- LocalPipelineLock for single-thread non-networked pipeline applications. (#80)
- Local lock interface is now exposed through the DataContext for use of arbitrary locks in the pipeline. (#80)
- Bumped
redisdependency to 4.3.1. (#94) - Pipeline class now uses middleware architecture for customizable pre/post run actions. (#80)
1.1.1 - 2022-05-05
- Support for Python 3.6 through 3.10. (#90)
StandalonePipelineno longer has a dependency onredis. SA and HA pipeline class references must now use the fully-qualified class name. (#89)
1.1.0 - 2022-05-01
ConsoleMetricsExporterfor locally debugging pipelines without an APM service. (#46)- Built-in Elastic APM metrics exporter. (#47)
Sequenceclass for logical groupings of steps. (#53)
- Bumped
redisdependency to 4.2.2. (#50) - Bumped
elastic-apmto 6.9.1. (#56)
- Dependencies for
MetricsExporterandPipelineLockmodules are no longer included in the base package and must now be installed separately throughwatergrid[...]metapackages. (#54) - Documentation from Readme. Docs now located at https://armmaster17.github.io/watergrid-python/. (#65)
- Resolved issue with objects inside a
DataContextnot being copied when the output mode of a step is set toSPLIT. (#6) - Pipeline now stops running if there are no more contexts to process in a new step. (#78)
1.0.1 - 2022-04-01
- Broken release pipeline to PyPI.
1.0.0 - 2022-04-01
- Core library functionality.
- High availability pipeline using the
HAPipelineclass. RedisLockfor high availability mode using Redis.- Package distribution on PyPI.
- Shared step contexts.
- Context manipulation through the
OutputModeenum. - Base class for building a metric exporter to an external APM solution.
- Staggered node startup no longer causes mid-interval pipeline runs on other nodes in HA mode.