Marcel Rieger

In our high-energy physics experiments, we're dealing with large amounts data and very complex objects that describe them. For one of our interface tools we are playing around a bit...

Have to pile up on this one ... I tested with a fresh MacOS 11.4 (BigSur) on a M1 MBP, a non-M1 MBP and a M1 Mini, all with latest...

The failing test reports ```shell ModuleNotFoundError: No module named 'openapi_spec_validator' ``` and connected to that ```shell _pytest.nodes.Collector.CollectError: ImportError while importing test module '/home/runner/work/luigi/luigi/test/contrib/ecs_test.py'. ``` Is it possible there's something wrong...

Seems like a temporary glitch, all tests pass now 🎉

Thanks for the feedback @lallea ! > What would the pros and cons be of DynamicRequirements vs similar functionality in the storage layer? Interesting, we also work with in-memory /...

@dlstadther Kind ping :)

I essentially have all changes I mentioned above ready in personal branches (and then work kicked in and couldn't follow up on this anymore). @dlstadther If you agree with the...

I committed the fix of the loop doing the completeness checks during initial scheduling. The main change is that queue objects are no longer required for storing the results of...

Workflow proxies are a way to tell luigi not to use the `run()`, `output()` and `requires()` methods defined in the task itself by the user when `branch` is `-1`. In...

I don't think this will fly on standard machines (like lxplus), since we also install gfal2 + bindings via conda that we strictly need.