mirror of
https://github.com/kellyjonbrazil/jc.git
synced 2025-06-17 00:07:37 +02:00
add support for strings with UTC offset suffixes
This commit is contained in:
Kelly Brazil
133
jc/utils.py
133
jc/utils.py
@ -370,6 +370,74 @@ class timestamp:
|
||||
|
||||
If the conversion completely fails, all fields will be None.
|
||||
"""
|
||||
formats = (
|
||||
{'id': 1000, 'format': '%a %b %d %H:%M:%S %Y', 'locale': None}, # manual C locale format conversion: Tue Mar 23 16:12:11 2021 or Tue Mar 23 16:12:11 IST 2021
|
||||
{'id': 1100, 'format': '%a %b %d %H:%M:%S %Y %z', 'locale': None}, # git date output: Thu Mar 5 09:17:40 2020 -0800
|
||||
{'id': 1300, 'format': '%Y-%m-%dT%H:%M:%S.%f%Z', 'locale': None}, # ISO Format with UTC (found in syslog 5424): 2003-10-11T22:14:15.003Z
|
||||
{'id': 1310, 'format': '%Y-%m-%dT%H:%M:%S.%f', 'locale': None}, # ISO Format without TZ (found in syslog 5424): 2003-10-11T22:14:15.003
|
||||
{'id': 1500, 'format': '%Y-%m-%d %H:%M', 'locale': None}, # en_US.UTF-8 local format (found in who cli output): 2021-03-23 00:14
|
||||
{'id': 1600, 'format': '%m/%d/%Y %I:%M %p', 'locale': None}, # Windows english format (found in dir cli output): 12/07/2019 02:09 AM
|
||||
{'id': 1700, 'format': '%m/%d/%Y, %I:%M:%S %p', 'locale': None}, # Windows english format wint non-UTC tz (found in systeminfo cli output): 3/22/2021, 1:15:51 PM (UTC-0600)
|
||||
{'id': 1705, 'format': '%m/%d/%Y, %I:%M:%S %p %Z', 'locale': None}, # Windows english format with UTC tz (found in systeminfo cli output): 3/22/2021, 1:15:51 PM (UTC)
|
||||
{'id': 1710, 'format': '%m/%d/%Y, %I:%M:%S %p UTC%z', 'locale': None}, # Windows english format with UTC tz (found in systeminfo cli output): 3/22/2021, 1:15:51 PM (UTC+0000)
|
||||
{'id': 2000, 'format': '%a %d %b %Y %I:%M:%S %p %Z', 'locale': None}, # en_US.UTF-8 local format (found in upower cli output): Tue 23 Mar 2021 04:12:11 PM UTC
|
||||
{'id': 3000, 'format': '%a %d %b %Y %I:%M:%S %p', 'locale': None}, # en_US.UTF-8 local format with non-UTC tz (found in upower cli output): Tue 23 Mar 2021 04:12:11 PM IST
|
||||
{'id': 4000, 'format': '%A %d %B %Y %I:%M:%S %p %Z', 'locale': None}, # European-style local format (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM UTC
|
||||
{'id': 5000, 'format': '%A %d %B %Y %I:%M:%S %p', 'locale': None}, # European-style local format with non-UTC tz (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM IST
|
||||
{'id': 6000, 'format': '%a %b %d %I:%M:%S %p %Z %Y', 'locale': None}, # en_US.UTF-8 format (found in date cli): Wed Mar 24 06:16:19 PM UTC 2021
|
||||
{'id': 7000, 'format': '%a %b %d %H:%M:%S %Z %Y', 'locale': None}, # C locale format (found in date cli): Wed Mar 24 11:11:30 UTC 2021
|
||||
{'id': 7100, 'format': '%b %d %H:%M:%S %Y', 'locale': None}, # C locale format (found in stat cli output - osx): # Mar 29 11:49:05 2021
|
||||
