Merge pull request #929 from static-frame/928/mypy-110

Update `mypy==1.10`, `pyright==1.1.361`

requirements-dev-3_11.txt

@@ -21,8 +21,8 @@ msgpack-numpy==0.4.8
 visidata==2.11
 frame-fixtures==0.2.1
 
-mypy==1.9.0
-pyright==1.1.331
+mypy==1.10.0
+pyright==1.1.361
 pylint==2.17.5
 isort==5.12.0
 memray==1.3.1 # only linux

requirements-dev-3_12.txt

@@ -21,8 +21,8 @@ msgpack-numpy==0.4.8
 visidata==2.11
 frame-fixtures==0.2.1
 
-mypy==1.9.0
-pyright==1.1.331
+mypy==1.10.0
+pyright==1.1.361
 pylint==2.17.5
 isort==5.12.0
 memray==1.11.0 # only linux

static_frame/core/display.py

@@ -334,7 +334,7 @@ def type_attributes(
         '''
         Return the `type_input` as a string, applying delimters to either numpy dtypes or Python classes.
         '''
-        type_ref: THeaderSpecifier
+        type_ref: tp.Union[type, TDtypeAny]
         if isinstance(type_input, np.dtype):
             type_str = str(type_input)
             type_ref = type_input

static_frame/core/frame.py

@@ -5204,7 +5204,9 @@ def _extract_iloc(self, key: TILocSelectorCompound) -> tp.Any: # pyright: ignore
         return self._extract(key)
 
     def _compound_loc_to_iloc(self,
-            key: TLocSelectorCompound) -> TILocSelectorCompound:
+            key: TLocSelectorCompound,
+            # ) -> TILocSelectorCompound:
+            ) -> tp.Tuple[TILocSelector, TILocSelector]:
         '''
         Given a compound iloc key, return a tuple of row, column keys. Assumes the first argument is always a row extractor.
         '''
@@ -5219,7 +5221,8 @@ def _compound_loc_to_iloc(self,
         return iloc_row_key, iloc_column_key
 
     def _extract_loc(self, key: TLocSelectorCompound) -> tp.Any:
-        return self._extract(*self._compound_loc_to_iloc(key))
+        r, c = self._compound_loc_to_iloc(key)
+        return self._extract(r, c)
 
     def _extract_loc_columns(self, key: TLocSelector) -> TFrameOrSeries:
         '''Alternate extract of a columns only selection.

static_frame/core/index.py

@@ -345,6 +345,7 @@ def __init__(self,
 
         self._name = None if name is NAME_DEFAULT else name_filter(name) # pyright: ignore
 
+        size: int
         if self._map is None: # if _map not shared from another Index
             if not loc_is_iloc:
                 if isinstance(labels, str):
@@ -355,7 +356,7 @@ def __init__(self,
                 except NonUniqueError: # Automap will raise ValueError of non-unique values are encountered
                     raise self._error_init_index_non_unique(labels) from None
                 # must take length after map as might be iterator
-                size = len(self._map)
+                size = len(self._map) # pyright: ignore
             else:
                 # if loc_is_iloc, labels must be positions and we assume that internal clients that provided loc_is_iloc will not give a generator
                 size = len(labels) #type: ignore

static_frame/core/store_xlsx.py

@@ -413,16 +413,16 @@ def read_many(self,
                 ws = wb[label_encoded] # pyright: ignore
                 name = label # set name to the un-encoded hashable
 
-            if ws.max_column <= 1 or ws.max_row <= 1:
+            if ws.max_column <= 1 or ws.max_row <= 1: # pyright: ignore
                 # https://openpyxl.readthedocs.io/en/stable/optimized.html
                 # says that some clients might not report correct dimensions
-                ws.calculate_dimension()
+                ws.calculate_dimension() # pyright: ignore
 
-            max_column = ws.max_column
-            max_row = ws.max_row
+            max_column: int = ws.max_column # pyright: ignore
+            max_row: int = ws.max_row # pyright: ignore
 
             # adjust for downward shift for skipping header, then reduce for footer; at this value and beyond we stop
-            last_row_count = max_row - skip_header - skip_footer
+            last_row_count: int = max_row - skip_header - skip_footer
 
             index_values: tp.List[tp.Any] = []
             columns_values: tp.List[tp.Any] = []
@@ -433,7 +433,7 @@ def read_many(self,
                 mask = np.full((last_row_count, max_column), False)
 
             for row_count, row in enumerate(
-                    ws.iter_rows(max_row=max_row), start=-skip_header):
+                    ws.iter_rows(max_row=max_row), start=-skip_header): # pyright: ignore
                 if row_count < 0:
                     continue # due to skip header; preserves comparison to columns_depth
                 if row_count >= last_row_count:

static_frame/core/type_clinic.py

@@ -1476,7 +1476,7 @@ class CallGuard:
 
     @tp.overload
     @staticmethod
-    def check(func: TVFunc) -> TVFunc: ...
+    def check(func: TVFunc, /) -> TVFunc: ...
 
