- core
- cli.worker: make sure LocalCluster is closed appropriately (#51)
- cli.worker: slack messages now get sent before JobCancelledError is being raised (#51)
- app: also allow cancelling job which is in a retrying state; update slack message without waiting for job manager to do so (#51)
- job_handler.k8s_worker: don't fail if IPv6 address cannot be parsed by requests (#51)
- CLI
- manager: add --skip-dashboard-check flag (#51)
- CI
- pip install hatch "virtualenv<21.0.0" to mitigate hatch build issues with current virtualenv versions
- core
- add "type": "application/json" acrossthe board to have fastapi accept json as content type
- CI
- run python package tests in parallel with integration tests and docker build
- the test-suite aligned/based on mapchete's latest test-suite
- docker base image version bumped to gdal:ubuntu-small-3.12.0
- dependencies
- removed deployment dependency group
- the dependencies source of truth is now pyproject.toml under [project.dependencies]
dependencies
- release the dask and distributed packages versions in the deployment group
- dependencies
- manage pyproject.toml primarily with uv
- release some deps versions, let uv try to resolve them on daily basis via CI
- packaging
- bump dask to 2025.11.0
- bump dask-kubernetes to 2025.7.0
- bump distributed to 2025.11.0
- bump fastapi to 0.121.3
- bump mapchete to 2025.11.0
- bump mapchete-eo to 2025.11.0
- bump shapely to 2.1.2
- packaging
- automatically build and push images to registry: ghcr.io/mapchete/mapchete-hub
- packaging
- first PyPI release after migrating to GitHub
- packaging
- bump mapchete_internal_processes to 2025.5.3
- set Programming Language in pyproject.toml to: 3.10, 3.11, 3.12, 3.13
- CI
- bump build image in Dockerfile to mapchete:2025.5.1 as well as test CI images
- packaging
- bump dask to 2025.5.0
- bump distributed to 2025.5.0
- bump mapchete to 2025.5.1
- bump mapchete_internal_processes to 2025.5.2
- packaging
- bump mapchete_eo to 2025.5.0
- bump mapchete_internal_processes to 2025.5.1
- CI
- bump build image in Dockerfile to mapchete:2025.5.0 as well as test CI images
- packaging
- bump mapchete to 2025.5.0
- bump mapchete_internal_processes to 2025.5.0
- CI
- bump build image in Dockerfile to mapchete:2025.4.2 (using NumPy 2.2.5)
- packaging
- bump dask to 2025.4.1
- bump dask-gateway to 2025.4.0
- bump dask-gateway-server to 2025.4.0
- bump dask-kubernetes to 2025.4.3
- bump distributed to 2025.4.1
- bump fastapi to 0.115.12
- bump mapchete_internal_processes to 2025.4.1
- bump zarr to 3.0.7
- add httpx to dependencies
- CI
- bump build image in Dockerfile to mapchete:2025.4.1 (now runs on Python 3.13)
- remove mapchete_xarray dependency from build
- CI
- bump build image in Dockerfile to mapchete:2025.4.0
- fix mapchete_xarray and numcodecs deps in Dockerfile for build
- CI
- bump build and test image to mapchete:2025.4.0
- packaging
- bump mapchete to 2025.4.0
- bump mapchete_eo to 2025.4.0
- bump mapchete_internal_processes to 2025.4.0
- packaging
- bump mapchete to 2025.3.1
- packaging
- bump fastapi to 0.115.7
- bump mapchete_eo to 2025.1.2
- CI
- bump test image to mapchete:2025.1.1
- packaging
- add and freeze zarr==2.18.4
- bump fastapi to 0.115.6
- bump mapchete_eo to 2025.1.1
- packaging
- bump mapchete_eo to 2025.1.0
- packaging
- bump mapchete_eo to 2024.12.0
- core
- mhub-worker: also run job if it is in retrying status; add log info messages
- packaging
- bump mapchete to 2024.12.0
- packaging
- bump mapchete_eo to 2024.11.6
