2229536127

Build Type

github

Committed by GitHub

Commit Message

Merge c519fcbcb into d076d8475

Pull Request Pull Request #33: Support CORDEX and CMIP5/6

Run Details

32 of 170 new or added lines in 15 files covered. (18.82%)

6 existing lines in 5 files now uncovered.

659 of 3229 relevant lines covered (20.41%)

0.61 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

24.56

/miranda/decode/_time.py

import logging
from logging import config
from typing import Union

import cftime
import pandas as pd
from pandas._libs.tslibs import NaTType  # noqa

from miranda.scripting import LOGGING_CONFIG

logging.config.dictConfig(LOGGING_CONFIG)

__all__ = [
    "date_parser",
    "DecoderError",
    "TIME_UNITS_TO_FREQUENCY",
    "TIME_UNITS_TO_TIMEDELTA",
]

TIME_UNITS_TO_FREQUENCY = {
    "subhrPt": "sub-hr",
    "hourly": "1hr",
    "hours": "1hr",
    "1hr": "1hr",
    "3-hourly": "3hr",
    "3hr": "3hr",
    "6-hourly": "6hr",
    "6hr": "6hr",
    "daily": "day",
    "days": "day",
    "day": "day",
    "weekly": "sem",
    "weeks": "sem",
    "sem": "sem",
    "monthly": "mon",
    "months": "mon",
    "mon": "mon",
    "monC": "monC",
    "Amon": "mon",
    "Omon": "mon",
    "yearly": "yr",
    "years": "yr",
    "annual": "yr",
    "yr": "yr",
    "yrPt": "yrPt",
    "decadal": "dec",
    "decades": "dec",
    "dec": "dec",
    "fixed": "fx",
    "fx": "fx",
}
TIME_UNITS_TO_TIMEDELTA = {
    "hourly": "1h",
    "hours": "1h",
    "1hr": "1h",
    "1hrCM": "1h",
    "1hrPt": "1h",
    "3hr": "3h",
    "3hrPt": "3h",
    "6-hourly": "6h",
    "6hr": "6h",
    "6hrPt": "6h",
    "daily": "1d",
    "day": "1d",
    "days": "1d",
    "weekly": "7d",
    "weeks": "7d",
    "sem": "7d",
    "mon": "30d",
    "monC": "30d",
    "monPt": "30d",
    "Amon": "30d",
    "QS": "90d",
    "qtr": "90d",
    "yearly": "365d",
    "years": "365d",
    "year": "365d",
    "yr": "365d",
    "yrPt": "365d",
}


class DecoderError(Exception):
    pass


def date_parser(
    date: str,
    *,
    end_of_period: bool = False,
    output_type: str = "str",
    strftime_format: str = "%Y-%m-%d",
) -> Union[str, pd.Timestamp, NaTType]:
    """Returns a datetime from a string.

    Parameters
    ----------
    date : str
      Date to be converted.
    end_of_period : bool
      If True, the date will be the end of month or year depending on what's most appropriate.
    output_type: {"datetime", "str"}
      Returned object type.
    strftime_format: str
      If output_type=='str', this sets the strftime format.

    Returns
    -------
    pd.Timestamp or str or pd.NaT
      Parsed date.

    Notes
    -----
    Adapted from code written by Gabriel Rondeau-Genesse (@RondeauG)
    """

    # Formats, ordered depending on string length
    formats = {
        4: ["%Y"],
        6: ["%Y%m"],
        7: ["%Y-%m"],
        8: ["%Y%m%d"],
        10: ["%Y%m%d%H", "%Y-%m-%d"],
        12: ["%Y%m%d%H%M"],
        13: ["%Y%m-%Y%m"],
        17: ["%Y%m%d-%Y%m%d"],
        19: ["%Y-%m-%dT%H:%M:%S"],
        21: ["%Y%m%d%H-%Y%m%d%H"],
        25: ["%Y%m%d%H%M-%Y%m%d%H%M"],
    }
    end_date_found = False

