I. Love. You.
Now... if we could add the extraction to the Edit Book Details window (like to the right of the ISBN text box) and then have an option to download metadata if an ISBN is found... I would have your baby.
(Although, yes, I can edit a batch and then download a batch. I tend to edit one at a time so I think one at a time.
)
This has worked beautifully on 480 out of 500 books. And the 20 that didn't work I confirmed were PDFs where the contents were JPG images rather than text -- so no way for the regex to pick up the ISBN.
Oh, some sort of progress indicator would be beneficial. (Dunno if possible.)