`dask.datasets.timeseries`: dask-expr does not support the DataFrameIOFunction protocol

Currently working on getting distributed CI green again (https://github.com/dask/distributed/actions/runs/8187119566/job/22386948701?pr=8549#step:19:7242) there are a number of these UserWarnings:

> UserWarning: dask_expr does not support the DataFrameIOFunction protocol for column projection. To enable column projection, please ensure that the signature of `func` includes a `columns=` keyword argument instead.

---

Example w/ traceback

```
@pytest.mark.skipif(MACOS, reason="dask/distributed#8075")
@pytest.mark.parametrize(
"Worker", [Worker, pytest.param(Nanny, marks=[pytest.mark.slow])]
)
@gen_test()
async def test_file_descriptors_dont_leak(Worker):
pytest.importorskip("pandas")
> df = dask.datasets.timeseries(freq="10s", dtypes={"x": int, "y": float})

distributed/tests/test_client.py:6455:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
../../mambaforge/envs/distributed/lib/python3.10/site-packages/dask/datasets.py:63: in timeseries
return make_timeseries(
../../mambaforge/envs/distributed/lib/python3.10/site-packages/dask/dataframe/io/demo.py:434: in make_timese
return from_map(
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

func = , args = None
meta = x y
timestamp
2000-01-01 1035 0.431179
divisions = [Timestamp('2000-01-01 00:00:00'), Timestamp('2000-01-02 00:00:00'), Timestamp('2000-01-03 00:00
label = 'make-timeseries', enforce_metadata = False
iterables = [[([Timestamp('2000-01-01 00:00:00'), Timestamp('2000-01-02 00:00:00')], 261786203), ([Timestamp
kwargs = {}, DataFrameIOFunction =
FromMap =
FromMapProjectable = , lengths = {30}, i = 0

def from_map(
func,
*iterables,
args=None,
meta=no_default,
divisions=None,
label=None,
enforce_metadata=False,
**kwargs,
):
"""Create a dask-expr collection from a custom function map

NOTE: The underlying ``Expr`` object produced by this API
will support column projection (via ``simplify``) if
the ``func`` argument has "columns" in its signature.

"""
from dask.dataframe.io.utils import DataFrameIOFunction

from dask_expr.io import FromMap, FromMapProjectable

if "token" in kwargs:
# This option doesn't really make sense in dask-expr
raise NotImplementedError("dask_expr does not support a token argument.")

lengths = set()
iterables = list(iterables)
for i, iterable in enumerate(iterables):
if not isinstance(iterable, Iterable):
raise ValueError(
f"All elements of `iterables` must be Iterable, got {type(iterable)}"
)
try:
lengths.add(len(iterable))
except (AttributeError, TypeError):
iterables[i] = list(iterable)
lengths.add(len(iterables[i]))
if len(lengths) == 0:
raise ValueError("`from_map` requires at least one Iterable input")
elif len(lengths) > 1:
raise ValueError("All `iterables` must have the same length")
if lengths == {0}:
raise ValueError("All `iterables` must have a non-zero length")

# Check if `func` supports column projection
allow_projection = False
columns_arg_required = False
if param := inspect.signature(func).parameters.get("columns", None):
allow_projection = True
columns_arg_required = param.default is param.empty
if meta is no_default and columns_arg_required:
raise TypeError(
"Argument `func` of `from_map` has a required `columns` "
" parameter and not `meta` provided."
"Either provide `meta` yourself or make `columns` an optional argument."
)
elif isinstance(func, DataFrameIOFunction):
> warnings.warn(
"dask_expr does not support the DataFrameIOFunction "
"protocol for column projection. To enable column "
"projection, please ensure that the signature of `func` "
"includes a `columns=` keyword argument instead."
)
E UserWarning: dask_expr does not support the DataFrameIOFunction protocol for column projection. To enable column projection, please ensure that the signature of `func` includes a `columns=` keyword argument instead.
```

Start with dask/dataframe/io/demo.py, the MakeDataframePart used by dask.datasets.timeseries, and the from_map path shown in the traceback. Reproduce the warning through the distributed test case, then verify that the DataFrameIOFunction protocol is handled without the UserWarning and that the relevant test still passes.

Written by the indexing model from the issue text.