CLN, STYLE: remove unused variables and imports in Cython files (#48290)

* remove unused variables and imports

* wip

* bump to 0.1.4

.pre-commit-config.yaml

@@ -26,6 +26,10 @@ repos:
     hooks:
     -   id: codespell
         types_or: [python, rst, markdown]
+-   repo: https://github.com/MarcoGorelli/cython-lint
+    rev: v0.1.4
+    hooks:
+    -   id: cython-lint
 -   repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.3.0
     hooks:

pandas/_libs/algos.pyx

@@ -14,18 +14,12 @@ import numpy as np
 
 cimport numpy as cnp
 from numpy cimport (
-    NPY_COMPLEX64,
-    NPY_COMPLEX128,
-    NPY_FLOAT32,
     NPY_FLOAT64,
     NPY_INT8,
     NPY_INT16,
     NPY_INT32,
     NPY_INT64,
     NPY_OBJECT,
-    NPY_UINT8,
-    NPY_UINT16,
-    NPY_UINT32,
     NPY_UINT64,
     float32_t,
     float64_t,

pandas/_libs/groupby.pyx

@@ -17,14 +17,10 @@ from numpy cimport (
     float32_t,
     float64_t,
     int8_t,
-    int16_t,
-    int32_t,
     int64_t,
     intp_t,
     ndarray,
     uint8_t,
-    uint16_t,
-    uint32_t,
     uint64_t,
 )
 from numpy.math cimport NAN
@@ -38,7 +34,6 @@ from pandas._libs.algos cimport (
 )
 
 from pandas._libs.algos import (
-    ensure_platform_int,
     groupsort_indexer,
     rank_1d,
     take_2d_axis1_bool_bool,

pandas/_libs/hashing.pyx

@@ -13,7 +13,6 @@ from numpy cimport (
     import_array,
     ndarray,
     uint8_t,
-    uint32_t,
     uint64_t,
 )
 
@@ -158,7 +157,6 @@ cdef uint64_t low_level_siphash(uint8_t* data, size_t datalen,
     cdef int i
     cdef uint8_t* end = data + datalen - (datalen % sizeof(uint64_t))
     cdef int left = datalen & 7
-    cdef int left_byte
     cdef int cROUNDS = 2
     cdef int dROUNDS = 4
 

pandas/_libs/hashtable.pyx

@@ -15,13 +15,7 @@ from libc.stdlib cimport (
 import numpy as np
 
 cimport numpy as cnp
-from numpy cimport (
-    float64_t,
-    ndarray,
-    uint8_t,
-    uint32_t,
-)
-from numpy.math cimport NAN
+from numpy cimport ndarray
 
 cnp.import_array()
 
@@ -37,9 +31,6 @@ from pandas._libs.khash cimport (
     kh_needed_n_buckets,
     kh_python_hash_equal,
     kh_python_hash_func,
-    kh_str_t,
-    khcomplex64_t,
-    khcomplex128_t,
     khiter_t,
 )
 from pandas._libs.missing cimport checknull

pandas/_libs/index.pyx

@@ -4,17 +4,10 @@ import numpy as np
 
 cimport numpy as cnp
 from numpy cimport (
-    float32_t,
-    float64_t,
-    int8_t,
-    int16_t,
-    int32_t,
     int64_t,
     intp_t,
     ndarray,
     uint8_t,
-    uint16_t,
-    uint32_t,
     uint64_t,
 )
 
@@ -35,7 +28,6 @@ from pandas._libs import (
 
 from pandas._libs.lib cimport eq_NA_compat
 from pandas._libs.missing cimport (
-    C_NA as NA,
     checknull,
     is_matching_na,
 )

pandas/_libs/internals.pyx

@@ -226,7 +226,7 @@ cdef class BlockPlacement:
         """
         cdef:
             slice nv, s = self._ensure_has_slice()
-            Py_ssize_t other_int, start, stop, step
+            Py_ssize_t start, stop, step
             ndarray[intp_t, ndim=1] newarr
 
         if s is not None:

pandas/_libs/interval.pyx

@@ -1,4 +1,3 @@
-import inspect
 import numbers
 from operator import (
     le,
@@ -13,15 +12,7 @@ from cpython.datetime cimport (
 import_datetime()
 
 cimport cython
-from cpython.object cimport (
-    Py_EQ,
-    Py_GE,
-    Py_GT,
-    Py_LE,
-    Py_LT,
-    Py_NE,
-    PyObject_RichCompare,
-)
+from cpython.object cimport PyObject_RichCompare
 from cython cimport Py_ssize_t
 
