v1.1.2

This version just cleans up some of the code and fixes some of the mix ups of job references and job ids. This does not resolve all issues but is a step in the right direction.

v1.1.1

This release just contains a bug fix regarding an error that can occur if there is an unexpected output by just importing your code, such as deprecation warnings. One of the last versions introduced a feature for checking the compatibility of your code structure, which can be affected by this. If such a thing happens, it will now just deactivate this feature, until I had the time for a more robust implementation.

Long term, I plan to make the syntax of Slurminade closer to FastAPI, using Pydantic and function annotations to communicate savely with the workers. Unfortunately, I am crazy busy and cannot predict when I will be able to do this, as it is rather a "nice to have" but things kind of work for us as they are, so there is no immediate pressure. This will go hand in hand with a better integration into AlgBench, which I also plan to move towards using more pyndatic models instead of just wild jsons.

v1.1.0

What's Changed

• v1.1.0 by @d-krupke in #23

Slurminade can now also be called from iPython! In this case, it will just use the defining file as entry point. This will still have problems with relative imports, though.

Full Changelog: v1.0.1...v1.1.0

v1.0.1

What's Changed

Instead of JobIds (ints), the dispatcher will now return JobReferences which will allow to do fancier stuff in the near future.

• v1.0.1 by @d-krupke in #22

Full Changelog: v0.10.1...v1.0.1

v0.10.1

What's Changed

• FIX: Don't call setup on listfuncs by @d-krupke in #21

Full Changelog: v0.10.0...v0.10.1

v0.10.0

What's Changed

• v0.10.0 by @d-krupke in #20

Full Changelog: v0.9.0...v0.10.0

Batch is now named JobBundling. There is a method join for easier synchronization. exec allows to executed commands just like srun and sbatch, but uniform syntax with other slurmified functions. Functions can now also be called with distribute_and_wait. If you call python3 -m slurminade.check --partition YOUR_PARTITION --constraint YOUR_CONSTRAINT you can check if your slurm configuration is running correctly.

v0.9.0

Contains a set of improvements. Further improvements are planned but may be breaking changes. Still thinking about the best way to do it.

v0.8.1

This is just a tiny improvement when calling wait_for with an empty list. This is often in combination with a batch.flush() outside the context (as the batch gets automatically flushed at the end of the context and all further flushes will be empty.

v0.8.0

Some users seem to struggle with understanding what is happening. This update adds some logging. You can easily deactivate it if you don't need it, as it uses Python's logging framework. Additionally, some type warning were resolved.

v0.7.1

Just updating the CI, see v0.7.0 for the actual changes.