Support multiplication and division of datasets. #792
Conversation
tests/test_dataset.py
Outdated
| def test_multiply_datasets_must_match(): | ||
| dsa = mikeio.Dataset({"Foo": mikeio.DataArray([1, 2, 3])}) | ||
| dsb = mikeio.Dataset({"Bar": mikeio.DataArray([4, 5, 6])}) | ||
| with pytest.raises(ValueError): |
There was a problem hiding this comment.
Why do the name/item need to match on multiplication/division?
There was a problem hiding this comment.
Let's try with minimal requirements. If two datasets have the same number of items, and the shapes of the data is identical, we can multiply them.
There was a problem hiding this comment.
Additional thought: could this feature also be used for scaling several items in a file with the same spatially/temporally varying field?
a) Should we allow that shapes conform but items in "other" is either the same as "this" or 1 (meaning same scaling applied to all variables).
b) Even further should we allow that as long as broadcasting rules allow multiplication then shape of "other" is okay? (so I could e.g. scale a dfsu with a dfs0)
There was a problem hiding this comment.
If we decide to go down this rabbit hole, I think we should consider restructuring MIKE IO to rely on xarray instead of numpy and to make use of the broadcasting in xarray. https://docs.xarray.dev/en/stable/user-guide/computation.html#broadcasting-by-dimension-name
There was a problem hiding this comment.
Let's wait until someone has the need
mikeio/dataset/_dataset.py
Outdated
| data = [self[x].to_numpy() / value for x in self.items] | ||
| except TypeError: | ||
| raise TypeError(f"{value} could not be divided to Dataset") | ||
| items = deepcopy(self.items) |
There was a problem hiding this comment.
I guess the unit is often not the same after division 🤔
There was a problem hiding this comment.
How to handle units in a pragmatic way is what I am struggling with, and I suppose one of the reasons that this is not already supported.
tests/test_dataset.py
Outdated
| def test_divide_dataset_must_match(): | ||
| dsa = mikeio.Dataset({"Foo": mikeio.DataArray([1, 2, 3])}) | ||
| dsb = mikeio.Dataset({"Bar": mikeio.DataArray([4, 5, 6])}) | ||
| with pytest.raises(ValueError): |
No description provided.