- packaging
- bump mapchete to 2024.11.1
- bump mapchete_eo to 2024.11.5
- core
- mhub-manager: refactored; retry kubernetes jobs if they are stalled
- packaging
- bump mapchete_eo to 2024.11.5
- packaging
- use base image 2024.11.0
- bump mapchete to 2024.11.0
- bump mapchete_eo to 2024.11.3
core
- mhub-manager: improve log messages
packaging
- bump mapchete_eo to 2024.10.4
- core
- mhub-manager: ignore job if job cannot be found on cluster
- core
- mhub-manager: submit job to k8s cluster if job cannot be found on cluster
- core
- mhub-manager now has subcommands (mhub-manager watch replaces the current behavior using mhub-manager --watch)
- mhub-manager clean can now be run separately
- consolidate mhub-* logger configurations
- extend list of env var prefixes to be passed on to workers
- fix docker-compose backend db setting
- implement first version of kubernetes jobs retry mechanism (untested)
- core
- add proper logging to mhub-manager and mhub-worker; also let log level be set via env settings
- packaging
- bump
mapchete_eoto2024.10.4
- bump
packaging
- bump
mapchete_eoto2024.10.3
- bump
packaging
- bump
mapchete_eoto2024.10.2
- bump
packaging
- bump
mapchete_eoto2024.10.1
- bump
packaging
- bump
mapchete_eoto2024.10.0
- bump
core
- do not retry k8s jobs by default, we will have this in mhub itself; in job_handler.k8s_worker backoff_limit=retry_job_x_times (default: 0)
- terminate k8s job as soon as it finishes; in job_handler.k8s_worker ttl_seconds_after_finished=remove_job_after_seconds (default: 0)
core
- fix bug in SlackMessenger text split
core
- make sure there is always an init message on slack to append updates to
- split traceback messages to make sure it is properly formated (#247)
core
- add k8s-managed-worker as job_handler setting
- submit jobs from app using a background task
CLI
- add mhub-manager command
core
- introduce JobHandlers to create abstraction layer between default background thread tasks, new kubernetes job tasks and mhub-worker tasks
- timestamps now all have UTC timezone
CLI
- add mhub-worker command
packaging
- use ruff and mypy
CI/CD
- add codecheck stage
packaging
- bump
mapchete_eoto2024.9.3 - bump
mapchete_internal_processesto2024.9.3
- bump
packaging
- bump
mapchete_eoto2024.9.2
- bump
packaging
- bump
mapchete_internal_processesto2024.9.3
- bump
packaging
- bump
mapchete_internal_processesto2024.9.2
- bump
packaging
- bump
mapchete_eoto2024.9.1 - removed
eox_preprocessing - added
mapchete_internal_processes
- bump
core
- add timeouts for pymongo connections
packaging
- use base image
2024.9.1 - bump
daskto2024.8.2 - bump
dask-gatewayto2024.1.0 - bump
dask-gateway-serverto2024.1.0 - bump
dask-kubernetesto2024.8.0 - bump
distributedto2024.8.2 - bump
fastapito0.114.1 - bump
mapcheteto2024.9.0 - bump
mapchete_eoto2024.9.0
- use base image
packaging
- bump
mapchete-eoto2024.6.0
- bump
packaging
- bump
mapchete-eoto2024.5.9
- bump
packaging
- bump
mapchete-eoto2024.5.8
- bump
packaging
- bump
mapchete-eoto2024.5.7
- bump
packaging
- bump
mapchete-eoto2024.5.6
- bump
packaging
- bump
mapchete-eoto2024.5.5
- bump
packaging
- bump
mapchete-eoto2024.5.4
- bump
packaging
- bump
mapchete-eoto2024.5.3
- bump
packaging
- bump
mapchete-eoto2024.5.2
- bump
core
- settings.get_das_specs(): make sure worker_environment setting values are strings, otherwise k8s breaks
packaging
- bump
mapchete-eoto2024.5.1
- bump
packaging
- bump
mapchete-eoto2024.5.0
- bump
packaging
- bump
mapcheteto2024.5.0
- bump
packaging
- bump
mapchete-eoto2024.4.3
