Revectorize¶

Re-vectorize the source segments. After the request is submitted successfully, you still need to read the source or process the task status to confirm the actual result.

POST https://moi.matrixorigin.cn/newmoi/semantic-models/{model_id}/sources/{source_row_id}/segments/re-embedding

Preparation before calling¶

First query the document details, and obtain the current segment version ID and index version as the baseline. Prepare a personal access token and target workspace ID that has access to the target workspace.

The example below uses:

$AI_STUDIO_API_KEY: The actual personal access token, passed through the X-API-Key Header.
$WORKSPACE_ID: Target workspace ID, passed through X-Workspace-ID Header.
$MODEL_ID: Knowledge Base ID.
$SOURCE_ROW_ID: Source record ID.

Request example¶

curl -X POST "https://moi.matrixorigin.cn/newmoi/semantic-models/$MODEL_ID/sources/$SOURCE_ROW_ID/segments/re-embedding" \
  -H "X-API-Key: $AI_STUDIO_API_KEY" \
  -H "X-Workspace-ID: $WORKSPACE_ID" \
  -H 'Content-Type: application/json' \
  -d '{
    "base_segment_version_id": "<SEGMENT_VERSION_ID>",
    "base_index_version": <INDEX_VERSION>
  }'

Path parameters¶

Parameters	Type	Description
`model_id`	integer	Knowledge base ID.
`source_row_id`	string	Source record ID.

Request body¶

The current version baseline of the commit source.

Field	Type	Required	Description
`base_segment_version_id`	string	Yes	The current segment version ID read before the call.
`base_index_version`	integer	Yes	The current index version read before the call.

Successful response¶

Returns 200 and an updated source document snapshot on success; the final result of the vectorization should still be confirmed by source or task status.

{
  "code": "OK",
  "msg": "OK",
  "data": {
    "document": {
      "source": {
        "row_id": "src_01",
        "model_id": 401,
        "ingest_status": "processing"
      },
      "segment_status": {
        "available": true,
        "total": 2
      },
      "current_segment_version_id": "ver_03",
      "current_index_version": 3,
      "segment_versions": [
        {
          "version_id": "ver_03",
          "current": true,
          "index_version": 3,
          "chunk_count": 2
        }
      ],
      "segments": [
        {
          "segment_id": "seg_01",
          "enabled": true
        }
      ]
    }
  }
}

The response fields are as follows.

Field	Type	Description
`code`	string	`OK` on success.
`msg`	string	`OK` on success.
`data.document`	object	Snapshot of the updated source document.
`data.document.source.ingest_status`	string	Current processing status; does not replace processing completion with HTTP success.
`data.document.segment_status`	object	Available status and number of segments.
`data.document.current_segment_version_id`	string	Current segment version.
`data.document.current_index_version`	integer	Current index version.
`data.document.segment_versions`	object[]	Currently available segment versions.
`data.document.segments`	object[]	Segmentation in the current version.

[] after a type means an array. For example, object[] is an array of objects.

Error response¶

{
  "code": "ErrConflict",
  "msg": "segment version conflict",
  "data": null
}

Common HTTP errors¶

HTTP status code	error code	Common causes	Recommended actions
`400`	`ErrParamInvalid`	Invalid path ID or staging version baseline.	Use current baseline after rereading document.
`401`	`ErrUnauthorized`	The API Key is invalid or has expired.	Check API Key.
`403`	`ErrForbidden`	The caller does not have permission to update the source.	Check workspace and object authorization.
`404`	`ErrNotFound`	The knowledge base or source does not exist or is not visible to the current caller.	Reconfirm the path ID.
`409`	`ErrConflict`	The current segment version has changed.	Retry after reading the latest document.
`500`	`ErrServer`	The service failed to perform revectorization.	Keep the desensitized response information and try again.

Follow-up operations¶

Vectorization is performed asynchronously. Confirm the final result of vectorization through Query data processing tasks or Query document details.

Last updated on Sep 3, 2026