Pipeline and loads

The pipeline chains all the steps to update the knowledge graph:

Extraction → Transformations → Mapping → Load

Run a load

1. Choose what should run: sources, extractions, transformations
2. Run the pipeline

Progress is displayed live.

The load keeps the data history: when a value changes (e.g. a contract’s price), the previous version is kept with its validity period.

Run history

Find the details of each past run (extractions and pipelines alike), with any errors.

Once the load is complete, the data is available in Explore and in AI Assistant.