Problem
Developing and testing the habitat pipeline (frs_habitat) requires a PostgreSQL instance with fwapg data. Currently depends on the shared remote db_newgraph over SSH tunnel. This limits connection count, adds latency, and creates contention with the weekly bcfishpass ng-prod build.
Proposed Solution
Ship a lightweight Docker setup in the package for local development. A docker-compose.yml that spins up PostGIS with FWA data for a few test WSGs (ADMS, BULK).
Contents
- docker-compose.yml (postgis/postgis image)
- FWA data loader script (base streams, channel width, falls, WSG boundaries)
- PG tuning for fresh workload (work_mem, max_parallel_workers_per_gather)
- ltree indexes on wscode_ltree, localcode_ltree
Scope
Development and CI only. A few WSGs, fast to spin up. Province-wide production infrastructure lives in awshak (NewGraphEnvironment/awshak).
Relates to NewGraphEnvironment/awshak#TBD
Problem
Developing and testing the habitat pipeline (frs_habitat) requires a PostgreSQL instance with fwapg data. Currently depends on the shared remote db_newgraph over SSH tunnel. This limits connection count, adds latency, and creates contention with the weekly bcfishpass ng-prod build.
Proposed Solution
Ship a lightweight Docker setup in the package for local development. A docker-compose.yml that spins up PostGIS with FWA data for a few test WSGs (ADMS, BULK).
Contents
Scope
Development and CI only. A few WSGs, fast to spin up. Province-wide production infrastructure lives in awshak (NewGraphEnvironment/awshak).
Relates to NewGraphEnvironment/awshak#TBD