- bump
packaging
- bump
mapchete-eoto2024.4.2
- bump
packaging
- repackaging
2024.4.2but with corrected version
- repackaging
packaging
- bump
mapchete-eoto2024.4.0
- bump
packaging
- bump
mapcheteto2024.2.1same as in docker-base
- bump
- core
- settings.MhubSettings: replace cancellederror_tries setting with retries and add a retry_on_exception setting including more exceptions than just a CancelledError
- packaging
- accidentally tagged 2024.3.5 with 2024.3.6, so fixing this by aligning both versions again
packaging
- bump
mapchete-eoto2024.3.6
- bump
packaging
- bump
mapchete-eoto2024.3.5
- bump
packaging
- bump
mapchete-eoto2024.3.4
- bump
packaging
- bump
mapchete-eoto2024.3.3
- bump
packaging
- bump
mapchete-eoto2024.3.1
- bump
packaging
- bump
mapchete-eoto2024.3.0 - use base image
2024.2.1
- bump
- core
- fix slack messaging
- offload job creation to background task
- core
- fix slack_sdk.WebClient call
- core
- add lifespan resources for FastAPI app (status DB handler, job threadpool, optional local dask cluster)
- use concurrent.futures.ThreadPool instead of FastAPI background tasks to run jobs
packaging
- bump
mapchete-eoto2024.2.6
- bump
packaging
- bump
mapchete-eoto2024.2.5
- bump
packaging
- bump
mapchete-eoto2024.2.4
- bump
packaging
- bump
mapchete-eoto2024.2.3
- bump
core
- only close connections to dask Gateway while not in use
packaging
- bump
mapcheteto2024.2.0 - bump
mapchete-eoto2024.2.2 - use base image
2024.2.0
- bump
- core
- close connections to dask Gateway and GatewayCluster while not in use
- core
- keep connection to GatewayCluster open
packaging
- bump
mapchete-eoto2024.2.1
- bump
- core
- close connections to dask Gateway and GatewayCluster while not in use
packaging
- bump
mapchete-eoto2024.2.0
- bump
- core
- fix Slack messages
CI
- run all jobs on privileged` runners with docker
packaging
- bump
mapcheteto2024.1.5
- bump
core
- slack messages: report in slack threads instead of single messages
- fix worker settings when adapting cluster
packaging
- replace
Slackerdependency withslack_sdk
- replace
core
- slack messages: also print exception representation, not just traceback
packaging
- bump
mapcheteto2024.1.3
- bump
core
- allow aborting jobs in pending mode
- differentiate between submitted and started time stamps
- add retry mechanism on requesting the dask cluster
- track Exception in DB using repr() instead of str() to better keep track of exception type
packaging
- bump
mapchete_eoto2024.1.4
- bump
packaging
- bump
mapchete_eoto2024.1.3
- bump
CI
- use base image
2024.1.2
- use base image
packaging
- use base image
2024.1.2 - bump
mapcheteto2024.1.2 - bump
mapchete_eoto2024.1.2
- use base image
CI
- add pushing to aws registry to eox gitlab CI
CI
- use base image
2024.1.0
- use base image
packaging
- use base image
2024.1.0 - bump
mapcheteto2024.1.0 - bump
mapchete_eoto2024.1.0
- use base image
CI
- use base image
2023.12.2
- use base image
core
- adaptive DaskSpecs and DaskSettings now can also be passed to mapchete under dask_specs in the yaml config
packaging
- use base image
2023.12.2 - bump
mapcheteto2023.12.2 - add
eox_preprocessingversion2023.12.0for backwards compability
- use base image
CI
- use podman layer caching
core
- fix db.mongodb jobs parsing
- pass on DaskSpecs and DaskSettings to cluster.get_dask_executor
- rewrite and use cluster.cluster_adapt
- minor fixes to Slack messages
packaging
- bump
daskto2023.12.0 - bump
distributedto2023.12.0 - bump
mapcheteto2023.12.1 - bump