     @tp.overload
     @staticmethod
@@ -1512,7 +1512,7 @@ def wrapper(*args: tp.Any, **kwargs: tp.Any) -> tp.Any:
 
     @tp.overload
     @staticmethod
-    def warn(func: TVFunc) -> TVFunc: ...
+    def warn(func: TVFunc, /) -> TVFunc: ...
 
     @tp.overload
     @staticmethod

static_frame/test/typing/test_frame.py

@@ -164,13 +164,14 @@ def test_frame_astype_a() -> None:
 
 
 #-------------------------------------------------------------------------------
+h1: tp.TypeAlias = sf.Frame[sf.Index[np.int64], sf.Index[np.str_], np.int64, np.int64, np.bool_]
+
 def test_frame_interface_a() -> None:
     records = (
             (1, 2, False),
             (30, 34,True),
             (54, 95, False),
             )
-    h1 = sf.Frame[sf.Index[np.int64], sf.Index[np.str_], np.int64, np.int64, np.bool_]
     f: h1 = sf.Frame.from_records(records,
         columns=('a', 'b', 'c'),
         index=sf.Index((10, 20, 30), dtype=np.int64),
@@ -190,6 +191,10 @@ def proc1(f: h1) -> sf.Series[sf.Index[np.int64], np.int64]:
 
     # s2 = proc2(f) # pyright: Type parameter "TVDtypes@Frame" is invariant, but "*tuple[int_, int_, bool_]" is not the same as "*tuple[int_, bool_]" (reportGeneralTypeIssues)
 
+hf1: tp.TypeAlias = sf.Frame[sf.IndexDate, sf.Index[np.str_], np.int_, np.int_, np.bool_]
+hs: tp.TypeAlias = sf.Series[sf.IndexDate, np.int_]
+hf2: tp.TypeAlias = sf.Frame[sf.Index[np.int_], sf.Index[np.str_], np.int_, np.int_, np.bool_]
+hf3: tp.TypeAlias = sf.Frame[sf.IndexDate, sf.Index[np.str_], np.int_, np.bool_]
 
 def test_frame_interface_b() -> None:
 
@@ -198,8 +203,6 @@ def test_frame_interface_b() -> None:
             (30, 34, False),
             (54, 95, True),
             )
-    hf1 = sf.Frame[sf.IndexDate, sf.Index[np.str_], np.int_, np.int_, np.bool_]
-    hs = sf.Series[sf.IndexDate, np.int_]
 
     f1: hf1 = sf.Frame.from_records(records, columns=('a', 'b', 'c'), index=sf.IndexDate(('2022-01-03', '2022-02-05', '2018-04-02')))
 
@@ -211,7 +214,6 @@ def proc(f: hf1) -> hs:
 
     # if we define a Frame with a different index type, we can statically check it
 
-    hf2 = sf.Frame[sf.Index[np.int_], sf.Index[np.str_], np.int_, np.int_, np.bool_]
     f2: hf2 = sf.Frame.from_records(records, columns=('a', 'b', 'c'))
 
     # s = proc(f2)  # pyright: error: Argument of type "Frame[Index[int_], Index[str_], int_, int_, bool_]" cannot be assigned to parameter "f" of type "Frame[IndexDate, Index[str_], int_, int_, bool_]" in function "proc"
@@ -220,14 +222,36 @@ def proc(f: hf1) -> hs:
 
     # if we define a Frame with different column typing, we can statically check it
 
-    hf3 = sf.Frame[sf.IndexDate, sf.Index[np.str_], np.int_, np.bool_]
     f3: hf3 = sf.Frame.from_records((r[1:] for r in records), columns=('b', 'c'), index=sf.IndexDate(('2022-01-03', '2022-02-05', '2018-04-02')))
 
     # s = proc(f3) #pyright: error: Argument of type "Frame[IndexDate, Index[str_], int_, bool_]" cannot be assigned to parameter "f" of type "Frame[IndexDate, Index[str_], int_, int_, bool_]" in function "proc"
     # "Frame[IndexDate, Index[str_], int_, bool_]" is incompatible with "Frame[IndexDate, Index[str_], int_, int_, bool_]"
     #   Type parameter "TVDtypes@Frame" is invariant, but "*tuple[int_, bool_]" is not the same as "*tuple[int_, int_, bool_]" (reportGeneralTypeIssues)
 
