More control over frame number identification

[rsgalloway]

Need the abiliy for more control over identifying what is and what is not a frame number for given filenames that may contain other numeric characters which can be falsely identified as a frame number and grouped into a sequence.

• It would be good to be able to both positively and negatively identify frame numbers for the following cases:
  • a known fixed, frame number pattern
  • a known non-frame number pattern, but unknown or variable frame number patterns
• It would be good if the control could be set without changing any code, for example using an environment variable.

( Duplicate of #4 )

[rsgalloway]

I think the simplest thing to do here would be to add a new optional env var to positively identify frame number patterns and a corresponding one for non-frame pattern.

[broganross]

What about one env var with a ":" or ";" separated list of patterns? Seems unlikely that someone would put one of those characters in the path.
Example:
PYSEQ_PATTERN="\w+.\d{6}.\w{3,4}:left|right"

[rsgalloway]

I'd like to stick to valid regex syntax if possible. On second thought though it should be possible to have both of these things expressed in the same regex string, e.g.

digits_re = re.compile(r'\d+(?!_v\d+_)')

would negatively identify files with _v001_ as the frame number. So I think the answer is to just support a env var to override the default, e.g.

digits_re = re.compile(os.getenv('PYSEQ_FRAME_REGEX', r'\d+'))

[broganross]

I might have been thinking of something else with the multiple regexs. After rereading the env var does make the most sense to me.