Add support for negative frame ranges

[BigRoy]

Problem

When pyseq encounters a frame range that is negative it doesn't parse it correctly.

For example I have an application that saves out to the following format:

file.-002.jpg
file.-001.jpg
file.0000.jpg
file.0001.jpg
file.0003.jpg

Testing that it parses:

data = ["file.-002.jpg", "file.-001.jpg", "file.0000.jpg", "file.0001.jpg", "file.0003.jpg"]
print pyseq.get_sequences(data)
# [<pyseq.Sequence "file.-1-2.jpg">, <pyseq.Sequence "file.0-3.jpg">]

Solution

Parse digits including the - (minus) sign to ensure it is parsed as it should.
I quickly tested to change:

digits_re = re.compile(r'\d+')

To:

digits_re = re.compile(r'-?\d+')

And it parsed the above file sequence correctly as a single sequence. YAY! (Dirty fix?)

Though to have this as a fully reliable implementation I'm imagining some changes need to be made to uncompress and alike to be perfect. Similarly tests should be implemented to also ensure the functionality remains reliable over time.

[rsgalloway]

Hi Roy, thanks for using pyseq and for reporting this issue.

I'll need to do a bit more testing and write some new unit tests, but I tried your solution and I think it should be fine. There are unit tests in test/test_pyseq.py, and they all pass with this change, including uncompress.

[rsgalloway]

Looks like a little more work will be required here,

>>> s = pyseq.get_sequences(data)[0]
>>> s
<pyseq.Sequence "file.-2-3.jpg">
>>> for i in s: print i
... 
file.-001.jpg
file.-002.jpg
file.0000.jpg
file.0001.jpg
file.0003.jpg

The issue seems to be the sort here, so a fancier sorting function will be required:

https://github.com/rsgalloway/pyseq/blob/master/pyseq.py#L1045

>>> s = pyseq.Sequence(data)
>>> s[0]
<pyseq.Item "file.-002.jpg">