mapchete_eoto2023.12.0 - clean up unused dependencies in
pyproject.toml
- bump
CI
- use base image
2023.12.1 - add
isortto pre-commit
- use base image
core
- settings: use pydantic_settings to collect mhub configuration from environment and defaults
- use job states from mapchete.enums.Status
- use completely refactored mapchete.commands.execute function with now integrated retries & observer classes in newly created job_wrapper module
- use observer classes (in observers module) to update status DB and send Slack messages
- create db submodule for mongo DB and memory status handlers
- define models.JobEntry model to ship around job metadata in from status handlers
- extract some functionality from app to job_wrapper and cluster modules
packaging
- use base image
2023.11.0also for CI - bump
dask-gatewayto2023.9.0 - bump
dask-gateway-serverto2023.9.0 - bump
dask-kubernetesto2023.10.0 - bump
daskto2023.11.0 - bump
distributedto2023.11.0 - bump
fastapito0.104.1 - bump
mapcheteto2023.11.0 - bump
mapchete_eoto2023.11.0
- use base image
packaging
- use base image
2023.8.0for tests as well - bump
fastapito0.103.1 - restrict
pydanticto<2.0.0 - add
httpxto dependencies
- use base image
packaging
- use base image
2023.8.0 - bump
mapcheteto2023.8.1
- use base image
packaging
- bump
eox_preprocessingto2023.8.0
- bump
packaging
- use base image
2023.7.1 - bump
daskto2023.5.0 - bump
distributedto2023.5.0 - bump
dask-gatewayto2023.1.1 - bump
dask-kubernetesto2023.3.2 - bump
eox_preprocessingto2023.7.3 - bump
fastapito0.100.0 - bump
mapcheteto2023.7.1
- use base image
packaging
- use base image
2023.7.0 - bump
mapcheteto2023.7.0
- use base image
core
- add retry_flag to only wait for newly started jobs, if retried by CancelledError, go ahead and start processing without delay
core
- wait for jobs in states MHUB_PROCESSING_STATES for 10 seconds with up to MHUB_MAX_PARALLEL_JOBS (default: 2)
- add wait parameter MHUB_PREPROCESSING_WAIT (default: 0) after preprocessing to offset possible lag for mongoDB and DaskExecutor connection
core
- add fiona.open to read --area files (tested and works on s3 stored files)
- if bounds and area given use intersection as geometry
packaging
- use base image
2023.5.0 - revert
mapcheteto2023.4.1
- use base image
core
- add passing area param to mhub config to initialize job geometry
- add test for area as WKT geometry and as fgb vector file
packaging
- bump
daskto2023.5.0 - bump
distributedto2023.5.0
- bump
packaging
- use base image
2023.6.0 - bump
mapcheteto2023.6.1
- use base image
CI/CD
- deactivate integration tests
core
- fix retry of CancelledError by reinitializing the whole job to skip existing output
packaging
- bump
daskis2023.4.0 - bump
distributedis2023.4.0 - bump
mapcheteto2023.4.1 - bump
mapchete_satelliteto2023.5.5
- bump
core
- integrate URLs into text of Slack messages
- enable retrying jobs when they raise a CancelledError configurable by environment variable MHUB_CANCELLEDERROR_TRIES
- try to get dask scheduler logs after a failed job
packaging
- use base image
2023.1.1 - bump
mapchete_satelliteto2023.1.9 - bump
s2brdfto2023.1.0
- use base image
CI/CD
- remove
mhub-s1image
- remove
packaging
- bump
daskis2022.12.1 - bump
dask-kubernetesis2022.12.0 - bump
distributedis2022.12.1 - bump
eox_preprocessingto2022.12.0 - bump
mapcheteto2022.12.0 - bump
mapchete_satelliteto2022.12.2 - use base image
2022.12.0
- bump
packaging
- bump
mapcheteto2022.11.2 - bump
mapchete_satelliteto2022.12.1 - use base image
2022.11.2
- bump
packaging
- bump
