Metadata Editor

Appearance

Metadata reviewer ​

The metadata reviewer is an AI-assisted review that helps you improve project metadata. It sends your metadata to the configured reviewer service, which returns findings — typos, missing fields, inconsistencies, unclear wording, and similar — often with suggested text you can apply to the project.

Findings appear as issues on the project Issues page. You work through them the same way as other issues: comment, accept, reject, dismiss, or apply a suggestion.

The metadata reviewer is built on ai4data. For pipeline design, categories, and severity scoring, see the Metadata Reviewer overview. For installation, API usage, and troubleshooting the reviewer package itself, see the Metadata Reviewer User Manual.

How it works ​

  1. Open the project Issues page and click Assess metadata.
  2. The reviewer analyzes the metadata (this runs in the background; you can leave the page).
  3. New issues appear on the Open tab when the run completes.
  4. Review and resolve items using the Project issues workflow.

Your site must have the metadata reviewer installed and enabled. See Install the metadata reviewer.

When your site has it enabled ​

The Assess metadata button appears only when a site administrator has turned on Metadata assessment in site configurations and the required services are running.

Before starting, the application checks:

  • The assessment service (FastAPI with the metadata reviewer add-on) is available.
  • The background worker is running (review runs as a background job).

If the button is missing, the feature is disabled for your site. If it is disabled with a tooltip, contact your administrator.

Running a review ​

  1. Open the project Issues page.
  2. Click Assess metadata.
  3. Confirm in the dialog. It explains that metadata is sent to the quality assessment service and that new issues will be added to this list.
  4. The job runs in the background. You may leave the page and return later; open Assessment status from the button while a run is in progress.

When the review completes, refresh the issue list. New items are categorized (for example typo/wording, inconsistency, missing data) and may include suggested metadata to apply.

Monitor jobs on Managing background jobs (metadata_assessment_result).

Privacy ​

The metadata reviewer sends project metadata to the configured LLM provider (Azure OpenAI, OpenAI, Ollama, Anthropic, or similar). Review your organization’s data-handling and AI policies before using it on sensitive content.