{'id': 7200, 'format': '%Y-%m-%d %H:%M:%S.%f %z', 'locale': None}, # C locale format (found in stat cli output - linux): 2019-08-13 18:13:43.555604315 -0400
|
||||
{'id': 7250, 'format': '%Y-%m-%d %H:%M:%S', 'locale': None}, # C locale format with non-UTC tz (found in modified vmstat cli output): # 2021-09-16 20:32:28 PDT
|
||||
{'id': 7255, 'format': '%Y-%m-%d %H:%M:%S %Z', 'locale': None}, # C locale format (found in modified vmstat cli output): # 2021-09-16 20:32:28 UTC
|
||||
{'id': 7300, 'format': '%a %Y-%m-%d %H:%M:%S %Z', 'locale': None}, # C locale format (found in timedatectl cli output): # Wed 2020-03-11 00:53:21 UTC
|
||||
# attempt locale changes last
|
||||
{'id': 8000, 'format': '%a %d %b %Y %H:%M:%S %Z', 'locale': ''}, # current locale format (found in upower cli output): # mar. 23 mars 2021 23:12:11 UTC
|
||||
{'id': 8100, 'format': '%a %d %b %Y %H:%M:%S', 'locale': ''}, # current locale format with non-UTC tz (found in upower cli output): # mar. 23 mars 2021 19:12:11 EDT
|
||||
{'id': 8200, 'format': '%A %d %B %Y, %H:%M:%S UTC%z', 'locale': ''}, # fr_FR.utf8 locale format (found in date cli output): vendredi 26 mars 2021, 13:26:46 (UTC+0000)
|
||||
{'id': 8300, 'format': '%A %d %B %Y, %H:%M:%S', 'locale': ''}, # fr_FR.utf8 locale format with non-UTC tz (found in date cli output): vendredi 26 mars 2021, 13:26:46 (UTC-0400)
|
||||
{'id': 9000, 'format': '%c', 'locale': ''} # locally configured locale format conversion: Could be anything :) this is a last-gasp attempt
|
||||
)
|
||||
|
||||
# from https://www.timeanddate.com/time/zones/
|
||||
# only removed UTC timezone and added known non-UTC offsets
|
||||
tz_abbr = {
|
||||
'A', 'ACDT', 'ACST', 'ACT', 'ACWST', 'ADT', 'AEDT', 'AEST', 'AET', 'AFT', 'AKDT',
|
||||
'AKST', 'ALMT', 'AMST', 'AMT', 'ANAST', 'ANAT', 'AQTT', 'ART', 'AST', 'AT', 'AWDT',
|
||||
'AWST', 'AZOST', 'AZOT', 'AZST', 'AZT', 'AoE', 'B', 'BNT', 'BOT', 'BRST', 'BRT', 'BST',
|
||||
'BTT', 'C', 'CAST', 'CAT', 'CCT', 'CDT', 'CEST', 'CET', 'CHADT', 'CHAST', 'CHOST',
|
||||
'CHOT', 'CHUT', 'CIDST', 'CIST', 'CKT', 'CLST', 'CLT', 'COT', 'CST', 'CT', 'CVT', 'CXT',
|
||||
'ChST', 'D', 'DAVT', 'DDUT', 'E', 'EASST', 'EAST', 'EAT', 'ECT', 'EDT', 'EEST', 'EET',
|
||||
'EGST', 'EGT', 'EST', 'ET', 'F', 'FET', 'FJST', 'FJT', 'FKST', 'FKT', 'FNT', 'G',
|
||||
'GALT', 'GAMT', 'GET', 'GFT', 'GILT', 'GMT', 'GST', 'GYT', 'H', 'HDT', 'HKT', 'HOVST',
|
||||
'HOVT', 'HST', 'I', 'ICT', 'IDT', 'IOT', 'IRDT', 'IRKST', 'IRKT', 'IRST', 'IST', 'JST',
|
||||
'K', 'KGT', 'KOST', 'KRAST', 'KRAT', 'KST', 'KUYT', 'L', 'LHDT', 'LHST', 'LINT', 'M',
|
||||
'MAGST', 'MAGT', 'MART', 'MAWT', 'MDT', 'MHT', 'MMT', 'MSD', 'MSK', 'MST', 'MT', 'MUT',
|
||||
'MVT', 'MYT', 'N', 'NCT', 'NDT', 'NFDT', 'NFT', 'NOVST', 'NOVT', 'NPT', 'NRT', 'NST',
|
||||
'NUT', 'NZDT', 'NZST', 'O', 'OMSST', 'OMST', 'ORAT', 'P', 'PDT', 'PET', 'PETST', 'PETT',
|
||||
'PGT', 'PHOT', 'PHT', 'PKT', 'PMDT', 'PMST', 'PONT', 'PST', 'PT', 'PWT', 'PYST', 'PYT',
|
||||