mapcheteto2022.11.1 - bump
mapchete_satelliteto2022.11.4
- bump
packaging
- use hatch instead of setuptools
- build, test and upload python package to registry for every release
packaging
- bump
eox_preprocessingto2022.11.1 - bump
mapcheteto2022.11.0 - bump
mapchete_satelliteto2022.11.3 - use base image
2022.11.0
- bump
packaging
- bump
daskto2022.11.0 - bump
dask-gatewayto2022.11.0 - bump
dask-gateway-serverto2022.11.0 - bump
dask-kubernetesto2022.10.1 - bump
distributedto2022.11.0 - bump
eox_preprocessingto2022.11.0 - bump
fastapito0.87.0 - bump
mapchete_satelliteto2022.11.2 - run pip check after image build
- bump
packaging
- bump
planet-signals-generationto2022.10.3 - add and freeze
mapchete_xarrayto2022.10.0
- bump
core
- reinstall and use
numcodecsfrom pypi as to fill any library or dependency gaps
- reinstall and use
packaging
- bump
mapchete-satelliteto2022.10.1
- bump
packaging
- bump
planet-signals-generationto2022.10.2 - bump
daskanddistributedto2022.10.0`
- bump
core
- add environment parser for dask_gateway.options
- add test for dask_spec in settings.py
- enable parsing of AWS, DASK, GDAL, MHUB, MAPCHETE, MP ENV variable for dask-scheduler and dask-worker
- fix docker-compose.yml the --nprocs to --nworkers to fit newer
daskanddistributedversions
packaging
- bump
planet-signals-generationto2022.10.1
- bump
packaging
- bump
mapchete_satelliteto2022.10.0
- bump
packaging
- bump
dask-kubernetesto2022.9.0 - bump
planet-signals-generationto2022.10.0
- bump
core
- add an optional in-memory status DB if no MongoDB is present
- dask LocalCluster() now uses processes & threads by default
- add mhub-server CLI to quickly start an mhub instance
- adaptive scaling is now deactivated by default unles MHUB_DASK_ADAPTIVE_SCALING is set to TRUE
CI/CD
- run only integration tests on integration test stage
- start mhub by using new mhub-server CLI instead of uvicorn`
- use in-memory status DB in tests instead of mongomock.MongoClient()
packaging
- use base image
2022.9.0 - don't tag
latestimages anymore - use dask-gateway pypi release instead of building from source
- bump
daskto2022.9.0 - bump
dask-gatewayto2022.6.1 - bump
dask-gateway-serverto2022.6.1 - bump
dask-kubernetesto2022.7.0 - bump
distributedto2022.9.0 - bump
fastapito0.85.0 - bump
mapcheteto2022.9.0
- use base image
- CI/CD
- every pushed commit now generates a docker image with the short commit hash as tag
- split up into unit and integration tests
- dump pip installed packages and versions as build job artefacts
- packaging
- update dependencies: dask==2022.5.0, dask-kubernetes==2022.4.1, distributed==2022.5.0, mapchete_satellite>=2022.5.0
- add planet-signals-generation
core
- fix cluster size adaption
packaging
- use base image
2022.4.0
- use base image
core
- align <job_id>/results with current OAPI standard
packaging
- use base image
2022.3.2
- use base image
packaging
- bump
dask-gatewayto0a69d3d711a7bd472c724ad5d58c11d5a8ced61d - bump
daskto2022.3.0
- bump
packaging
- use base image
2022.3.1
- use base image
core
- request dask cluster after job was initialized
- process dask task graph per default
- use different adapt_options if dask task graph is used
packaging
- set
mapcheteto2022.2.2 - set
mapchete_satelliteto2022.2.0
- set
packaging
- set
mapcheteto2022.2.0 - freeze
dask-gatewayto commitbee9255e5ea0d77f456985cd91b2622bb3776dbb
- set
packaging
- set
daskand distributed to2022.1.1 - set
dask-kubernetesto2022.1.0 - set
mapcheteto2022.1.2
- set
packaging
- set
mapchete_satelliteto2022.1.2
- set
