I have no idea about a PDF Recovery Software.

When the downloaded file is corrupted, that means there was an error in the downloading process itself and file rebuild after downloading (for resume-supported sites) fails to integrate the parsed file locations. In cases such as these, what I do is simply restart the file download and let it run its due course until the file is integrally complete.

Attempting to reconstruct a PDF file by extracting its contents into a text editor then recreate it using a PDF writer (I use CutePDF; and when the document production is mine, there always is a watermark) is tedious... the integrity of the original file (most specially when there are graphics and in multiple columns) is compromised. Personally, that's a no-no since I need to pinpoint the exact page (plus the URL when necessary) when citing the document as a reference.

[Aside 1: I literally had to ask the National Academy of Sciences on how to cite their books that I download in PDF format using the MLA format given the length of the list of editors, the committees and all.]

[Aside 2: I have a healthy fear of PDF attachments in unsolicited emails.]

But when I browse e-books that are displayed in HTML format, I print it in PDF format to make sure I extract it all. The disadvantage of this approach is the configuration of the PDF writer with regard to fonts prevails over my personal preference.