Script

The script processor runs a Tengo script on every record. The script gets full read-write access to the entity, so you can rename fields, rewrite URNs, build descriptions, or enrich properties with data fetched over HTTP.

For a walkthrough with real examples, see the scripting guide.

Usage

processors:
  - name: script
    config:
      engine: tengo
      script: |
        entity.name = entity.name + " (processed)"

Configuration

The script environment

• The script sees one global, entity, as a map with the fields urn, type, name, description, source, properties, create_time, and update_time. All fields can be read and written.
• properties can be undefined when the extractor set none — guard with entity.properties || {}.
• Setting a key outside the entity fields fails the run.
• Edges are not exposed. They pass through the processor unchanged.
• Scripts cannot drop a record; every record continues to the sinks.
• Importable modules: math, text, times, rand, fmt, json, base64, hex, enum, plus a built-in http module with http.get(url) and http.get(url, headers) (GET only, 5-second timeout, returns {code, body}).

Error behavior

The script is compiled when the recipe starts; a compile error fails the run before extraction. A runtime error in the script fails the whole run — records are not skipped, and processors are not retried.