STYLE fixup cython linting (#42142)

Author: MarcoGorelli

.pre-commit-config.yaml

@@ -44,6 +44,7 @@ repos:
             - flake8-bugbear==21.3.2
             - pandas-dev-flaker==0.2.0
     -   id: flake8
+        alias: flake8-cython
         name: flake8 (cython)
         types: [cython]
         args: [--append-config=flake8/cython.cfg]

flake8/cython.cfg

@@ -1,3 +1,17 @@
 [flake8]
 filename = *.pyx,*.pxd
-select=E501,E302,E203,E111,E114,E221,E303,E128,E231,E126,E265,E305,E301,E127,E261,E271,E129,W291,E222,E241,E123,F403,C400,C401,C402,C403,C404,C405,C406,C407,C408,C409,C410,C411
+extend_ignore=
+    # whitespace before '('
+    E211,
+    # missing whitespace around operator
+    E225,
+    # missing whitespace around arithmetic operator
+    E226,
+    # missing whitespace around bitwise or shift operator
+    E227,
+    # ambiguous variable name (# FIXME maybe this one can be fixed)
+    E741,
+    # invalid syntax
+    E999,
+    # invalid escape sequence (# FIXME maybe this one can be fixed)
+    W605,

pandas/_libs/algos.pyx

@@ -274,16 +274,22 @@ cdef inline numeric kth_smallest_c(numeric* arr, Py_ssize_t k, Py_ssize_t n) nog
         j = m
 
         while 1:
-            while arr[i] < x: i += 1
-            while x < arr[j]: j -= 1
+            while arr[i] < x:
+                i += 1
+            while x < arr[j]:
+                j -= 1
             if i <= j:
                 swap(&arr[i], &arr[j])
-                i += 1; j -= 1
+                i += 1
+                j -= 1
 
-            if i > j: break
+            if i > j:
+                break
 
-        if j < k: l = i
-        if k < i: m = j
+        if j < k:
+            l = i
+        if k < i:
+            m = j
     return arr[k]
 
 

pandas/_libs/khash.pxd

@@ -26,20 +26,20 @@ cdef extern from "khash_python.h":
         double imag
 
     bint are_equivalent_khcomplex128_t \
-    "kh_complex_hash_equal" (khcomplex128_t a, khcomplex128_t b) nogil
+        "kh_complex_hash_equal" (khcomplex128_t a, khcomplex128_t b) nogil
 
     ctypedef struct khcomplex64_t:
         float real
         float imag
 
     bint are_equivalent_khcomplex64_t \
-    "kh_complex_hash_equal" (khcomplex64_t a, khcomplex64_t b) nogil
+        "kh_complex_hash_equal" (khcomplex64_t a, khcomplex64_t b) nogil
 
     bint are_equivalent_float64_t \
-    "kh_floats_hash_equal" (float64_t a, float64_t b) nogil
+        "kh_floats_hash_equal" (float64_t a, float64_t b) nogil
 
     bint are_equivalent_float32_t \
-    "kh_floats_hash_equal" (float32_t a, float32_t b) nogil
+        "kh_floats_hash_equal" (float32_t a, float32_t b) nogil
 
     uint32_t kh_python_hash_func(object key)
     bint kh_python_hash_equal(object a, object b)

pandas/_libs/lib.pyx

@@ -2107,9 +2107,9 @@ cpdef bint is_interval_array(ndarray values):
                 return False
             elif numeric:
                 if not (
-                        util.is_float_object(val.left)
-                        or util.is_integer_object(val.left)
-                    ):
+                    util.is_float_object(val.left)
+                    or util.is_integer_object(val.left)
+                ):
                     # i.e. datetime64 or timedelta64
                     return False
             elif td64:

pandas/_libs/parsers.pyx

@@ -356,7 +356,7 @@ cdef class TextReader:
                   thousands=None,       # bytes | str
                   dtype=None,
                   usecols=None,
-                  on_bad_lines = ERROR,
+                  on_bad_lines=ERROR,
                   bint na_filter=True,
                   na_values=None,
                   na_fvalues=None,
@@ -1442,7 +1442,7 @@ cdef _categorical_convert(parser_t *parser, int64_t col,
 
             if na_filter:
                 if kh_get_str_starts_item(na_hashset, word):
-                # is in NA values
+                    # is in NA values
                     na_count += 1
                     codes[i] = NA
                     continue
@@ -1578,7 +1578,7 @@ cdef inline int _try_double_nogil(parser_t *parser,
                             strcasecmp(word, cposinfty) == 0):
                         data[0] = INF
                     elif (strcasecmp(word, cneginf) == 0 or
-                            strcasecmp(word, cneginfty) == 0 ):
+                            strcasecmp(word, cneginfty) == 0):
                         data[0] = NEGINF
                     else:
                         return 1

pandas/_libs/tslibs/conversion.pyx

@@ -200,7 +200,7 @@ cdef inline int64_t get_datetime64_nanos(object val) except? -1:
 
