How to extract same parts of file name with different patterns? - python-3.x

Question

I have the following file name formats:
2020-01-05-ABC1111_001.jpg
2020_02_06_B444444_MN_004.jpg
2020_03_20_KUKU44223222-STAFF_005.jpg
2020-04-03-LULU4444211-MN_018.jpg
Most (99%) of the files are of the following format:
2020_04_03_LULU4444211_018.jpg
And I just use rsplit("_", 2) and get what I need.
Where the first part is date, second is and ID, MN or STAFF, page number.
How to build a good regex or split function to somehow split it to date, id, and page?
From all the above examples I would like to get:
{
"2020-01-05-ABC1111_001.jpg": {"date": 2020-01-05, "id": ABC1111, "page_num": 1},
"2020_02_06_B444444_MN_004.jpg": {"date": 2020_02_06, "id": B444444, "page_num": 4},
"2020_03_20_KUKU44223222-STAFF_005.jpg": {"date": 2020_03_20, "id": KUKU44223222, "page_num": 5},
"2020-04-03-LULU4444211-MN_018.jpg": {"date": 2020-04-03, "id": LULU4444211, "page_num": 18}
}
I am have tried rsplit, I know there is an annotation option + Spacy NER model but maybe there is another way to do it more simply?

Answer 1

You might use code like
import re
strings = ['2020-01-05-ABC1111_001.jpg','2020_02_06_B444444_MN_004.jpg','2020_03_20_KUKU44223222-STAFF_005.jpg','2020-04-03-LULU4444211-MN_018.jpg']
rx = re.compile(r'(?P<date>\d{4}[-_]\d{2}[-_]\d{2})[-_](?P<id>[^_-]+)(?:[_-](?:MN|STAFF))?[_-](?P<page_num>\d+)')
d = {}
for s in strings:
m = rx.search(s)
if m:
d[s] = m.groupdict()
print(d)
See the Python demo, yielding
{'2020-01-05-ABC1111_001.jpg': {'date': '2020-01-05', 'id': 'ABC1111', 'page_num': '001'}, '2020_02_06_B444444_MN_004.jpg': {'date': '2020_02_06', 'id': 'B444444', 'page_num': '004'}, '2020_03_20_KUKU44223222-STAFF_005.jpg': {'date': '2020_03_20', 'id': 'KUKU44223222', 'page_num': '005'}, '2020-04-03-LULU4444211-MN_018.jpg': {'date': '2020-04-03', 'id': 'LULU4444211', 'page_num': '018'}}
Note the regex used contains named capturing groups so that you could get access to .groupdict() after a match is found, it looks like
(?P<date>\d{4}[-_]\d{2}[-_]\d{2})[-_](?P<id>[^_-]+)(?:[_-](?:MN|STAFF))?[_-](?P<page_num>\d+)
See the regex demo.
Regex details
(?P<date>\d{4}[-_]\d{2}[-_]\d{2}) - Group "date": 4 digits, _ or -, 2 digits, _ or - and then again 2 digits
[-_] - a hyphen or underscore
(?P<id>[^_-]+) - Group "id": 1 or more chars other than - and _
(?:[_-](?:MN|STAFF))? - an optional non-capturing group matching - or _ and then MN or STAFF
[_-] - a - or _
(?P<page_num>\d+) - Group "page_num": 1 or more digits.

Answer 2

Regexp:
(\d{4}[-_]\d{2}[-_]\d{2})[-_](.*)[-_](\d+)\.[a-zA-Z]+
It contains three regexp groups:
date
id
page number
Explanation:
(\d{4}[-_]\d{2}[-_]\d{2}) # date (yyyy-mm-dd or yyyy_mm_dd) - group 1
[-_] # separator (dash or underscore)
(.+) # id (any character) - group 2
[-_] # separator
(\d+) # page number - group 3
\.[a-zA-Z]+ # file extension
Demo: https://regex101.com/r/IPF7QE/1
You can read groups in Python this way:
if match := re.search(regexp, text_line, re.IGNORECASE):
date = match.group(1)
id = match.group(2)
page_number = match.group(3)