Tslibs resolution (#18034)

Author: jbrockmendel

doc/source/whatsnew/v0.22.0.txt

@@ -45,6 +45,7 @@ Other API Changes
 - :class:`Timestamp` will no longer silently ignore unused or invalid ``tz`` or ``tzinfo`` keyword arguments (:issue:`17690`)
 - :class:`Timestamp` will no longer silently ignore invalid ``freq`` arguments (:issue:`5168`)
 - :class:`CacheableOffset` and :class:`WeekDay` are no longer available in the ``pandas.tseries.offsets`` module (:issue:`17830`)
+- `tseries.frequencies.get_freq_group()` and `tseries.frequencies.DAYS` are removed from the public API (:issue:`18034`)
 
 .. _whatsnew_0220.deprecations:
 

pandas/_libs/period.pyx

@@ -1,4 +1,5 @@
 # -*- coding: utf-8 -*-
+# cython: profile=False
 from datetime import datetime, date, timedelta
 import operator
 
@@ -27,28 +28,23 @@ from util cimport is_period_object, is_string_object, INT32_MIN
 
 from lib cimport is_null_datetimelike
 from pandas._libs import tslib
-from pandas._libs.tslib import Timestamp, iNaT, NaT
+from pandas._libs.tslib import Timestamp, iNaT
 from tslibs.timezones cimport (
     is_utc, is_tzlocal, get_utcoffset, get_dst_info, maybe_get_tz)
 from tslibs.timedeltas cimport delta_to_nanoseconds
 
-from tslibs.parsing import parse_time_string, NAT_SENTINEL
+from tslibs.parsing import (parse_time_string, NAT_SENTINEL,
+                            _get_rule_month, _MONTH_NUMBERS)
 from tslibs.frequencies cimport get_freq_code
-from tslibs.nattype import nat_strings
+from tslibs.resolution import resolution, Resolution
+from tslibs.nattype import nat_strings, NaT
 from tslibs.nattype cimport _nat_scalar_rules
 
 from pandas.tseries import offsets
 from pandas.tseries import frequencies
 
 cdef int64_t NPY_NAT = util.get_nat()
 
-cdef int RESO_US = frequencies.RESO_US
-cdef int RESO_MS = frequencies.RESO_MS
-cdef int RESO_SEC = frequencies.RESO_SEC
-cdef int RESO_MIN = frequencies.RESO_MIN
-cdef int RESO_HR = frequencies.RESO_HR
-cdef int RESO_DAY = frequencies.RESO_DAY
-
 cdef extern from "period_helper.h":
     ctypedef struct date_info:
         int64_t absdate
@@ -487,98 +483,10 @@ def extract_freq(ndarray[object] values):
     raise ValueError('freq not specified and cannot be inferred')
 
 
-cpdef resolution(ndarray[int64_t] stamps, tz=None):
-    cdef:
-        Py_ssize_t i, n = len(stamps)
-        pandas_datetimestruct dts
-        int reso = RESO_DAY, curr_reso
-
-    if tz is not None:
-        tz = maybe_get_tz(tz)
-        return _reso_local(stamps, tz)
-    else:
-        for i in range(n):
-            if stamps[i] == NPY_NAT:
-                continue
-            dt64_to_dtstruct(stamps[i], &dts)
-            curr_reso = _reso_stamp(&dts)
-            if curr_reso < reso:
-                reso = curr_reso
-        return reso
-
-
-cdef inline int _reso_stamp(pandas_datetimestruct *dts):
-    if dts.us != 0:
-        if dts.us % 1000 == 0:
-            return RESO_MS
-        return RESO_US
-    elif dts.sec != 0:
-        return RESO_SEC
-    elif dts.min != 0:
-        return RESO_MIN
-    elif dts.hour != 0:
-        return RESO_HR
-    return RESO_DAY
-
-cdef _reso_local(ndarray[int64_t] stamps, object tz):
-    cdef:
-        Py_ssize_t n = len(stamps)
-        int reso = RESO_DAY, curr_reso
-        ndarray[int64_t] trans, deltas, pos
-        pandas_datetimestruct dts
-
-    if is_utc(tz):
-        for i in range(n):
-            if stamps[i] == NPY_NAT:
-                continue
-            dt64_to_dtstruct(stamps[i], &dts)
-            curr_reso = _reso_stamp(&dts)
-            if curr_reso < reso:
-                reso = curr_reso
-    elif is_tzlocal(tz):
-        for i in range(n):
-            if stamps[i] == NPY_NAT:
-                continue
-            dt64_to_dtstruct(stamps[i], &dts)
-            dt = datetime(dts.year, dts.month, dts.day, dts.hour,
-                          dts.min, dts.sec, dts.us, tz)
-            delta = int(get_utcoffset(tz, dt).total_seconds()) * 1000000000
-            dt64_to_dtstruct(stamps[i] + delta, &dts)
-            curr_reso = _reso_stamp(&dts)
-            if curr_reso < reso:
-                reso = curr_reso
-    else:
-        # Adjust datetime64 timestamp, recompute datetimestruct
-        trans, deltas, typ = get_dst_info(tz)
-
-        _pos = trans.searchsorted(stamps, side='right') - 1
-        if _pos.dtype != np.int64:
-            _pos = _pos.astype(np.int64)
-        pos = _pos
-
-        # statictzinfo
-        if typ not in ['pytz', 'dateutil']:
-            for i in range(n):
-                if stamps[i] == NPY_NAT:
-                    continue
-                dt64_to_dtstruct(stamps[i] + deltas[0], &dts)
-                curr_reso = _reso_stamp(&dts)
-                if curr_reso < reso:
-                    reso = curr_reso
-        else:
-            for i in range(n):
-                if stamps[i] == NPY_NAT:
-                    continue
-                dt64_to_dtstruct(stamps[i] + deltas[pos[i]], &dts)
-                curr_reso = _reso_stamp(&dts)
-                if curr_reso < reso:
-                    reso = curr_reso
-
-    return reso
-
-
+# -----------------------------------------------------------------------
 # period helpers
 
