REST API

Everything in the web app is available over a JSON REST API, so you can refresh test data from a CI/CD pipeline.

Authenticate

curl -s -X POST https://<host>/auth/signin \
  -H 'Content-Type: application/json' \
  -d '{"username":"ci.bot","password":"…"}'
# → {"access_token":"eyJ…","token_type":"bearer"}

Send the token on every call: Authorization: Bearer <access_token>. Use a dedicated Member account for automation, and save the passwords of the connections it uses.

Main endpoints

Area	Endpoint
Connections	/api/v1/connections
Virtual data sets & prune	/api/v1/vds · POST /api/v1/vds/shrink to run · GET /api/v1/vds/shrink/{id}/status · /result · /artifacts
Bloom plans	POST /api/v1/bloom/plans {vds_id, learn, learn_columns} (from a VDS) · POST /api/v1/bloom/plans/from-sql (DDL or template) · GET /api/v1/bloom/templates · POST /api/v1/bloom/plans/{id}/preflight
Bloom runs	POST /api/v1/bloom/plans/{id}/runs {seed, scale, formats} · GET /api/v1/bloom/runs/{id}/status · /runs/{id}/artifacts
Load	POST /api/v1/bloom/runs/{id}/load/preflight to check · POST /api/v1/bloom/runs/{id}/load to load

Long jobs return a task ID. Their …/status endpoints stream progress as server-sent events.

Load options

Field	Meaning
connection_id	The target. It must be approved as a test database, or the call returns 403 Approve this connection as a test database before loading into it.
schema_name	Target schema; the connection's default if empty
mode	append (default) or replace
create_missing	Create missing tables
keep_schemas	For plans read from several schemas: load each table into its own schema
create_schema	Create the target schema if it's missing
recreate_tables	Plan tables to drop and create again before loading
confirm_database	The target database's exact name. Required for replace and recreate_tables.

learn_columns lists the columns whose value mixes a plan may learn. Leave it empty to learn table sizes only.

Example: refresh test data in a pipeline

TOKEN=$(curl -s -X POST https://<host>/auth/signin -H 'Content-Type: application/json' \
  -d '{"username":"ci.bot","password":"'"$DF_PASSWORD"'"}' | jq -r .access_token)

# Generate a fresh, repeatable data set from an existing plan
curl -s -X POST https://<host>/api/v1/bloom/plans/$PLAN_ID/runs \
  -H "Authorization: Bearer $TOKEN" -H 'Content-Type: application/json' \
  -d '{"seed": 42, "scale": 1, "formats": ["csv"]}'

Responses to handle

Status	Meaning
402	The licence is missing or expired, so DataFactory is read-only. See Licence.
403	Not allowed, e.g. the load target isn't an approved test database.
429	Three runs are already queued or running for this account. Wait for one to finish.

Interactive reference

The full, interactive API reference is at https://<host>/api/docs when your admin sets API_DOCS=true (it's off by default).