Matthew Barber

added spaces inbetween "python -m" and package_name templates

3

I found two instances of `python -m` in docstrings generated by this cookiecutter template. Note that there's no space inbetween `-m` and ``! I have thus modified the corresponding template...

Interchange `Column.null_count` is a NumPy scalar, not a builtin `int`

1

A `PandasProtocolColumn` returns a `null_count` as a 0d integer array (specifically a NumPy scalar), as opposed to [`int` as specified in the interchange protocol](https://github.com/data-apis/dataframe-api/blob/4f7c1e0c425643d57120c5b73434d992b3e83595/protocol/dataframe_protocol.py#L216). ```python >>> from modin import pandas...

Support complex dtypes in `extra.array_api`

2

The Array API has had it's [first spec release `2021.12`](https://data-apis.org/array-api/2021.12/) a while now. The next release will add complex numbers, which would be nice to support in `extra.array_api`. We could...

In the [relevant docs page](https://hypothesis.readthedocs.io/en/latest/numpy.html#array-api), we currently exclaim: > If you want to test with CuPy, Dask, JAX, MXNet, PyTorch, TensorFlow, or Xarray, this is the extension for you! This...

[BUG-REPORT] Column names prefixed with double underscores (`"__"`) don't show in `df.get_column_names()`

1

```python >>> df = vaex.from_items(("__", np.asarray([0]))) >>> df.get_column_names() [] ``` Same goes for seemingly any string after the double underscores ```python >>> df = vaex.from_items(("__foo", np.asarray([0]))) >>> df.get_column_names() [] ```...

Interchange protocol fixes and updates

5

The interchange protocol had some spec changes recently (https://github.com/data-apis/dataframe-api/pull/74), so this PR namely updates vaex to conform with them. * `Column.size()` * Resolves https://github.com/vaexio/vaex/issues/2093 (was useful so I could test...

[BUG-REPORT] Interchange `Column.size` returns 0d arrays as opposed to Python `int`

1

Dataframes via the [interchange protocol](https://data-apis.org/dataframe-protocol/latest) will store return interchange columns which give an erroneous `size`. ```python >>> import vaex >>> df = vaex.from_dict({"a": ["foo", "bar"]}) >>> interchange_df = df.__dataframe__() >>>...

[BUG-REPORT] `describe_categorical` in interchange columns is a tuple, not a dict

1

In the [interchange protocol](https://data-apis.org/dataframe-protocol/latest/API.html), `describe_categorical` should return a dict (mind the spec's API type annotation is faulty), but Vaex returns a tuple https://github.com/vaexio/vaex/blob/35c250d585f889272b8ef1096de6fa5462816f52/packages/vaex-core/vaex/dataframe_protocol.py#L443 This prevents interchanging dataframes with categorical columns,...

Weird bug(?) I stumbled upon when using (string) numbers as names for categorical columns, and then trying to use the interchange protocol on it. ```python >>> df = vaex.from_items(("42", np.asarray([3,...

In the [interchange protocol](https://data-apis.org/dataframe-protocol/latest/API.html), `Column.dtype` should return an [Arrow-style](https://arrow.apache.org/docs/format/CDataInterface.html#data-type-description-format-strings) format string, but instead a NumPy-styled one is returned ```python >>> df = vaex.from_items(("foo", np.asarray([0, 1, 2], dtype="int64"))) >>> interchange_df =...