Regex match specific word in string but exclude indexed versions

Question

I'm sure that if a solution exists for this then its out there somewhere but I can't find it. I've followed Python regex to match a specific word and had success in the first aspect but now am struggling with the second aspect.

I've inherited a horrible file format where each test result is on its own line. They are limited to 12 chars per record so some results are split into groups of lines e.g SITE, SITE1 and SITE2. I'm trying to parse the file into a dictionary so I can do more analysis with it and ultimately produce a formatted report.

The link above / code below allows me to match each SITE and concatenate them together but its giving me problems matching INS, INS 1 and INS 2 correctly. Yes the space is intentional - its what I have to deal with. INS is the test result and INS 1 is the limit of the test for a pass.

Is there a regular expression that would match

SITE > SITE True but SITE > SITE1 false

and

INS > INS True but INS to INS 1 false?

Here is the python code.

import re    
lines = ['SITE start', 'SITE1 more', 'SITE2 end','INS value1', 'INS 1 value2']
headings = ['SITE','SITE1',"SITE2", "INS", "INS 1"]
for line in lines:
    for heading in headings:
        headregex = r"\b" + heading + r"\b"
        match = re.search(headregex,heading)
        if match:
            print "Found " + heading + " " + line
        else:
            print "Not Found " + heading + " " + line

And here is some dummy data:

TEST MODE 131 AUTO SITE startaddy SITE1 middle addy SITE2 end addy
USER DB VISUAL CHECK P BOND RANGE 25A EARTH 0.09 OHM P LIMIT 0.10 OHM INS 500 V INS 1 >299 MEG P ... TEST MODE 231 AUTO SITE startaddy SITE1 middle addy SITE2 end addy USER DB VISUAL CHECK P INS 500 V INS 2 >299 MEG P ...

Sorry for the horrid formatting - its copied and pasted from what I am dealing with!

Answer 1

The problem is that INS pattern finds a partial match in INS in INS 1 or INS 2 etc.

In cases when you extract alternatives, it is customary to use alternations starting with the longest value (like INS \d+|INS), but in this case you are looking to obtain a list of all regex matches only excluding some overlapping heading matches.

To achieve that, there is a way to exclude that match by treating all headings items as regular expressions, and define the INS pattern as INS(?! \d) to make sure INS is not matched if it is followed with a space and a digit.

See the Python demo:

import re    
lines = ['SITE start', 'SITE1 more', 'SITE2 end','INS value1', 'INS 1 value2']
headings = ['SITE','SITE1',"SITE2", r"INS(?! \d)", "INS 1"]
headings=sorted(headings, key=lambda x: len(x), reverse=True)
for line in lines:
    print("----")
    for heading in headings:
        headregex = r"\b{}\b".format(heading)
        match = re.search(headregex,heading)
        if match:
            print "Found " + heading + " " + line
        else:
            print "Not Found " + heading + " " + line

Answer 2

Just to give an answer that might solve the problem while avoiding some of the tediousness, is this what you are trying to achieve?

import re

lines = ['SITE start', 'SITE1 more', 'SITE2 end','INS value1', 'INS 1 value2']
headings = ['SITE','SITE1',"SITE2", "INS", "INS 1"]

headings_re = re.compile(r"(SITE\d? )?(INS( \d)?)? (.*)") 
# build by hand, only works if SITE and INS are the literal identifiers 

site = []
ins = []

for line in lines:
  match = headings_re.match(line)
  if match:
    if match.group(1):
      site.append(match.group(4))
    elif match.group(2):
      ins.append(match.group(4))
    else:
      print("something weird happened")
      print(match.group(0))
  else:
    print("something weird happened")
    print(line)

print("SITE: {}".format(" ".join(site)))
>> SITE: start more end
print("INS: {}".format(" ".join(ins)))
>> INS:  value1  value2