Migration Guide
Breaking changes and migration steps between releases. Read the section for the version you are upgrading to before updating your scripts and existing jobs.
v0.9 → v0.10
v0.10 has two breaking changes. Both require action; neither alters your FHIR data.
1. aether.yaml is now a required positional argument
The --config flag has been removed, and Aether no longer discovers a config file implicitly. v0.9 searched --config → ./aether.yaml → ~/.config/aether/aether.yaml. v0.10 requires you to pass the config path explicitly, as the first positional argument of every command that needs configuration.
Before (v0.9):
# Relied on the --config flag or implicit ./aether.yaml discovery
aether pipeline start your-crtdl.json
aether pipeline status <job-id>
aether pipeline continue <job-id>
aether job list
# or with the explicit flag
aether --config /path/aether.yaml pipeline start your-crtdl.jsonAfter (v0.10):
aether pipeline start aether.yaml your-crtdl.json
aether pipeline status aether.yaml <job-id>
aether pipeline continue aether.yaml <job-id>
aether job list aether.yamlWhat to update:
- Scripts, cron jobs, and CI: insert the config path as the first argument of each
aethersubcommand, and remove any--configflag — it no longer exists. - No implicit discovery: a bare
./aether.yamlor~/.config/aether/aether.yamlis not picked up automatically. Point at the file explicitly.
2. DIMP output directory renamed pseudonymized/ → dimp/
The DIMP step now writes to jobs/<job-id>/dimp/ instead of jobs/<job-id>/pseudonymized/, matching the step name and the rest of the per-step layout. The file contents are unchanged — only the directory name moved.
What to update:
Downstream tooling and scripts that read pseudonymized output by path must switch
jobs/<job-id>/pseudonymized/→jobs/<job-id>/dimp/.In-flight jobs from v0.9: a job whose DIMP step already ran has its data under
pseudonymized/, but v0.10 looks underdimp/. To resume such a job, rename the directory first:bashmv jobs/<job-id>/pseudonymized jobs/<job-id>/dimpJobs started fresh on v0.10 need no action.
Worth knowing (non-breaking)
- Per-job logs: each job now writes its own log file into its job directory, so logs live alongside that job's data.
- TORCH URL handling: scheme-less, host-prefixed URLs returned by TORCH are handled correctly, fixing double-host URLs.
- Extraction reporting: status now shows the extraction summary instead of a stale cohort-size diagnostic.
See the v0.10.0 release notes for the full changelog.