需要Python Regex多个模式匹配和反模式匹配

Question

I have a python string, that has the following pattern of data.

(a) be given by XXX for its election per scenario 3.3(a). (b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially

I might have the above string of pattern which needs to be split by (a) through (z) but it should not split if (a) through (z) occurs in the middle of the statement and particularly the scenario XX(a) through XX(z) should not be selected.

I need it split as (2 lines)

1. (a) be given by XXX for its election per scenario 3.3(a).

2. (b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially

I am trying to pattern-match using python re

import re patterns=["[^0-9] (a) ","[^0-9] (b) ","[^0-9] (c) ","[^0-9] (d) "]

textData="(a) be given by XXX for its election per scenario 3.3(a). (b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially" regexPattern = '|'.join(map(re.escape, patterns)) splitList=re.split(regexPattern,textData) print(splitList)

This is the output Iam getting from executing

['(a) be given by XXX for its election per scenario 3.3(a). (b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially']

The space before and after the '.' varies in previous section and a new section lets say (b) after previous section (a) begins in a new line.

Answer 1

Although your requirements are a bit fuzzy, a reasonable shot given your particular input string seems to be to split on any space that is preceded by a literal . and followed by the literal (letter) pattern.

import re

s = "(a) be given by XXX for its election per scenario 3.3(a). (b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially"

print(re.split(r"(?<=\.) (?=\([a-z]\))", s))

Output:

['(a) be given by XXX for its election per scenario 3.3(a).', 
 '(b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially']

I'd caution using this on a large or complex input because the likelihood of false positives is high.

---

Another idea: if you are guaranteed to have every letter of the alphabet to extract, are sure each letter will show up eventually and in order, and want to treat anything out of order as normal content, you could try building a mammoth regex:

import re
from string import ascii_lowercase

s = "(a) be given by XXX for its election per scenario 3.3(a). (b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially. (c) blah blah  (c) blah blah (d) asd ad(a) (b) (e) ee (b) (a) (d) (f) (f) fff f ff (g) (a) gggg (h) hhhh (b) (i) iii i i (i) i (j) jjj (k) k (l) ll (a) (b) (x) (m) mm (n) nn (o) oo) () () (p) ppp (A) (B) (Q) (q) qq (r) rr (s) ss (t) tt( u ) (u) uu (v) vvv (ww) (w) ww (x) xx (y) yy (z) zzz"

pattern = "".join([f"((?: |^)\({l}\) .+)" for l in ascii_lowercase])

for result in re.findall(pattern, s)[0]:
    print(result.strip())

Output:

(a) be given by XXX for its election per scenario 3.3(a).
(b) Second statement has a section 4.2(b) which might elect for scenario 2.4(a) potentially. (c) blah blah
(c) blah blah
(d) asd ad(a) (b)
(e) ee (b) (a) (d) (f)
(f) fff f ff
(g) (a) gggg
(h) hhhh (b) (i) iii i i
(i) i
(j) jjj
(k) k
(l) ll (a) (b) (x)
(m) mm
(n) nn
(o) oo) () ()
(p) ppp (A) (B) (Q)
(q) qq
(r) rr
(s) ss
(t) tt( u )
(u) uu
(v) vvv (ww)
(w) ww
(x) xx
(y) yy
(z) zzz

This still makes some sweeping assumptions about the input, but might be worth playing around with; consider it a proof of concept.

Newlines are another issue to think about, if present (among many other things). Long story short, writing a parser by hand might be a better bet than regex.