Build and Push Docker Images / build-api (push) Waiting to runDetails
Build and Push Docker Images / build-web (push) Waiting to runDetails
Build and Push Docker Images / build-worker (push) Failing after 14m30sDetails
- Updated workflow to use parallel jobs for worker, api, web builds
- Added GitHub Actions cache for faster builds
- Fixed Gitea runner to properly support Docker-in-Docker with TLS
- Worker fixes: STAC S3 URL patching, MinIO internal endpoint, model path
- storage.py: Fixed endpoint, credentials, and secure=false
Build and Push Docker Images / build-and-push (push) Failing after 4m27sDetails
- Deleted inference.py (sklearn path) in favor of hybrid_inference.py
- Worker now uses ThreadPoolExecutor for async DW baseline loading
- DW baseline URL sent to client as soon as ready, parallel to inference
- Removed sklearn model fallback (only Hybrid_SpatioTemporal supported)
- Updated docstring to reflect current module dependencies
Build and Push Docker Images / build-and-push (push) Waiting to runDetails
- Adds -v /var/run/docker.sock:/var/run/docker.sock to act_runner options
- Sets DOCKER_HOST=unix:///var/run/docker.sock for runner
- Fixes missing_outputs initialization bug in worker.py
- Implements standard inference path in worker.py using inference.py
Build and Push Docker Images / build-and-push (push) Failing after 1sDetails
- Adds act_runner config.yaml to enable Docker-in-Docker in jobs
- Configures gitea-runner deployment to mount config via ConfigMap
- Enables privileged mode and mounting of /var/run/docker.sock for jobs
- Updates worker with Feature Order V2 and vectorized computation
Build and Push Docker Images / build-and-push (push) Failing after 24sDetails
- Add hybrid PyTorch (TemporalFCN) + CatBoost ensemble logic in new hybrid_inference.py
- Update worker.py to support 'Hybrid' model type with artifact syncing from MinIO
- Integrate odc-stac for raw spectral index fetching from DE Africa STAC
- Update requirements.txt with torch, odc-stac, and rioxarray dependencies
- Include ntfy deployment in k8s manifests
fchinembiri