+
 cdef ndarray[int64_t] localize_dt64arr_to_period(ndarray[int64_t] stamps,
                                                  int freq, object tz):
     cdef:
@@ -1191,7 +1099,7 @@ class Period(_Period):
 
             if freq is None:
                 try:
-                    freq = frequencies.Resolution.get_freq(reso)
+                    freq = Resolution.get_freq(reso)
                 except KeyError:
                     raise ValueError(
                         "Invalid frequency or could not infer: %s" % reso)
@@ -1236,7 +1144,7 @@ def _quarter_to_myear(year, quarter, freq):
         if quarter <= 0 or quarter > 4:
             raise ValueError('Quarter must be 1 <= q <= 4')
 
-        mnum = tslib._MONTH_NUMBERS[tslib._get_rule_month(freq)] + 1
+        mnum = _MONTH_NUMBERS[_get_rule_month(freq)] + 1
         month = (mnum + (quarter - 1) * 3) % 12 + 1
         if month > mnum:
             year -= 1

pandas/_libs/tslib.pyx

@@ -50,6 +50,7 @@ from datetime cimport (
 # stdlib datetime imports
 from datetime import time as datetime_time
 
+
 from tslibs.np_datetime cimport (check_dts_bounds,
                                  reverse_ops,
                                  cmp_scalar,
@@ -61,12 +62,6 @@ from tslibs.np_datetime cimport (check_dts_bounds,
                                  get_timedelta64_value)
 from tslibs.np_datetime import OutOfBoundsDatetime
 
-from khash cimport (
-    khiter_t,
-    kh_destroy_int64, kh_put_int64,
-    kh_init_int64, kh_int64_t,
-    kh_resize_int64, kh_get_int64)
-
 from .tslibs.parsing import parse_datetime_string
 
 cimport cython
@@ -877,33 +872,6 @@ Timestamp.min = Timestamp(_NS_LOWER_BOUND)
 Timestamp.max = Timestamp(_NS_UPPER_BOUND)
 
 
-# ----------------------------------------------------------------------
-# Frequency inference
-
-def unique_deltas(ndarray[int64_t] arr):
-    cdef:
-        Py_ssize_t i, n = len(arr)
-        int64_t val
-        khiter_t k
-        kh_int64_t *table
-        int ret = 0
-        list uniques = []
-
-    table = kh_init_int64()
-    kh_resize_int64(table, 10)
-    for i in range(n - 1):
-        val = arr[i + 1] - arr[i]
-        k = kh_get_int64(table, val)
-        if k == table.n_buckets:
-            kh_put_int64(table, val, &ret)
-            uniques.append(val)
-    kh_destroy_int64(table)
-
-    result = np.array(uniques, dtype=np.int64)
-    result.sort()
-    return result
-
-
 cdef str _NDIM_STRING = "ndim"
 
 # This is PITA. Because we inherit from datetime, which has very specific
@@ -1388,27 +1356,6 @@ _MONTH_NUMBERS = {k: i for i, k in enumerate(_MONTHS)}
 _MONTH_ALIASES = {(k + 1): v for k, v in enumerate(_MONTHS)}
 
 
-cpdef object _get_rule_month(object source, object default='DEC'):
-    """
-    Return starting month of given freq, default is December.
-
-    Example
-    -------
-    >>> _get_rule_month('D')
-    'DEC'
-
-    >>> _get_rule_month('A-JAN')
-    'JAN'
-    """
-    if hasattr(source, 'freqstr'):
-        source = source.freqstr
-    source = source.upper()
-    if '-' not in source:
-        return default
-    else:
-        return source.split('-')[1]
-
-
 cpdef array_with_unit_to_datetime(ndarray values, unit, errors='coerce'):
     """
     convert the ndarray according to the unit