packaging
- set
mapchete_satelliteto2022.1.1
- set
packaging
- use base image
2022.1.0 - set
mapchete_satelliteto2022.1.0
- use base image
core
- use context managers for all dask Client and Cluster instances
- add more meaningful logger.info messages
packaging
- set
eox_preprocessingto2021.1.0 - set
fastAPIto0.72.0
- set
core
- use async for all fastAPI request functions
core
- pass on cluster.adapt() kwargs via 'adapt_options' section in dask specs JSON
packaging
- use base image
2021.12.3 - set
mapcheteto2021.12.3
- use base image
packaging
- use base image
2021.12.2 - set
mapcheteto2021.12.2
- use base image
packaging
- use base image
2021.12.1 - set
mapcheteto2021.12.1
- use base image
packaging
- set
mapchete_satelliteto2021.12.4(fix prior release)
- set
packaging
- set
mapchete_satelliteto2021.12.4
- set
packaging
- set
mapchete_satelliteto2021.12.3
- set
packaging
- set
mapchete_satelliteto2021.12.2
- set
core
- use 8 worker 2 threads (on an 8 core machine) default specification for dask workers
packaging
- set
eox_preprocessingto2021.12.0
- set
packaging
- set
mapchete_satelliteto2021.12.1
- set
core
- improve slack messages
packaging
- use base image
2021.12.0 - set
mapcheteto2021.12.0
- use base image
packaging
- set
mapchete_satelliteto2021.12.0
- set
dockerfile
- add step with go to build wheels of dask-gateway packages
packaging
- use latest/main dask-gateway version
core
- cache BackendDB connection
packaging
- use base image
2021.11.3 - add
orgonite>=2021.11.0to dependencies
- use base image
core
- add slack notifications
packaging
- set dask and distributed versions to
2021.11.1 - set dask-kubernetes to
2021.10.0
- set dask and distributed versions to
core
- set cluster worker minimum as either default or tiles tasks
- submit tasks in chunks, not one by one (see mapchete/mapchete#387)
packaging
- set minimum mapchete version to
2021.11.2 - use base image
2021.11.2
- set minimum mapchete version to
core
- set cluster worker maximum as maximum of preprocessing and tiles tasks
- large jobs now start earlier and use less ressources (mapchete/mapchete#384)
packaging
- set minimum mapchete version to
2021.11.1 - use base image
2021.11.1
- set minimum mapchete version to
core
- fix cluster initialization
core
- enable posting custom dask specs as JSON
packaging
- re-enable
mapchete_xarray - use base image
2021.11.0
- re-enable
core
- add updated timestamp also on new job
dependencies
mapchete_satelliteversion to2021.10.3
dependencies
mapchete_satelliteversion to2021.10.2
dependencies
mapcheteversion to2021.10.3
dependencies
mapchete_satelliteversion to2021.10.1
core
- set worker threads to 1 in default dask specs
dependencies
mapchete_satelliteversion to2021.10.0mapcheteversion to2021.10.1
packaging
- change version numbering scheme to
YYYY.MM.x
- change version numbering scheme to
Docker
- update base image
docker-base mapchete:2021.10.1for mhubsnap-mapchete-ost:2021.10.1for mhub-s1
- update base image
- fix GeoJSON creation if
boundsfield is not available.
fix default random job names
fix dask specs write into metadata
add
boundsto GeoJSONDocker
update base image
docker-basemapchete:0.24for mhubsnap-mapchete-ost:0.24for mhub-s1
dependencies
mapchete_satelliteversion to0.17daskversion to2021.9.1distributedversion to2021.9.1
- add
dask_dashboard_linkto job metadata - enable configuration of dask scheduler & workers via env variables when using dask gateway
- use black & flake8 for code
- re-enable full integration tests
NOTE: major code changes!
