Julius Busecke

Interesting question @hhweiCT. Re the skin function I will raise a detailed issue shortly. I am concerned about this: > When I use skin function, it crashed without any error...

> Or ideally, the below would run, with the xarray version > > ```python > # Valid data range according to documentation > # https://github.com/xgcm/aerobulk-python/blob/main/source/aerobulk/flux.py > VALID_VALUE_RANGES = {'sst': [270,...

I will have to jump off for a bit but here are some suggestions how to proceed: The error you are getting suggests that the issue lies in [these lines](https://github.com/xgcm/aerobulk-python/blob/f1affeb275ae86ea21f3b5a59d83f9238d380cce/source/aerobulk/flux.py#L187C1-L194C6)...

> Could there be something telling in the way that data is created in the aerobulk-python test py file, here: (https://github.com/xgcm/aerobulk-python/blob/f1affeb275ae86ea21f3b5a59d83f9238d380cce/tests/create_test_data.py)? Yes comparing these and figuring out what the difference...

Ok so the issue with the xarray wrapper was that the inputs were lists (not numpy arrays or as required here xarray.Dataarrays). This works for me with the newest (v0.4.0)...

I have raised #80 to clarify the input for the range checking. Also very importantly the results in #81 shows that even if this works, it will **likely produce the...

I suspect that either the wrapper or the fortran code will just pick the first value of any iterable (including a list as given here), since these inputs are not...

Hey @mgorfer thank you so much for getting involved with xgcm more. I think this is a great suggestion. How would you imagine this to happen? By default or with...

I remembered and found online that the relevant attribute is `tracking_id` which usually has values that look like `'hdl:xxx/xxx'`. I had trouble finding any info on how to lookup the...

😭 There are datasets without the `tracking_id` field?!?! Oh well thats a problem I cannot solve today, but I now have prototype code ```python import requests def handle_to_url(handle): # convert...