typing fixes. (#292)

* typing fixes.

* update typing ci

* more fixes

Author: dcherian

.github/workflows/ci-additional.yaml

@@ -59,7 +59,7 @@ jobs:
           environment-name: flox-tests
           init-shell: bash
           cache-environment: true
-          cache-env-key: "${{runner.os}}-${{runner.arch}}-py${{env.PYTHON_VERSION}}-${{env.TODAY}}-${{hashFiles(env.CONDA_ENV_FILE)}}"
+          cache-environment-key: "${{runner.os}}-${{runner.arch}}-py${{env.PYTHON_VERSION}}-${{env.TODAY}}-${{hashFiles(env.CONDA_ENV_FILE)}}"
           create-args: |
             python=${{ env.PYTHON_VERSION }}
 
@@ -112,7 +112,7 @@ jobs:
           environment-name: flox-tests
           init-shell: bash
           cache-environment: true
-          cache-env-key: "${{runner.os}}-${{runner.arch}}-py${{env.PYTHON_VERSION}}-${{env.TODAY}}-${{hashFiles(env.CONDA_ENV_FILE)}}"
+          cache-environment-key: "${{runner.os}}-${{runner.arch}}-py${{env.PYTHON_VERSION}}-${{env.TODAY}}-${{hashFiles(env.CONDA_ENV_FILE)}}"
           create-args: |
             python=${{ env.PYTHON_VERSION }}
       - name: Install flox

flox/core.py

@@ -54,7 +54,7 @@
     T_DuckArray = Union[np.ndarray, DaskArray]  # Any ?
     T_By = T_DuckArray
     T_Bys = tuple[T_By, ...]
-    T_ExpectIndex = Union[pd.Index]
+    T_ExpectIndex = pd.Index
     T_ExpectIndexTuple = tuple[T_ExpectIndex, ...]
     T_ExpectIndexOpt = Union[T_ExpectIndex, None]
     T_ExpectIndexOptTuple = tuple[T_ExpectIndexOpt, ...]
@@ -314,7 +314,7 @@ def invert(x) -> tuple[np.ndarray, ...]:
         items = tuple((k, set(k), v) for k, v in sorted_chunks_cohorts.items() if k)
 
         merged_cohorts = {}
-        merged_keys = set()
+        merged_keys: set[tuple] = set()
 
         # Now we iterate starting with the longest number of chunks,
         # and then merge in cohorts that are present in a subset of those chunks
@@ -1895,7 +1895,7 @@ def groupby_reduce(
     engine: T_EngineOpt = None,
     reindex: bool | None = None,
     finalize_kwargs: dict[Any, Any] | None = None,
-) -> tuple[DaskArray, Unpack[tuple[np.ndarray | DaskArray, ...]]]:  # type: ignore[misc]  # Unpack not in mypy yet
+) -> tuple[DaskArray, Unpack[tuple[np.ndarray | DaskArray, ...]]]:
     """
     GroupBy reductions using tree reductions for dask.array
 
@@ -2223,4 +2223,4 @@ def groupby_reduce(
 
     if is_bool_array and (_is_minmax_reduction(func) or _is_first_last_reduction(func)):
         result = result.astype(bool)
-    return (result, *groups)  # type: ignore[return-value]  # Unpack not in mypy yet
+    return (result, *groups)

tests/test_core.py

@@ -179,14 +179,15 @@ def test_groupby_reduce(
     elif func == "count":
         expected_result = np.array(expected, dtype=np.intp)
 
-    (result, groups) = groupby_reduce(
+    (result, *groups) = groupby_reduce(
         array,
         by,
         func=func,
         expected_groups=expected_groups,
         fill_value=123,
         engine=engine,
     )
+    (groups_array,) = groups
     # we use pd.Index(expected_groups).to_numpy() which is always int64
     # for the values in this test
     if expected_groups is None:
@@ -196,7 +197,7 @@ def test_groupby_reduce(
     else:
         g_dtype = np.int64
 
-    assert_equal(groups, np.array([0, 1, 2], g_dtype))
+    assert_equal(groups_array, np.array([0, 1, 2], g_dtype))
     assert_equal(expected_result, result)
 
 
@@ -795,11 +796,14 @@ def test_groupby_bins(chunk_labels, kwargs, chunks, engine, method) -> None:
             labels = dask.array.from_array(labels, chunks=chunks)
 
     with raise_if_dask_computes():
-        actual, groups = groupby_reduce(
+        actual, *groups = groupby_reduce(
             array, labels, func="count", fill_value=0, engine=engine, method=method, **kwargs
         )
+    (groups_array,) = groups
     expected = np.array([3, 1, 0], dtype=np.intp)
-    for left, right in zip(groups, pd.IntervalIndex.from_arrays([1, 2, 4], [2, 4, 5]).to_numpy()):
+    for left, right in zip(
+        groups_array, pd.IntervalIndex.from_arrays([1, 2, 4], [2, 4, 5]).to_numpy()
+    ):
         assert left == right
     assert_equal(actual, expected)
 
@@ -1034,13 +1038,13 @@ def test_bool_reductions(func, engine):
 def test_map_reduce_blockwise_mixed() -> None:
     t = pd.date_range("2000-01-01", "2000-12-31", freq="D").to_series()
     data = t.dt.dayofyear
-    actual, _ = groupby_reduce(
+    actual, *_ = groupby_reduce(
         dask.array.from_array(data.values, chunks=365),
         t.dt.month,
         func="mean",
         method="map-reduce",
     )
-    expected, _ = groupby_reduce(data, t.dt.month, func="mean")
+    expected, *_ = groupby_reduce(data, t.dt.month, func="mean")
     assert_equal(expected, actual)