- replaced Celery with dask
- moved CLI functionality and api module into separate
mapchete_hub_clipackage - replaced
flaskwithFastAPI - deactivated xarray and Sentinel-1 support/tests(!) for now
fixed the mhub state query (#120)
Docker
pip-compile is now used to resolve dependeny graph before installing requirements
dependencies
- update
mapcheteto>=0.38 - update
mapchete_satelliteto0.15 - update
eox_preprocessingto0.13
- update
update base image
docker-basemapchete:0:17for mhubsnap-mapchete-ost:0:17for mhub-s1
Docker
dependencies
update mapchete_satellite to 0.14
pass
AWS_REQUEST_PAYERto fiona cloudmask reading to enable reading of L1C masksread_cloudmasks functions now support cloud_types arg
- default: ['opaque', 'cirrus']
- this allows to choose which cloudmasks will be read in all read functions
Docker
update to 0.16 base image
version updates
- mapchete 0.37
dependencies
update mapchete_satellite to 0.13
replace catching all rasterio errors using
mapchete.errors.MapcheteIOErrorclass (!96)make retry decoration settings configurable via env (!96)
MP_SATELLITE_IO_RETRY_TRIES(default: 3)MP_SATELLITE_RETRY_DELAY(default: 1)MP_SATELLITE_IO_RETRY_BACKOFF(default: 1)
packaging:
- increase mapchete minimum dependency to 0.37 (!96)
Docker
update to 0.14 base image
version updates
- GDAL 3.2.0
dependencies
update mapchete_satellite to 0.12
S2AWS_COG:
- switch off catalog concurency for S2 STAC search endpoint (#82)
- retry rasterio.errors.CRSError and rasterio.errors.CRSError (#83, #84)
make Dockerfile more dev-friendly (!101)
remove requester pays ENV setting as it should be provided on deployment (!101)
starter scripts
- added AWS_DEFAULT_REGION to starter scripts (#124)
main package
pin Celery dependency to <5.0.0 because of breaking changes in API
API
- remove default progress timeout
CLI
- better make use of tqdm api
Docker
update to 0.13 base image
version updates
- Fiona 1.8.17
- GDAL 3.1.3
- GEOS 3.7.1 (downgraded from 3.8.1)
- OpenSAR Toolkit 0.9.8
- proj 7.1.1
- pyproj 2.6.1
dependencies
updated mapchete_satellite to 0.11
- enable S2AWS_COG archive
- enable BRDF correction
use new internal PyPi instance from EOX GitLab to install internal packages
testing
- use CI_JOB_ID instead of random hash for docker-compose project in order to clean up running containers & volumes properly after test run
main package
add worker event rate limit (!85, #67)
CLI
- fix missing output_path in verbose mode (!81)
- fix worker count (!83)
- add mhub workers subcommand (!84)
starter scripts (#106)
restructured directories
added
- idle_workers.sh
- live_worker_info.sh
Docker
use base image 0.12 which updates
- GDAL 2.4.4 (downgraded from 3.0.4)
- proj 5.2.0 (downgraded from 6.3.2)
install latest boto3 version
testing
- use random ports and use unique name for docker-compose project (!88)
main package
- fix job termination (#108)
Docker
use base image 0.11 which updates
- GDAL 3.0.4
- Fiona 1.8.13.post1
- mapchete 0.35
main package
- use a MongoDB instance as message broker (!69)
- use a MongoDB instance as backend database for jobs (!69, !70)
- cancel jobs (!69, #4)
- monitor does not have to run on same machine than server anymore (!69)
- mapchete_satellite: increase remote timeout to 30s (!74, #88)
testing
- run integration tests using docker-compose (!69, #44)
- retry test stage (!72)
Docker
use mapchete_satellite 0.10
use base image 0.10 which updates
- GDAL 3.1.2
- Fiona 1.8.13
- GEOS 3.8.1
- mapchete 0.34
- proj 6.3.2
- rasterio 1.1.4
- spatialite 5.0.0-beta0
- SQLite 3310100
Docker
- use mapchete_satellite 0.9