 import numpy as np
@@ -31,9 +22,7 @@ from numpy cimport (
     NPY_QUICKSORT,
     PyArray_ArgSort,
     PyArray_Take,
-    float32_t,
     float64_t,
-    int32_t,
     int64_t,
     ndarray,
     uint64_t,

pandas/_libs/lib.pyx

@@ -47,7 +47,6 @@ from numpy cimport (
     PyArray_IterNew,
     complex128_t,
     flatiter,
-    float32_t,
     float64_t,
     int64_t,
     intp_t,

pandas/_libs/sparse.pyx

@@ -3,10 +3,8 @@ import numpy as np
 
 cimport numpy as cnp
 from numpy cimport (
-    float32_t,
     float64_t,
     int8_t,
-    int16_t,
     int32_t,
     int64_t,
     ndarray,
@@ -127,7 +125,7 @@ cdef class IntIndex(SparseIndex):
 
     cpdef IntIndex intersect(self, SparseIndex y_):
         cdef:
-            Py_ssize_t out_length, xi, yi = 0, result_indexer = 0
+            Py_ssize_t xi, yi = 0, result_indexer = 0
             int32_t xind
             ndarray[int32_t, ndim=1] xindices, yindices, new_indices
             IntIndex y
@@ -205,7 +203,7 @@ cdef class IntIndex(SparseIndex):
         Vectorized lookup, returns ndarray[int32_t]
         """
         cdef:
-            Py_ssize_t n, i, ind_val
+            Py_ssize_t n
             ndarray[int32_t, ndim=1] inds
             ndarray[uint8_t, ndim=1, cast=True] mask
             ndarray[int32_t, ndim=1] masked
@@ -232,7 +230,7 @@ cdef class IntIndex(SparseIndex):
 
 cpdef get_blocks(ndarray[int32_t, ndim=1] indices):
     cdef:
-        Py_ssize_t init_len, i, npoints, result_indexer = 0
+        Py_ssize_t i, npoints, result_indexer = 0
         int32_t block, length = 1, cur, prev
         ndarray[int32_t, ndim=1] locs, lens
 
@@ -606,7 +604,7 @@ cdef class BlockUnion(BlockMerge):
         cdef:
             ndarray[int32_t, ndim=1] xstart, xend, ystart
             ndarray[int32_t, ndim=1] yend, out_bloc, out_blen
-            int32_t nstart, nend, diff
+            int32_t nstart, nend
             Py_ssize_t max_len, result_indexer = 0
 
         xstart = self.xstart
@@ -659,7 +657,7 @@ cdef class BlockUnion(BlockMerge):
         """
         cdef:
             ndarray[int32_t, ndim=1] xstart, xend, ystart, yend
-            int32_t xi, yi, xnblocks, ynblocks, nend
+            int32_t xi, yi, ynblocks, nend
 
         if mode != 0 and mode != 1:
             raise Exception('Mode must be 0 or 1')

pandas/_libs/tslib.pyx

@@ -474,12 +474,10 @@ cpdef array_to_datetime(
         Py_ssize_t i, n = len(values)
         object val, tz
         ndarray[int64_t] iresult
-        ndarray[object] oresult
         npy_datetimestruct dts
         NPY_DATETIMEUNIT out_bestunit
         bint utc_convert = bool(utc)
         bint seen_integer = False
-        bint seen_string = False
         bint seen_datetime = False
         bint seen_datetime_offset = False
         bint is_raise = errors=='raise'
@@ -489,7 +487,7 @@ cpdef array_to_datetime(
         _TSObject _ts
         int64_t value
         int out_local = 0, out_tzoffset = 0
-        float offset_seconds, tz_offset
+        float tz_offset
         set out_tzoffset_vals = set()
         bint string_to_dts_failed
         datetime py_dt
@@ -572,7 +570,6 @@ cpdef array_to_datetime(
 
                 elif isinstance(val, str):
                     # string
-                    seen_string = True
                     if type(val) is not str:
                         # GH#32264 np.str_ object
                         val = str(val)

pandas/_libs/tslibs/conversion.pyx

@@ -1,20 +1,14 @@
 import inspect
-
-cimport cython
-
 import warnings
 
 import numpy as np
 
 from pandas.util._exceptions import find_stack_level
 
 cimport numpy as cnp
-from cpython.object cimport PyObject
 from numpy cimport (
     int32_t,
     int64_t,
-    intp_t,
-    ndarray,
 )
 
