oh sorry - i didnŽt notice, that it no longer catches for phrases - just for the digits.
Anyway - as i am sorting my collection: i got some books where the isbn could not be extracted but can easily be found via okular (kde pdf viewer - so it is ocrŽd and not only an image). + they are on the first 10 pages.
by the way - i also got a lot of ocrŽd djvuŽs where they could not be extracted but found via strg+f in my pdf-viewer. Did i read right that djvus wonŽt work at all?
Anyway: If you want i can collect those pdfs (and djvus) for you (i will simply print out the single page where i find the isbn with cups (linux pdf printer) to keep the data size small). But as i am doing like "5 books a day" this may take a while.
as for cbr/cbz - i know
- this was more a general comment than regarding isbn extraction. Only wanted to tell that i am not jet sure if calibre can help me with those ones. But as said - i may discuss this in another topic.