- use base image 0.9 which updates OpenSAR toolkit to 0.9.7
Docker
- use orgonite 0.6 and don't extra install Cython
- use base image 0.8 which fixes ost version mismatch for mhub_s1 image (#91)
repository
- removed deprecated Mapfiles
Docker
- use base image 0.7
- automate docker builds
- add full zarr support in builds
API
- require to encode custom process code as base64 string
- fix passing on query parameters to /jobs/ endpoint (#89)
CLI
- add
--timeoutparameter - increase verbose output
- add
--debugflag to all subcommands - add
remote-versionsquery
- add
monitor
- make sure job events have a
job_idbefore updating the database - add
job_namefilter - rename
StatusHandler.all()toStatusHandler.jobs()
- make sure job events have a
API
- don't append queue information in capabilities.json
- add queue length (i.e. number of jobs waiting) to response
- add /queues/<queue_name> to API
seeding
- added mercator configurations
- fixed compression setting in mapfiles
- increased
eox_preprocessingdependency to0.9 - mhub handles jobs with other CRSes than EPSG:4326 (fixes #59)
- fix query error when filtering by queues or commands (#73)
- enable posting of custom process file (#52)
- fix rendering artefacts by changing mapserver scaling
- remove AWS credentials from mapfile & adapted starter script to temporarily include credentials from environment
- increased
eox_preprocessingdependency to0.8 - increased base image version for Dockerfile to
0.3 - added AWS management scripts
- use multistage docker builds to reduce image size
- enable filters to better query jobs (#53)
- print more details using
mhub jobsandmhub status <job_id> - rename
mapchete_hub.workermodule tomapchete_hub.commands - remember timestamp on
task-receivedandtask-startedevents instartedproperty - add Celery-Slack integration (#26)
- add
mhub batchcommand - pass on mapchete config as
OrderedDict(#48) - serialize Cerlery messages as
JSONinstead of pickling - fix
announce_on_slacksetting (#66 #25)
- fixed preprocessing dependency from version 0.4 to 0.5
Docker image registry.gitlab.eox.at/maps/mapchete_hub/mhub:0.3
- bases on registry.gitlab.eox.at/maps/docker-base/mapchete:0.2
add
mhub executeandmhub indexcommands (#54)API returns more useful error message for client
automatically assign job ID (#64)
only use one docker image for all mhub services: registry.gitlab.eox.at/maps/mapchete_hub/mhub
Docker image registry.gitlab.eox.at/maps/mapchete_hub/mhub:0.2
- bases on registry.gitlab.eox.at/maps/docker-base/mapchete:0.1
celery worker now capture logs again (#62)
zone_worker and preview_worker modules renamed to execute and index like their mapchete counterparts (#60)
use tagged versions instead of branches in docker base images & depdendencies (#58)
move mapserver and mapcache docker images to docker-base repository (#57)
generate capabilities.json (#51)
filter jobs by process output path (#40)
list available processes
list active queues and workers
use built-in mapchete batch functions (#47)
added better unit test coverage for most flask & celery related code parts (#7)
start monitor in child process (#23)
use built-in mapchete batch functions (#47)
deploy application as WSGI using gunicorn (#20)
added mapchete_hub.api.API class which abstracts all the relevant requests to the API
when starting a worker, a queue can be specified which solves (#32)
switched to mapchete_satellite backend
added image filter functions
pyproj metis support 1.9.5.1
Sentinel-1 integration and images
mhub, broker, etc. s1processor for mundi
- first build