Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: Scalar typetracer after reduction #529

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 19 additions & 5 deletions src/dask_awkward/lib/core.py
Original file line number Diff line number Diff line change
Expand Up @@ -500,19 +500,17 @@ def f(self, other):
deps = [self]
plns = [self.name]
if is_dask_collection(other):
task = (op, self.key, *other.__dask_keys__())
deps.append(other)
plns.append(other.name)
if inv:
plns.insert(0, other.name)
task = (op, *other.__dask_keys__(), self.key)
else:
plns.append(other.name)
task = (op, self.key, *other.__dask_keys__())
else:
if inv:
task = (op, other, self.key)
else:
task = (op, self.key, other)
if inv:
plns.reverse()
graph = HighLevelGraph.from_collections(
name,
layer=AwkwardMaterializedLayer(
Expand All @@ -532,6 +530,11 @@ def f(self, other):
meta = op(other, self._meta)
else:
meta = op(self._meta, other)
if meta.ndim:
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This applies to the scenario, which apparently we hadn't anticipated, that an operation on a scalar produces an array.

divisions = other.divisions if is_dask_collection(other) else [0, 1]
return new_array_object(
graph, name, meta=ak.Array(meta), divisions=divisions
)
return new_scalar_object(graph, name, meta=meta)

return f
Expand Down Expand Up @@ -570,6 +573,15 @@ def f(*args):
args = tuple(
ak.Array(arg.content) if isinstance(arg, MaybeNone) else arg for arg in args
)
args = tuple(
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here we account for getting the low-level typetracer through during column optimisation. This is the questionable part.

(
ak.Array(arg)
if isinstance(arg, ak._nplikes.typetracer.TypeTracerArray)
else arg
)
for arg in args
)

result = op(*args)
return result

Expand Down Expand Up @@ -2598,6 +2610,8 @@ def typetracer_array(a: ak.Array | Array) -> ak.Array:
behavior=a._behavior,
attrs=a._attrs,
)
elif isinstance(a, numbers.Number):
return ak.Array([a]).layout.to_typetracer()
else:
msg = (
"`a` should be an awkward array or a Dask awkward collection.\n"
Expand Down
2 changes: 0 additions & 2 deletions tests/test_core.py
Original file line number Diff line number Diff line change
Expand Up @@ -440,8 +440,6 @@ def test_typetracer_function(daa: Array) -> None:
tta = typetracer_array(aa)
assert tta is not None
assert tta.layout.form == aa.layout.form
with pytest.raises(TypeError, match="Got type <class 'int'>"):
typetracer_array(3)


def test_single_partition(ndjson_points_file: str) -> None:
Expand Down
Loading