Fix/Scatter plot density with few points

[jsmariegaard]

Scatter plot can fail when strange error message if you trying to plot only few points. The problem is the density calculation which anyway does not make sense for few points. The solution in this PR is to change the default behavior of show_density if number of points is less than 200 to False.

Note that if user manually sets show_density=True or a very high number of bins are used, the same error could still occur. So it would be even better to catch the problem in the code and issue a better error message that could help the user.

[jsmariegaard]

Closes #381

[ecomodeller]

Sort of related to this PR

Colorbar related to density is only shown is show_points=True

Not by default?

[ecomodeller]

Using show_density=True has no effect if show_points=False, I apparently don't understand how this is supposed to work.😐

[jsmariegaard]

Using show_density=True has no effect if show_points=False, I apparently don't understand how this is supposed to work.😐

density is a feature (color) of the points, it's "point density" (implicitly) - so maybe that combination should raise an error 🤔

[jsmariegaard]

@ecomodeller - why are the type checking suddenly failing... ?

mypy modelskill/ --config-file pyproject.toml
modelskill/metrics.py:229: error: Need type annotation for "residual" [var-annotated]
modelskill/metrics.py:231: error: "Never" has no attribute "mean" [attr-defined]
modelskill/metrics.py:341: error: Need type annotation for "residual" [var-annotated]
modelskill/metrics.py:466: error: Need type annotation for "residual" [var-annotated]
modelskill/metrics.py:467: error: "Never" has no attribute "mean" [attr-defined]
modelskill/metrics.py:664: error: Need type annotation for "residual" [var-annotated]
modelskill/metrics.py:894: error: Need type annotation for "resi" [var-annotated]
modelskill/metrics.py:895: error: Incompatible types in assignment (expression has type "int", variable has type "Never") [assignment]
modelskill/timeseries/_track.py:133: error: Argument "keep" to "drop_duplicates" of "Dataset" has incompatible type "Union[bool, str]"; expected "Literal['first', 'last', False]" [arg-type]
modelskill/timeseries/_point.py:89: error: Argument 1 to "Dataset" has incompatible type "DataArray"; expected "Union[Mapping[str, DataArray], Sequence[DataArray], Sequence[Any]]" [arg-type]
modelskill/model/dfsu.py:58: error: Incompatible types in assignment (expression has type "Union[Dfs0, Dfs1, Dfs2, Dfs3, Dfsu2DH, Dfsu2DV, Dfsu3D, DfsuSpectral, Mesh]", variable has type "Union[str, Path, Dfsu2DH, Dataset, DataArray]") [assignment]
modelskill/model/dfsu.py:77: error: Item "str" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr]
modelskill/model/dfsu.py:77: error: Item "Path" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr]
modelskill/model/dfsu.py:79: error: Item "str" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr]
modelskill/model/dfsu.py:79: error: Item "Path" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr]
modelskill/model/dfsu.py:88: error: Incompatible types in assignment (expression has type "Union[str, Path, Dataset]", variable has type "Union[Dfsu2DH, Dataset]") [assignment]
modelskill/model/dfsu.py:259: error: Argument "method" to "extract_track" of "Dfsu2DH" has incompatible type "str"; expected "Literal['nearest', 'inverse_distance']" [arg-type]
modelskill/model/dfsu.py:263: error: Argument "method" to "extract_track" of "Dataset" has incompatible type "str"; expected "Literal['nearest', 'inverse_distance']" [arg-type]
Found 18 errors in 4 files (checked 40 source files)
make: *** [Makefile:18: typecheck] Error 1
Error: Process completed with exit code 2.

[ecomodeller]

@ecomodeller - why are the type checking suddenly failing... ?

mypy modelskill/ --config-file pyproject.toml modelskill/metrics.py:229: error: Need type annotation for "residual" [var-annotated] modelskill/metrics.py:231: error: "Never" has no attribute "mean" [attr-defined] modelskill/metrics.py:341: error: Need type annotation for "residual" [var-annotated] modelskill/metrics.py:466: error: Need type annotation for "residual" [var-annotated] modelskill/metrics.py:467: error: "Never" has no attribute "mean" [attr-defined] modelskill/metrics.py:664: error: Need type annotation for "residual" [var-annotated] modelskill/metrics.py:894: error: Need type annotation for "resi" [var-annotated] modelskill/metrics.py:895: error: Incompatible types in assignment (expression has type "int", variable has type "Never") [assignment] modelskill/timeseries/_track.py:133: error: Argument "keep" to "drop_duplicates" of "Dataset" has incompatible type "Union[bool, str]"; expected "Literal['first', 'last', False]" [arg-type] modelskill/timeseries/_point.py:89: error: Argument 1 to "Dataset" has incompatible type "DataArray"; expected "Union[Mapping[str, DataArray], Sequence[DataArray], Sequence[Any]]" [arg-type] modelskill/model/dfsu.py:58: error: Incompatible types in assignment (expression has type "Union[Dfs0, Dfs1, Dfs2, Dfs3, Dfsu2DH, Dfsu2DV, Dfsu3D, DfsuSpectral, Mesh]", variable has type "Union[str, Path, Dfsu2DH, Dataset, DataArray]") [assignment] modelskill/model/dfsu.py:77: error: Item "str" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr] modelskill/model/dfsu.py:77: error: Item "Path" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr] modelskill/model/dfsu.py:79: error: Item "str" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr] modelskill/model/dfsu.py:79: error: Item "Path" of "Union[str, Path, Dataset]" has no attribute "items" [union-attr] modelskill/model/dfsu.py:88: error: Incompatible types in assignment (expression has type "Union[str, Path, Dataset]", variable has type "Union[Dfsu2DH, Dataset]") [assignment] modelskill/model/dfsu.py:259: error: Argument "method" to "extract_track" of "Dfsu2DH" has incompatible type "str"; expected "Literal['nearest', 'inverse_distance']" [arg-type] modelskill/model/dfsu.py:263: error: Argument "method" to "extract_track" of "Dataset" has incompatible type "str"; expected "Literal['nearest', 'inverse_distance']" [arg-type] Found 18 errors in 4 files (checked 40 source files) make: *** [Makefile:18: typecheck] Error 1 Error: Process completed with exit code 2.

The ones related to MIKE IO fails because the last release includes a py.typed file. See PR #470

The ones related to NumPy is likely related to some changes in the last release two days ago.https://pypi.org/project/numpy/2.2.0/

Seems like we should be using numpy.typing.NDArray instead of numpy.ndarray to make mypy happy.

[jsmariegaard]

Hmm, but that seems like something that should be fixed in another PR.

I looked at https://numpy.org/doc/stable/reference/typing.html and consider if we should actually use np.typing.ArrayLike?

[jsmariegaard]

@ecomodeller, I have updated the code to consistently show the colorbar when relevant (referring to your comment and screenshots above). Could you please review?