'Q', 'QYZT', 'R', 'RET', 'ROTT', 'S', 'SAKT', 'SAMT', 'SAST', 'SBT', 'SCT', 'SGT',
|
||||
'SRET', 'SRT', 'SST', 'SYOT', 'T', 'TAHT', 'TFT', 'TJT', 'TKT', 'TLT', 'TMT', 'TOST',
|
||||
'TOT', 'TRT', 'TVT', 'U', 'ULAST', 'ULAT', 'UYST', 'UYT', 'UZT', 'V', 'VET', 'VLAST',
|
||||
'VLAT', 'VOST', 'VUT', 'W', 'WAKT', 'WARST', 'WAST', 'WAT', 'WEST', 'WET', 'WFT',
|
||||
'WGST', 'WGT', 'WIB', 'WIT', 'WITA', 'WST', 'WT', 'X', 'Y', 'YAKST', 'YAKT', 'YAPT',
|
||||
'YEKST', 'YEKT', 'UTC-1200', 'UTC-1100', 'UTC-1000', 'UTC-0930', 'UTC-0900',
|
||||
'UTC-0800', 'UTC-0700', 'UTC-0600', 'UTC-0500', 'UTC-0400', 'UTC-0300', 'UTC-0230',
|
||||
'UTC-0200', 'UTC-0100', 'UTC+0100', 'UTC+0200', 'UTC+0300', 'UTC+0400', 'UTC+0430',
|
||||
'UTC+0500', 'UTC+0530', 'UTC+0545', 'UTC+0600', 'UTC+0630', 'UTC+0700', 'UTC+0800',
|
||||
'UTC+0845', 'UTC+0900', 'UTC+1000', 'UTC+1030', 'UTC+1100', 'UTC+1200', 'UTC+1300',
|
||||
'UTC+1345', 'UTC+1400'
|
||||
}
|
||||
|
||||
offset_suffixes = (
|
||||
'-12:00', '-11:00', '-10:00', '-09:30', '-09:00',
|
||||
'-08:00', '-07:00', '-06:00', '-05:00', '-04:00', '-03:00', '-02:30',
|
||||
'-02:00', '-01:00', '+01:00', '+02:00', '+03:00', '+04:00', '+04:30',
|
||||
'+05:00', '+05:30', '+05:45', '+06:00', '+06:30', '+07:00', '+08:00',
|
||||
'+08:45', '+09:00', '+10:00', '+10:30', '+11:00', '+12:00', '+13:00',
|
||||
'+13:45', '+14:00'
|
||||
)
|
||||
|
||||
data = dt_string or ''
|
||||
normalized_datetime = ''
|
||||
utc_tz = False
|
||||
@ -404,65 +472,6 @@ class timestamp:
|
||||
elif '+0000' in data or '-0000' in data:
|
||||
utc_tz = True
|
||||
|
||||
formats = [
|
||||
{'id': 1000, 'format': '%a %b %d %H:%M:%S %Y', 'locale': None}, # manual C locale format conversion: Tue Mar 23 16:12:11 2021 or Tue Mar 23 16:12:11 IST 2021
|
||||
{'id': 1100, 'format': '%a %b %d %H:%M:%S %Y %z', 'locale': None}, # git date output: Thu Mar 5 09:17:40 2020 -0800
|
||||
{'id': 1300, 'format': '%Y-%m-%dT%H:%M:%S.%f%Z', 'locale': None}, # ISO Format with UTC (found in syslog 5424): 2003-10-11T22:14:15.003Z
|
||||
{'id': 1310, 'format': '%Y-%m-%dT%H:%M:%S.%f', 'locale': None}, # ISO Format without UTC (found in syslog 5424): 2003-10-11T22:14:15.003
|
||||
{'id': 1500, 'format': '%Y-%m-%d %H:%M', 'locale': None}, # en_US.UTF-8 local format (found in who cli output): 2021-03-23 00:14
|
||||
{'id': 1600, 'format': '%m/%d/%Y %I:%M %p', 'locale': None}, # Windows english format (found in dir cli output): 12/07/2019 02:09 AM
|
||||
{'id': 1700, 'format': '%m/%d/%Y, %I:%M:%S %p', 'locale': None}, # Windows english format wint non-UTC tz (found in systeminfo cli output): 3/22/2021, 1:15:51 PM (UTC-0600)
|
||||
{'id': 1705, 'format': '%m/%d/%Y, %I:%M:%S %p %Z', 'locale': None}, # Windows english format with UTC tz (found in systeminfo cli output): 3/22/2021, 1:15:51 PM (UTC)
|
||||
{'id': 1710, 'format': '%m/%d/%Y, %I:%M:%S %p UTC%z', 'locale': None}, # Windows english format with UTC tz (found in systeminfo cli output): 3/22/2021, 1:15:51 PM (UTC+0000)
