getTypeName
- Returns:
- Restrict retrieval to records of a specific schema type. Applies only to records: setting it implicitly limits results to records unless
contentTypesalso includesdocuments, in which case documents are still returned unfiltered.
-
getCreatedAfter
- Returns:
- Restrict retrieval to content created at or after this ISO-8601 UTC timestamp. Useful for incremental retrieval (for example, answering based only on content ingested today). Pair with
createdBeforeto define a time window.
-
getCreatedBefore
- Returns:
- Restrict retrieval to content created at or before this ISO-8601 UTC timestamp. Pair with
createdAfterto define a time window, or use alone to scope retrieval to content older than a cutoff.
-
getRequireComplete
- Returns:
- Require complete retrieval. When true, if a search backend is unavailable the request returns HTTP 503 before any answer is streamed, rather than grounding the answer on partial results. Defaults to false, in which case an outage degrades to the surviving search engine and the
search_resultsevent reportsdegraded: true. Set this to true when an answer grounded on incomplete retrieval is unacceptable.
-
equals
-
getAdditionalProperties
-
hashCode
public int hashCode() -
toString
-
builder
-