 @cython.boundscheck(False)
 @cython.wraparound(False)
-def ensure_datetime64ns(arr: ndarray, copy: bool=True):
+def ensure_datetime64ns(arr: ndarray, copy: bool = True):
     """
     Ensure a np.datetime64 array has dtype specifically 'datetime64[ns]'
 
@@ -260,7 +260,7 @@ def ensure_datetime64ns(arr: ndarray, copy: bool=True):
     return result
 
 
-def ensure_timedelta64ns(arr: ndarray, copy: bool=True):
+def ensure_timedelta64ns(arr: ndarray, copy: bool = True):
     """
     Ensure a np.timedelta64 array has dtype specifically 'timedelta64[ns]'
 

pandas/_libs/tslibs/np_datetime.pyx

@@ -38,7 +38,7 @@ cdef extern from "src/datetime/np_datetime.h":
     void pandas_timedelta_to_timedeltastruct(npy_timedelta val,
                                              NPY_DATETIMEUNIT fr,
                                              pandas_timedeltastruct *result
-                                            ) nogil
+                                             ) nogil
 
     npy_datetimestruct _NS_MIN_DTS, _NS_MAX_DTS
 

pandas/_libs/tslibs/offsets.pyx

@@ -1454,7 +1454,7 @@ cdef class BusinessHour(BusinessMixin):
 
     def __init__(
             self, n=1, normalize=False, start="09:00", end="17:00", offset=timedelta(0)
-        ):
+    ):
         BusinessMixin.__init__(self, n, normalize, offset)
 
         # must be validated here to equality check
@@ -3897,7 +3897,7 @@ cdef ndarray[int64_t] _shift_bdays(const int64_t[:] i8other, int periods):
     return result.base
 
 
-def shift_month(stamp: datetime, months: int, day_opt: object=None) -> datetime:
+def shift_month(stamp: datetime, months: int, day_opt: object = None) -> datetime:
     """
     Given a datetime (or Timestamp) `stamp`, an integer `months` and an
     option `day_opt`, return a new datetimelike that many months later,

pandas/_libs/tslibs/period.pyx

@@ -197,7 +197,7 @@ cdef freq_conv_func get_asfreq_func(int from_freq, int to_freq) nogil:
             return <freq_conv_func>asfreq_BtoW
         elif to_group == FR_BUS:
             return <freq_conv_func>no_op
-        elif to_group  in [FR_DAY, FR_HR, FR_MIN, FR_SEC, FR_MS, FR_US, FR_NS]:
+        elif to_group in [FR_DAY, FR_HR, FR_MIN, FR_SEC, FR_MS, FR_US, FR_NS]:
             return <freq_conv_func>asfreq_BtoDT
         else:
             return <freq_conv_func>nofunc

pandas/_libs/tslibs/strptime.pyx

@@ -199,17 +199,17 @@ def array_strptime(ndarray[object] values, object fmt, bint exact=True, errors='
                 year = int(found_dict['Y'])
             elif parse_code == 2:
                 month = int(found_dict['m'])
-            elif parse_code == 3:
             # elif group_key == 'B':
+            elif parse_code == 3:
                 month = locale_time.f_month.index(found_dict['B'].lower())
-            elif parse_code == 4:
             # elif group_key == 'b':
+            elif parse_code == 4:
                 month = locale_time.a_month.index(found_dict['b'].lower())
-            elif parse_code == 5:
             # elif group_key == 'd':
+            elif parse_code == 5:
                 day = int(found_dict['d'])
-            elif parse_code == 6:
             # elif group_key == 'H':
+            elif parse_code == 6:
                 hour = int(found_dict['H'])
             elif parse_code == 7:
                 hour = int(found_dict['I'])

pandas/_libs/tslibs/timedeltas.pyx

@@ -641,7 +641,8 @@ def _binary_op_method_timedeltalike(op, name):
             return NaT
 
         elif is_datetime64_object(other) or (
-           PyDateTime_Check(other) and not isinstance(other, ABCTimestamp)):
+            PyDateTime_Check(other) and not isinstance(other, ABCTimestamp)
+        ):
             # this case is for a datetime object that is specifically
             # *not* a Timestamp, as the Timestamp case will be
             # handled after `_validate_ops_compat` returns False below

pandas/_libs/tslibs/timestamps.pyx

@@ -1958,7 +1958,7 @@ default 'raise'
                  self.second / 3600.0 +
                  self.microsecond / 3600.0 / 1e+6 +
                  self.nanosecond / 3600.0 / 1e+9
-                ) / 24.0)
+                 ) / 24.0)
 
 
 # Aliases