|
||||
{'id': 2000, 'format': '%a %d %b %Y %I:%M:%S %p %Z', 'locale': None}, # en_US.UTF-8 local format (found in upower cli output): Tue 23 Mar 2021 04:12:11 PM UTC
|
||||
{'id': 3000, 'format': '%a %d %b %Y %I:%M:%S %p', 'locale': None}, # en_US.UTF-8 local format with non-UTC tz (found in upower cli output): Tue 23 Mar 2021 04:12:11 PM IST
|
||||
{'id': 4000, 'format': '%A %d %B %Y %I:%M:%S %p %Z', 'locale': None}, # European-style local format (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM UTC
|
||||
{'id': 5000, 'format': '%A %d %B %Y %I:%M:%S %p', 'locale': None}, # European-style local format with non-UTC tz (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM IST
|
||||
{'id': 6000, 'format': '%a %b %d %I:%M:%S %p %Z %Y', 'locale': None}, # en_US.UTF-8 format (found in date cli): Wed Mar 24 06:16:19 PM UTC 2021
|
||||
{'id': 7000, 'format': '%a %b %d %H:%M:%S %Z %Y', 'locale': None}, # C locale format (found in date cli): Wed Mar 24 11:11:30 UTC 2021
|
||||
{'id': 7100, 'format': '%b %d %H:%M:%S %Y', 'locale': None}, # C locale format (found in stat cli output - osx): # Mar 29 11:49:05 2021
|
||||
{'id': 7200, 'format': '%Y-%m-%d %H:%M:%S.%f %z', 'locale': None}, # C locale format (found in stat cli output - linux): 2019-08-13 18:13:43.555604315 -0400
|
||||
{'id': 7250, 'format': '%Y-%m-%d %H:%M:%S', 'locale': None}, # C locale format with non-UTC tz (found in modified vmstat cli output): # 2021-09-16 20:32:28 PDT
|
||||
{'id': 7255, 'format': '%Y-%m-%d %H:%M:%S %Z', 'locale': None}, # C locale format (found in modified vmstat cli output): # 2021-09-16 20:32:28 UTC
|
||||
{'id': 7300, 'format': '%a %Y-%m-%d %H:%M:%S %Z', 'locale': None}, # C locale format (found in timedatectl cli output): # Wed 2020-03-11 00:53:21 UTC
|
||||
# attempt locale changes last
|
||||
{'id': 8000, 'format': '%a %d %b %Y %H:%M:%S %Z', 'locale': ''}, # current locale format (found in upower cli output): # mar. 23 mars 2021 23:12:11 UTC
|
||||
{'id': 8100, 'format': '%a %d %b %Y %H:%M:%S', 'locale': ''}, # current locale format with non-UTC tz (found in upower cli output): # mar. 23 mars 2021 19:12:11 EDT
|
||||
{'id': 8200, 'format': '%A %d %B %Y, %H:%M:%S UTC%z', 'locale': ''}, # fr_FR.utf8 locale format (found in date cli output): vendredi 26 mars 2021, 13:26:46 (UTC+0000)
|
||||
{'id': 8300, 'format': '%A %d %B %Y, %H:%M:%S', 'locale': ''}, # fr_FR.utf8 locale format with non-UTC tz (found in date cli output): vendredi 26 mars 2021, 13:26:46 (UTC-0400)
|
||||
{'id': 9000, 'format': '%c', 'locale': ''} # locally configured locale format conversion: Could be anything :) this is a last-gasp attempt
|
||||
]
|
||||
|
||||
# from https://www.timeanddate.com/time/zones/
|
||||
# only removed UTC timezone and added known non-UTC offsets
|
||||
tz_abbr = [
|
||||
'A', 'ACDT', 'ACST', 'ACT', 'ACWST', 'ADT', 'AEDT', 'AEST', 'AET', 'AFT', 'AKDT',
|
||||
