Skip to content

[pipelines] Core step library — map / filter / dedupe·upsert / paginate / reachability / geocode #96

Description

@serge-ivo

Child of #94.

Declarative pipelines (#94) compose reusable steps. Ship a first-party step catalog (each a registry tool, #85/#87) so the common data-shaping ops are config, not code:

  • map/transform — rename/derive fields, extract nested (e.g. addressComponents → {city,suburb,state,country}), format.
  • filter — keep/drop by predicate over fields.
  • dedupe / upsert — by key field; skip-if-seen or update-in-place (respects collection unique constraints, MCP: instance-scoped collection read/write (insert_record/query_records/list_collections) #91).
  • paginate / fan-out — iterate a source over pages or a generated set (e.g. grid cells around a centre), with a concurrency cap.
  • http_reachable — GET a URL (browser UA, timeout, retry) → {ok, code}; the "is this site alive" probe.
  • geocode — city/address → {lat,lng,country,state,locality} (via the generic HTTP connector + a maps provider).

Acceptance

  • The lead-finder's transform/filter/dedupe/reachability/geocode steps are all expressible from this catalog — no custom code.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or request

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions