I have a problem where I need to find if a word(for example "output") is present in a snake_case word.

Which means the regex must be capable of matching all of the following situations-

• output
• output_of_my_program
• my_output_from_program
• program_output

i.e. output, output_, output, and _output need to be matched

currently I have three individual regex patterns to cover all the cases, which are-

• "^output[^a-z]_?"
• "_output_"
• "_output$"

I however have tried to combine the three into one, which is [a-z]*_?[^a-z]output_?[a-z]* but this fails for certain cases. Is it possible to combine the three patterns into one in this case?

Edit:The other keywords I am interested in are "in", "input", "out" and the challenge is to avoid matches with words such as "introspection" and other cases such as "my_programoutput"

CodePudding user response：

You don't include it in your test cases, but I assume part of the intent is to explicitly not match a string like my_floutputs_from_program. You could do this with a tricky regex, but I'd just use split and in:

for s in (
    'output', 
    'output_of_my_program', 
    'my_output_from_program', 
    'program_output', 
    'input', 
    'my_floutputs_from_program'
):
    print(f"{s}: {'output' in s.split('_')}")

output: True
output_of_my_program: True
my_output_from_program: True
program_output: True
input: False
my_floutputs_from_program: False

CodePudding user response：

Is there a reason why 'output' in s would not work? You do not have to use regex to solve this, unless it is necessitated as part of the project.

>>> strings = [
...     'output',
...     'output_of_my_program',
...     'my_output_from_program',
...     'program_output',
...     'input'
... ]
>>>
>>> for s in strings:
...     print('output' in s)
...
True
True
True
True
False

CodePudding user response：

If you can use the re library, you can just use the search function with the string required to find.

import re

txt = "my_output_from_program"
result = re.search("output", txt)

CodePudding user response：

You can use a pattern with either a word boundary, or a postive assertion for _ on the left or the right if you don't want partial matches.

(?:\b|(?<=_))output(?:\b|(?=_))

The pattern matches:

• (?:\b|(?<=_)) Match either a word boundary or assert _ to the left
• output Match literally
• (?:\b|(?=_)) Match either a word boundary, or assert _ to the right

See a regex demo and a Python demo.

import re

pattern = r"(?:\b|(?<=_))output(?:\b|(?=_))"
strings = [
    "output",
    "output_of_my_program",
    "my_output_from_program",
    "program_output",
    "my_programoutput"
]

for s in strings:
    m = re.search(pattern, s)
    if m:
        print(f"{s} --> {m.group()}")

Output

output --> output
output_of_my_program --> output
my_output_from_program --> output
program_output --> output