Features:
- Report if framework not detected by
spark_submit.system_info()
Misc/Internal
- Better way to construct the spark-submit command
Features:
- Addition of
spark_submit.SparkJob.get_id()
method to obtain the Spark submission ID - Use all environment variables from
os.environ
ifuse_env_vars=True
inspark_submit.SparkJob.submit()
- Additional environment variables can be used in
spark_submit.SparkJob.submit()
withextra_env_vars
argument
Misc/Internal
spark_submit.SparkJob.submit()
now uses keyword arguments only- Argument
await_result
is changed topoll_time
inspark_submit.SparkJob.submit()
- No default environment variables to be used with
spark_submit.SparkJob.submit()
- Code standards improvements and general refactoring
Misc/Internal
- Refactor to comply with
pylint
,isort
andmypy
checks - Minor logic improvements
Features
- Added optional
timeout
argument forspark_submit.SparkJob.submit()
Improvements
- Improve support for Yarn and Kubernetes clusters on polling and killing submitted tasks
Misc/Internal
- Use dedicated file for default arguments and Spark states
- Use consistent path separators where possible
Features
- Addition of
spark_submit.SparkJob.get_submit_cmd()
method - Introduction of
spark_submit.SparkJob.concluded
attribute to keep track of submitted job
Improvements
- Improve
spark_submit.system_info()
functionality - More generic status polling
- Minor refactoring
Bug Fixes
- Fix return code bug in
spark_submit.SparkJob.get_code()
Improvements
- Change argument order in
spark_submit.SparkJob.submit()
- Minor improvements
- First release