'AKST', 'ALMT', 'AMST', 'AMT', 'ANAST', 'ANAT', 'AQTT', 'ART', 'AST', 'AT', 'AWDT',
|
||||
'AWST', 'AZOST', 'AZOT', 'AZST', 'AZT', 'AoE', 'B', 'BNT', 'BOT', 'BRST', 'BRT', 'BST',
|
||||
'BTT', 'C', 'CAST', 'CAT', 'CCT', 'CDT', 'CEST', 'CET', 'CHADT', 'CHAST', 'CHOST',
|
||||
'CHOT', 'CHUT', 'CIDST', 'CIST', 'CKT', 'CLST', 'CLT', 'COT', 'CST', 'CT', 'CVT', 'CXT',
|
||||
'ChST', 'D', 'DAVT', 'DDUT', 'E', 'EASST', 'EAST', 'EAT', 'ECT', 'EDT', 'EEST', 'EET',
|
||||
'EGST', 'EGT', 'EST', 'ET', 'F', 'FET', 'FJST', 'FJT', 'FKST', 'FKT', 'FNT', 'G',
|
||||
'GALT', 'GAMT', 'GET', 'GFT', 'GILT', 'GMT', 'GST', 'GYT', 'H', 'HDT', 'HKT', 'HOVST',
|
||||
'HOVT', 'HST', 'I', 'ICT', 'IDT', 'IOT', 'IRDT', 'IRKST', 'IRKT', 'IRST', 'IST', 'JST',
|
||||
'K', 'KGT', 'KOST', 'KRAST', 'KRAT', 'KST', 'KUYT', 'L', 'LHDT', 'LHST', 'LINT', 'M',
|
||||
'MAGST', 'MAGT', 'MART', 'MAWT', 'MDT', 'MHT', 'MMT', 'MSD', 'MSK', 'MST', 'MT', 'MUT',
|
||||
'MVT', 'MYT', 'N', 'NCT', 'NDT', 'NFDT', 'NFT', 'NOVST', 'NOVT', 'NPT', 'NRT', 'NST',
|
||||
'NUT', 'NZDT', 'NZST', 'O', 'OMSST', 'OMST', 'ORAT', 'P', 'PDT', 'PET', 'PETST', 'PETT',
|
||||
'PGT', 'PHOT', 'PHT', 'PKT', 'PMDT', 'PMST', 'PONT', 'PST', 'PT', 'PWT', 'PYST', 'PYT',
|
||||
'Q', 'QYZT', 'R', 'RET', 'ROTT', 'S', 'SAKT', 'SAMT', 'SAST', 'SBT', 'SCT', 'SGT',
|
||||
'SRET', 'SRT', 'SST', 'SYOT', 'T', 'TAHT', 'TFT', 'TJT', 'TKT', 'TLT', 'TMT', 'TOST',
|
||||
'TOT', 'TRT', 'TVT', 'U', 'ULAST', 'ULAT', 'UYST', 'UYT', 'UZT', 'V', 'VET', 'VLAST',
|
||||
'VLAT', 'VOST', 'VUT', 'W', 'WAKT', 'WARST', 'WAST', 'WAT', 'WEST', 'WET', 'WFT',
|
||||
'WGST', 'WGT', 'WIB', 'WIT', 'WITA', 'WST', 'WT', 'X', 'Y', 'YAKST', 'YAKT', 'YAPT',
|
||||
'YEKST', 'YEKT', 'UTC-1200', 'UTC-1100', 'UTC-1000', 'UTC-0930', 'UTC-0900',
|
||||
'UTC-0800', 'UTC-0700', 'UTC-0600', 'UTC-0500', 'UTC-0400', 'UTC-0300', 'UTC-0230',
|
||||
'UTC-0200', 'UTC-0100', 'UTC+0100', 'UTC+0200', 'UTC+0300', 'UTC+0400', 'UTC+0430',
|
||||
'UTC+0500', 'UTC+0530', 'UTC+0545', 'UTC+0600', 'UTC+0630', 'UTC+0700', 'UTC+0800',
|
||||
'UTC+0845', 'UTC+0900', 'UTC+1000', 'UTC+1030', 'UTC+1100', 'UTC+1200', 'UTC+1300',
|
||||
'UTC+1345', 'UTC+1400'
|
||||
]
|
||||
|
||||
# normalize the timezone by taking out any timezone reference, except UTC
|
||||
cleandata = data.replace('(', '').replace(')', '')
|
||||
normalized_datetime_list = []
|
||||
@ -472,6 +481,12 @@ class timestamp:
|
||||
|
||||
normalized_datetime = ' '.join(normalized_datetime_list)
|
||||
|
||||
# remove non UTC offset suffixes at the end of the string
|
||||
for suffix in offset_suffixes:
|
||||
if normalized_datetime.endswith(suffix):
|
||||
normalized_datetime = normalized_datetime[0:-len(suffix)]
|
||||
break
|
||||
|
||||
# normalize further by converting any greater-than 6-digit subsecond to 6-digits
|
||||
p = re.compile(r'(\W\d\d:\d\d:\d\d\.\d{6})\d+\W')
|
||||
normalized_datetime = p.sub(r'\g<1> ', normalized_datetime)
|
||||
|
||||
Reference in New Issue
Block a user