    def _parse_date(d, fmts):
        for fmt in fmts:
            try:
                s = pd.to_datetime(d, format=fmt)
                match = fmt
                break
            except ValueError:
                pass
        else:
            raise DecoderError(f"Can't parse date {d} with supported formats {fmts}.")
        return s, match

    date_format = None
    if isinstance(date, str):
        if len(date) in [13, 17, 21, 25]:
            dates = date.split("-")
            if not end_of_period:
                date = dates[0]
            else:
                date = dates[1]
                end_date_found = True

        try:
            possible_formats = formats[len(date)]
            date, date_format = _parse_date(date, possible_formats)
        except KeyError:
            # Return NaT for fixed/missing/ill-formatted date strings
            return pd.NaT

    elif isinstance(date, cftime.datetime):
        for n in range(3):
            try:
                date = pd.Timestamp.fromisoformat((date - pd.Timedelta(n)).isoformat())
            except ValueError:  # We are NOT catching OutOfBoundsDatetime.
                pass
            else:
                break
        else:
            raise DecoderError(
                f"Unable to parse cftime date {date}, even when moving back 2 days."
            )
    elif not isinstance(date, pd.Timestamp):
        date = pd.Timestamp(date)  # noqa

    if end_of_period and date_format and not end_date_found:
        if "m" not in date_format:
            date = date + pd.tseries.offsets.YearEnd(1)  # noqa
        elif "d" not in date_format:
            date = date + pd.tseries.offsets.MonthEnd(1)  # noqa
        # TODO: Implement sub-daily?

    if output_type == "str":
        return date.strftime(strftime_format)

