[KPhotoAlbum] Multiple versions of an image

[Joe]

On 12/10/2010 05:23 PM, Aleksi Halkola wrote:
>
> Hello again!
>
> I now got the solution I was looking for. The problem with my previous 
> regexp was that I had not considered that the filename consists also 
> of the folder the image is in. With the these regexps all the versions 
> are correctly found.
>
> "File versions search regexp:"
> (.*)/(.{8})(.*)
> "Original file replacement text:"
> \1/\2.nef
>
> The (.*)/ puts the folder into \1/, the (.{8}) takes the next 8 
> characters (the basename of the original) and puts them into \2, the 
> (.*) discards all the rest (which I could get with \3). The original 
> name can then be reconstructed using \1/\2.nef.
>
> I always wanted learn regular expressions but never really had any use 
> for them. Now I finally got the chance :).
>
> Cheers,
>
> Aleksi
>
>
> 09.12.2010 23:23, Aleksi Halkola kirjoitti:
>>
>> Hello Andreas,
>>
>> thanks for your reply. It gave me lots of new ideas :). One that works
>> with sed is:
>> echo dsc_8084.jpg | sed 's/\(........\).*/\1.nef/g' -> dsc_8084.nef
>> echo dsc_8084_v2.jpg | sed 's/\(........\).*/\1.nef/g' -> dsc_8084.nef
>> echo dsc_8084_v2.nef | sed 's/\(........\).*/\1.nef/g' -> dsc_8084.nef
>>
>> i.e.
>> "File versions search regexp:"
>> (........).*
>> "Original file replacement text:"
>> \1.nef
>>
>> This doesn't group the photos in kphotoalbum for some reason. My ideas
>> was that since most of my photos (and all from digital cameras?) have
>> 8.3 letter file names the (........) matches the base of the name so
>> that I can reconstruct the original with \1.nef and the .* just removes
>> the rest of the versioned file name (either .jpg or _v?.jpg).
>>
>> Anyways, the solution I found yesterday works for most cases so I'm not
>> too bothered.
>>
>> Cheers,
>>
>> Aleksi
>>
>>
>> 08.12.2010 23:59, Andreas Neustifter kirjoitti:
>>> Hi Aleksi!
>>>
>>> On 8 December 2010 23:09, Aleksi Halkola<halkola at gmail.com> wrote:
>>>>
>>>> I found a solution that I think will work pretty well for me. I took
>>>> inspiration from http://www.hardakers.net/kphotoalbum/#sec-4.2,
>>>>
>>>> I used the following settings:
>>>> "File versions search regexp:"
>>>> (.*)_v[0-9]*\.(.*)
>>>> "Original file replacement text:"
>>>> \1.nef
>>>> [...]
>>>> I learned
>>>> regular expressions only this evening so if anybody knows how I can
>>>> catch
>>>> both dsc_8237.jpg dsc_8237_v2.jpg as versions of dsc_8237.nef I'd be
>>>> very
>>>> happy.
>>>
>>> Those regular expression do not need to match the whole file, mine for
>>> example are "_v[0-9]*\." and "." (a single dot). This way one can
>>> insert a version between the file name and the extension and this file
>>> is then recognised as new version of an old file.
>>> (This is a little fiddly, it assumes that the filename has only one
>>> dot in it, which is the case with my files.)
>>>
>>> I was not aware of the regular expression supporting named groups, so
>>> in this case maybe "(.*)_v[0-9]*\.(.*)" and "\1.\2" would work? (The
>>> two "(.*)" each form a group of characters that can be accessed in the
>>> second expression as \1 and \2. At least it seems that this is
>>> possible from your example.)
>>>
>>> Hope that helps,
>>>
>>> Andi
>>> _______________________________________________
>>> KPhotoAlbum mailing list
>>> KPhotoAlbum at mail.kdab.com
>>> http://mail.kdab.com/mailman/listinfo/kphotoalbum
> _______________________________________________
> KPhotoAlbum mailing list
> KPhotoAlbum at mail.kdab.com
> http://mail.kdab.com/mailman/listinfo/kphotoalbum
>
Regular expressions are more than a bit cryptic, but if you ever need to 
manipulate text files, they're amazing.

If you ever do any work with Linux, they're invaluable parts of the bash 
shell, most text editors, and many of the command line tools.

In Windows, Cygwin gives you a Linux-like environment with a shell 
command line processor that gives you access to REs. Programs such as 
awk and sed (invaluable text processing programs in their own right - 
also available for Windows) also let you use regular expressions.

Joe