Hi there! I am running zavod locally, and I notice the exporters take a lot of time to run on my machine. Are all exporters required to run? Can yente work properly then if I exclude some of them?
No — yente only actually reads one output per dataset: entities.ftm.json, the FollowTheMoney entities file. That’s confirmed both in zavod’s own exporter code and in yente’s manifest docs, which explicitly set resource_name: entities.ftm.json when pointing at a catalog.
The other exporters in DEFAULT_EXPORTERS (nested targets JSON, names.txt, senzing.json, simple CSV, delta) exist for other consumers — Senzing entity-resolution import, a flat names list, spreadsheet-friendly CSV, incremental delta updates for bulk-data subscribers — not for yente’s own indexing. The statistics exporter always runs regardless of config (it’s not in that configurable set), but it’s cheap.
To skip the rest for local dev, add exports: [entities.ftm.json] to the dataset’s own YAML config — get_exporters() only falls back to the full default set when exports is empty, so this overrides it and zavod will build just the FtM file.
One thing to keep in mind: if you’re also relying on yente’s delta_url / incremental reindexing, you’d want to keep the delta exporter too — without it yente falls back to rebuilding from the full entities.ftm.json on every reindex instead of applying an incremental diff.