I am using the regex package to work with approximate regular expressions in Python. I do not understand the behavior I see when errors are counted in matches.
If I use the following regular expression with errors allowed:
>>> rex = '(HOUSE OF REPRESENTATIVES){e}'
I get exactly the behavior I expect, where a match is found with 20 errors. The 20 errors are due to the 20 lowercase letters:
>>> regex.match(rex, 'House of Representatives')
<regex.Match object; span=(0, 24), match='House of Representatives', fuzzy_counts=(20, 0, 0)>
When I attempt to something similar with added complexity, I get a result that makes no sense to me.
I am attempting to match strings that look like: "HOUSE OF REPRESENTATIVES\nWEDNESDAY, JANUARY 23, 1939\n".
>>> rex = '(HOUSE OF REPRESENTATIVES\n[A-Z]{3,6}DAY, [A-Z]{3,11} [0-9]{1,2}, [0-9]{4}\n){e}'
When I attempt to use this, I get an error number that I cannot understand.
>>> h1 = 'HOUSE OF REPRESENTATIVES\nWEDNESDAY, JANUARY 4, 1939\n'
>>> regex.match(rex, h1)
<regex.Match object; span=(0, 52), match='HOUSE OF REPRESENTATIVES\nWEDNESDAY, JANUARY 4, 1939\n', fuzzy_counts=(7, 0, 2)>
I get a match 9 fuzzy errors when I expect 0. What gives?
Note: When I use the BESTMATCH flag, I get 8 errors.