RegEx - 如何仅获取在大输出中重复的多行文本块？

Question

I'm parsing a large output (25 MB provided here ) from a quantum chemistry software. The software performs a calculation using two methods: CASSCF and NEVPT2. Each method performs the same calculation, leading to different results. Actually, I've set my script to run the calculation several times for different configurations, so I have something organized like this in the end:

JOB 1
CASSCF RESULTS
***
Lots of text
***
end
NEVPT2 RESULTS
***
Lots of text
***
end

JOB 2
CASSCF RESULTS
***
Lots of text
***
end
NEVPT2 RESULTS
***
Lots of text
***
end
………………
JOB 31
CASSCF RESULTS
***
Lots of text
***
end
NEVPT2 RESULTS
***
Lots of text
***
end

I only want the NEVPT2 results and I've set my regular expression as this one (applied to the actual output (my example above is just to show the organization):

NEVPT2_Section = r"(?:AILFT MATRIX ELEMENTS \(NEVPT2\)\n-+\n\n)([\s\S]*)(?:\n\n--------------\nCASSCF TIMINGS)"
NEVPT2_Section_mathes = re.finditer(NEVPT2_Section, inp_content, re.MULTILINE)

for xyz in NEVPT2_Section_mathes:
    my_xyz = xyz.group(1)
    print(my_xyz)

If I'm working with a file that has only one job it works fine, starting from “NEVPT2 RESULTS” and stopping at the first “end” but, the multi-job file finds the first “NEVPT2 RESULTS” and goes on until the last “end”, catching everything in between.

So, after wasting the whole Sunday trying to figure this out, I'm asking for your advice, guys. How can I get only the bits from each NEVPT2 section?

Answer 1

You could use

^NEVPT2.+?^end

in single and multiline mode, see a demo on regex101.com .

Answer 2

As an alternative you could match the line at the beginning ^NEVPT2.*\\n and continue matching all lines that do not start with end using a negative lookahead (?!end$) using the multiline flag.

^NEVPT2.*\n(?:(?!end$).*\n)*end$

Regex demo | Python demo

For example

NEVPT2_Section = r"^NEVPT2.*\n(?:(?!end$).*\n)*end$"
NEVPT2_Section_mathes = re.finditer(NEVPT2_Section, inp_content, re.MULTILINE)

for xyz in NEVPT2_Section_mathes:
    print(xyz.group())