 cnp.import_array()
@@ -42,33 +36,25 @@ from pandas._libs.tslibs.dtypes cimport (
 from pandas._libs.tslibs.np_datetime cimport (
     NPY_DATETIMEUNIT,
     NPY_FR_ns,
-    astype_overflowsafe,
     check_dts_bounds,
     dtstruct_to_dt64,
     get_datetime64_unit,
     get_datetime64_value,
     get_implementation_bounds,
-    get_unit_from_dtype,
     npy_datetime,
     npy_datetimestruct,
     npy_datetimestruct_to_datetime,
     pandas_datetime_to_datetimestruct,
-    pydatetime_to_dt64,
     pydatetime_to_dtstruct,
     string_to_dts,
 )
 
-from pandas._libs.tslibs.np_datetime import (
-    OutOfBoundsDatetime,
-    OutOfBoundsTimedelta,
-)
+from pandas._libs.tslibs.np_datetime import OutOfBoundsDatetime
 
 from pandas._libs.tslibs.timezones cimport (
     get_utcoffset,
     is_utc,
     maybe_get_tz,
-    tz_compare,
-    utc_pytz as UTC,
 )
 from pandas._libs.tslibs.util cimport (
     is_datetime64_object,
@@ -82,7 +68,6 @@ from pandas._libs.tslibs.nattype cimport (
     NPY_NAT,
     c_NaT as NaT,
     c_nat_strings as nat_strings,
-    checknull_with_nat,
 )
 from pandas._libs.tslibs.tzconversion cimport (
     Localizer,

pandas/_libs/tslibs/dtypes.pyx

@@ -1,7 +1,5 @@
 # period frequency constants corresponding to scikits timeseries
 # originals
-cimport cython
-
 from enum import Enum
 
 from pandas._libs.tslibs.np_datetime cimport (

pandas/_libs/tslibs/fields.pyx

@@ -37,15 +37,12 @@ from pandas._libs.tslibs.ccalendar cimport (
     get_iso_calendar,
     get_lastbday,
     get_week_of_year,
-    is_leapyear,
     iso_calendar_t,
-    month_offset,
 )
 from pandas._libs.tslibs.nattype cimport NPY_NAT
 from pandas._libs.tslibs.np_datetime cimport (
     NPY_DATETIMEUNIT,
     NPY_FR_ns,
-    get_unit_from_dtype,
     npy_datetimestruct,
     pandas_datetime_to_datetimestruct,
     pandas_timedelta_to_timedeltastruct,

pandas/_libs/tslibs/nattype.pyx

@@ -15,10 +15,6 @@ from cpython.datetime cimport (
 import_datetime()
 from cpython.object cimport (
     Py_EQ,
-    Py_GE,
-    Py_GT,
-    Py_LE,
-    Py_LT,
     Py_NE,
     PyObject_RichCompare,
 )

pandas/_libs/tslibs/offsets.pyx

@@ -1,5 +1,4 @@
 import inspect
-import operator
 import re
 import time
 import warnings

pandas/_libs/tslibs/parsing.pyx

@@ -503,7 +503,7 @@ cdef inline object _parse_dateabbr_string(object date_string, datetime default,
     cdef:
         object ret
         # year initialized to prevent compiler warnings
-        int year = -1, quarter = -1, month, mnum
+        int year = -1, quarter = -1, month
         Py_ssize_t date_len
 
     # special handling for possibilities eg, 2Q2005, 2Q05, 2005Q1, 05Q1

pandas/_libs/tslibs/period.pyx

@@ -25,7 +25,6 @@ cimport cython
 from cpython.datetime cimport (
     PyDate_Check,
     PyDateTime_Check,
-    PyDelta_Check,
     datetime,
     import_datetime,
 )
@@ -50,7 +49,6 @@ from pandas._libs.missing cimport C_NA
 from pandas._libs.tslibs.np_datetime cimport (
     NPY_DATETIMEUNIT,
     NPY_FR_D,
-    NPY_FR_us,
     astype_overflowsafe,
     check_dts_bounds,
     get_timedelta64_value,
@@ -780,7 +778,7 @@ cdef int64_t get_period_ordinal(npy_datetimestruct *dts, int freq) nogil:
     """
     cdef:
         int64_t unix_date
-        int freq_group, fmonth, mdiff
+        int freq_group, fmonth
         NPY_DATETIMEUNIT unit
 
     freq_group = get_freq_group(freq)