 
+h10: tp.TypeAlias = sf.Frame[sf.IndexDate,
+        sf.Index[np.str_],
+        np.int_,
+        np.bool_]
+
+h20: tp.TypeAlias = sf.Frame[sf.IndexDate,
+        sf.Index[np.str_],
+        np.int_,
+        np.int_,
+        np.int_,
+        np.bool_]
+h30: tp.TypeAlias = sf.Frame[sf.IndexDate,
+        sf.Index[np.str_],
+        np.bool_,
+        np.int_,
+        ]
+
+hflex: tp.TypeAlias = sf.Frame[sf.IndexDate,
+        sf.Index[np.str_],
+        tp.Unpack[tp.Tuple[np.int_, ...]],
+        np.bool_]
+
+
 def test_frame_interface_c() -> None:
 
 
@@ -247,32 +271,10 @@ def test_frame_interface_c() -> None:
             (True, 3),
             )
 
-    h1 = sf.Frame[sf.IndexDate,
-            sf.Index[np.str_],
-            np.int_,
-            np.bool_]
-
-    h2 = sf.Frame[sf.IndexDate,
-            sf.Index[np.str_],
-            np.int_,
-            np.int_,
-            np.int_,
-            np.bool_]
-    h3 = sf.Frame[sf.IndexDate,
-            sf.Index[np.str_],
-            np.bool_,
-            np.int_,
-            ]
-
     index = sf.IndexDate(('2022-01-03', '2022-02-05', '2018-04-02'))
-    f1: h1 = sf.Frame.from_records(records1, columns=('a', 'b'), index=index)
-    f2: h2 = sf.Frame.from_records(records2, columns=('a', 'b', 'c', 'd'), index=index)
-    f3: h3 = sf.Frame.from_records(records3, columns=('a', 'd'), index=index)
-
-    hflex = sf.Frame[sf.IndexDate,
-            sf.Index[np.str_],
-            tp.Unpack[tp.Tuple[np.int_, ...]],
-            np.bool_]
+    f1: h10 = sf.Frame.from_records(records1, columns=('a', 'b'), index=index)
+    f2: h20 = sf.Frame.from_records(records2, columns=('a', 'b', 'c', 'd'), index=index)
+    f3: h30 = sf.Frame.from_records(records3, columns=('a', 'd'), index=index)
 
     fflex1: hflex = f1
     fflex2: hflex = f2
@@ -284,11 +286,9 @@ def test_frame_interface_c() -> None:
 
 def test_frame_type_var_tuple_a() ->  None:
     records = ((1, 3, True), (3, 8, True),)
-    h1 = sf.Frame[sf.IndexDate,
-            sf.Index[np.str_],]
     index = sf.IndexDate(('2022-01-03', '2018-04-02'))
     # NOTE: this works because of default of TypeVarTuple
-    f: h1 = sf.Frame.from_records(records,
+    f:sf.Frame[sf.IndexDate, sf.Index[np.str_],] = sf.Frame.from_records(records,
             columns=('a', 'b', 'c'),
             index=index,
             )

static_frame/test/typing/test_series.py

@@ -54,13 +54,13 @@ def test_series_he_iloc_a() -> None:
     v3: sf.SeriesHE[sf.Index[np.str_], np.int64] = s.iloc[1:]
     assert len(v3) == 2
 
+TSeries0: tp.TypeAlias = sf.Series[sf.Index[np.str_], np.int64]
 
 def test_series_drop() -> None:
-    TSeries = sf.Series[sf.Index[np.str_], np.int64]
-    s1: TSeries = sf.Series((10, 20, 30), index=('a', 'b', 'c'))
-    s2: TSeries = s1.drop['b'] # dropping always returns a series
+    s1: TSeries0 = sf.Series((10, 20, 30), index=('a', 'b', 'c'))
+    s2: TSeries0 = s1.drop['b'] # dropping always returns a series
 
-    def proc1(x: TSeries) -> TSeries:
+    def proc1(x: TSeries0) -> TSeries0:
         return x.dropna()
 
     y1 = proc1(s1.drop['b'])
@@ -73,9 +73,10 @@ def proc2(x: sf.Series[sf.Index[np.str_], np.str_]) -> sf.Series[sf.Index[np.str
 
 
 
+TSeries1: tp.TypeAlias = sf.Series[sf.IndexDate, np.float64]
+TSeries2: tp.TypeAlias = sf.Series[sf.Index[np.int64], np.float64]
+
 def test_series_ih1() -> None:
-    TSeries1 = sf.Series[sf.IndexDate, np.float64]
-    TSeries2 = sf.Series[sf.Index[np.int64], np.float64]
 
     s1: TSeries1 = sf.Series((10, 20, np.nan), index=sf.IndexDate(('2022-01-01', '2022-01-02', '2022-01-03')))