Indexing

Configure content indexing.

Content Indexing

Graphlit will analyze all source content, including documents, audio or image files, web pages, or Slack messages. Metadata will be indexed into JSON and stored internally with the ingested content.

Metadata indexing is handled by built-in content parsers, but we do support some optional indexing functionality which can be configured by the content workflow.

If you want to analyze the ingested text for the written language, you can enable the indexing stage of the workflow, and create a connector with the type AZURE_AI_LANGUAGE.

This will use the Azure AI Language service to read through the extracted text and identify one or more languages in the text. Thelist of ISO codes for the languages will be stored in the content metadata.

Using Azure AI Language for language analysis will incur partial credit usage per-page.

Mutation:

mutation CreateWorkflow($workflow: WorkflowInput!) {
  createWorkflow(workflow: $workflow) {
    id
    name
    state
    indexing {
      jobs {
        connector {
          type
          contentType
          fileType
        }
      }
    }
  }
}

Variables:

{
  "workflow": {
    "indexing": {
      "jobs": [
        {
          "connector": {
            "type": "AZURE_AI_LANGUAGE"
          }
        }
      ]
    },
    "name": "Workflow"
  }
}

Response:

{
  "indexing": {
    "jobs": [
      {
        "connector": {
          "type": "AZURE_AI_LANGUAGE"
        }
      }
    ]
  },
  "id": "4ae85292-6579-49af-9f7f-f96fa0392b57",
  "name": "Workflow",
  "state": "ENABLED"
}