python/jc
1
0
Fork 0
mirror of https://github.com/kellyjonbrazil/jc.git synced 2025-06-17 00:07:37 +02:00
Code Issues Releases Activity

normalize datetime string to remove all timezones except UTC

Browse Source
This commit is contained in:
Kelly Brazil
2021-03-24 17:01:56 -07:00
parent a5c444587b
commit 1d41c46cc7
1 changed files with 34 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
jc/utils.py
View File

@ -124,16 +124,45 @@ def parse_datetime_to_timestamp(data):
formats = [
{'id': 1, 'format': '%c', 'locale': None}, # C locale format conversion: Tue Mar 23 16:12:11 2021
{'id': 2, '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': 3, 'format': '%A %d %B %Y %I:%M:%S %p %Z', 'locale': None}, # European local format (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM UTC
{'id': 4, 'format': '%a %b %d %I:%M:%S %p %Z %Y', 'locale': None}, # date cli command in en_US.UTF-8 format: Wed Mar 24 06:16:19 PM UTC 2021
{'id': 5, 'format': '%a %b %d %H:%M:%S %Z %Y', 'locale': None}, # date cli command in C locale format: Wed Mar 24 11:11:30 PDT 2021
{'id': 6, 'format': '%c', 'locale': ''} # locally configured locale format conversion: Could be anything :) this is a last-gasp attempt
{'id': 3, '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': 4, 'format': '%A %d %B %Y %I:%M:%S %p %Z', 'locale': None}, # European local format (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM UTC
{'id': 5, 'format': '%A %d %B %Y %I:%M:%S %p', 'locale': None}, # European local format with non-UTC tz (found in upower cli output): Tuesday 01 October 2019 12:50:41 PM IST
{'id': 6, 'format': '%a %b %d %I:%M:%S %p %Z %Y', 'locale': None}, # date cli command in en_US.UTF-8 format: Wed Mar 24 06:16:19 PM UTC 2021
{'id': 7, 'format': '%a %b %d %H:%M:%S %Z %Y', 'locale': None}, # date cli command in C locale format: Wed Mar 24 11:11:30 PDT 2021
{'id': 8, 'format': '%c', 'locale': ''} # locally configured locale format conversion: Could be anything :) this is a last-gasp attempt
]
# from https://www.timeanddate.com/time/zones/
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', 'Z']
# normalize the timezone by taking out any timezone reference, except UTC
normalized_datetime_list = []
for term in data.split():
if term not in tz_abbr:
normalized_datetime_list.append(term)
normalized_datetime = ' '.join(normalized_datetime_list)
for fmt in formats:
try:
locale.setlocale(locale.LC_TIME, fmt['locale'])
dt = datetime.strptime(data, fmt['format'])
dt = datetime.strptime(normalized_datetime, fmt['format'])
timestamp_naive = int(dt.replace(tzinfo=None).timestamp())
timestamp_obj['format'] = fmt['id']
locale.setlocale(locale.LC_TIME, None)