    return date

1	import logging	3✔
2	from logging import config	3✔
3	from typing import Union	3✔
4
5	import cftime	3✔
6	import pandas as pd	3✔
7	from pandas._libs.tslibs import NaTType # noqa	3✔
8
9	from miranda.scripting import LOGGING_CONFIG	3✔
10
11	logging.config.dictConfig(LOGGING_CONFIG)	3✔
12
13	__all__ = [	3✔
14	"date_parser",
15	"DecoderError",
16	"TIME_UNITS_TO_FREQUENCY",
17	"TIME_UNITS_TO_TIMEDELTA",
18	]
19
20	TIME_UNITS_TO_FREQUENCY = {	3✔
21	"subhrPt": "sub-hr",
22	"hourly": "1hr",
23	"hours": "1hr",
24	"1hr": "1hr",
25	"3-hourly": "3hr",
26	"3hr": "3hr",
27	"6-hourly": "6hr",
28	"6hr": "6hr",
29	"daily": "day",
30	"days": "day",
31	"day": "day",
32	"weekly": "sem",
33	"weeks": "sem",
34	"sem": "sem",
35	"monthly": "mon",
36	"months": "mon",
37	"mon": "mon",
38	"monC": "monC",
39	"Amon": "mon",
40	"Omon": "mon",
41	"yearly": "yr",
42	"years": "yr",
43	"annual": "yr",
44	"yr": "yr",
45	"yrPt": "yrPt",
46	"decadal": "dec",
47	"decades": "dec",
48	"dec": "dec",
49	"fixed": "fx",
50	"fx": "fx",
51	}
52	TIME_UNITS_TO_TIMEDELTA = {	3✔
53	"hourly": "1h",
54	"hours": "1h",
55	"1hr": "1h",
56	"1hrCM": "1h",
57	"1hrPt": "1h",
58	"3hr": "3h",
59	"3hrPt": "3h",
60	"6-hourly": "6h",
61	"6hr": "6h",
62	"6hrPt": "6h",
63	"daily": "1d",
64	"day": "1d",
65	"days": "1d",
66	"weekly": "7d",
67	"weeks": "7d",
68	"sem": "7d",
69	"mon": "30d",
70	"monC": "30d",
71	"monPt": "30d",
72	"Amon": "30d",
73	"QS": "90d",
74	"qtr": "90d",
75	"yearly": "365d",
76	"years": "365d",
77	"year": "365d",
78	"yr": "365d",
79	"yrPt": "365d",
80	}
81
82
83	class DecoderError(Exception):	3✔
84	pass	3✔
85
86
87	def date_parser(	3✔
88	date: str,
89	*,
90	end_of_period: bool = False,
91	output_type: str = "str",
92	strftime_format: str = "%Y-%m-%d",
93	) -> Union[str, pd.Timestamp, NaTType]:
94	"""Returns a datetime from a string.
95
96	Parameters
97	----------
98	date : str
99	Date to be converted.
100	end_of_period : bool
101	If True, the date will be the end of month or year depending on what's most appropriate.
102	output_type: {"datetime", "str"}
103	Returned object type.
104	strftime_format: str
105	If output_type=='str', this sets the strftime format.
106
107	Returns
108	-------
109	pd.Timestamp or str or pd.NaT
110	Parsed date.
111
112	Notes
113	-----
114	Adapted from code written by Gabriel Rondeau-Genesse (@RondeauG)
115	"""
116
117	# Formats, ordered depending on string length
118	formats = {	×
119	4: ["%Y"],
120	6: ["%Y%m"],
121	7: ["%Y-%m"],
122	8: ["%Y%m%d"],
123	10: ["%Y%m%d%H", "%Y-%m-%d"],
124	12: ["%Y%m%d%H%M"],
125	13: ["%Y%m-%Y%m"],
126	17: ["%Y%m%d-%Y%m%d"],
127	19: ["%Y-%m-%dT%H:%M:%S"],
128	21: ["%Y%m%d%H-%Y%m%d%H"],
129	25: ["%Y%m%d%H%M-%Y%m%d%H%M"],
130	}
131	end_date_found = False	×
132
133	def _parse_date(d, fmts):	×
134	for fmt in fmts:	×
135	try:	×
136	s = pd.to_datetime(d, format=fmt)	×
137	match = fmt	×
138	break	×
139	except ValueError:	×
140	pass	×
141	else:
NEW 142	raise DecoderError(f"Can't parse date {d} with supported formats {fmts}.")	×
143	return s, match	×
144
145	date_format = None	×
146	if isinstance(date, str):	×
NEW 147	if len(date) in [13, 17, 21, 25]:	×
148	dates = date.split("-")	×
149	if not end_of_period:	×
150	date = dates[0]	×
151	else:
152	date = dates[1]	×
153	end_date_found = True	×
154
155	try:	×
156	possible_formats = formats[len(date)]	×
157	date, date_format = _parse_date(date, possible_formats)	×
158	except KeyError:	×
159	# Return NaT for fixed/missing/ill-formatted date strings
160	return pd.NaT	×
161
162	elif isinstance(date, cftime.datetime):	×
163	for n in range(3):	×
164	try:	×
165	date = pd.Timestamp.fromisoformat((date - pd.Timedelta(n)).isoformat())	×
166	except ValueError: # We are NOT catching OutOfBoundsDatetime.	×
167	pass	×
168	else:
169	break	×
170	else:
171	raise DecoderError(	×
172	f"Unable to parse cftime date {date}, even when moving back 2 days."
173	)
174	elif not isinstance(date, pd.Timestamp):	×
175	date = pd.Timestamp(date) # noqa	×
176
177	if end_of_period and date_format and not end_date_found:	×
178	if "m" not in date_format:	×
NEW 179	date = date + pd.tseries.offsets.YearEnd(1) # noqa	×
180	elif "d" not in date_format:	×
NEW 181	date = date + pd.tseries.offsets.MonthEnd(1) # noqa	×
182	# TODO: Implement sub-daily?
183
184	if output_type == "str":	×
NEW 185	return date.strftime(strftime_format)	×
186
187	return date	×

Ouranosinc / miranda / 2229536127

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous