Positive identification of a fillable PDF form among non-PDF form files

I'm looking for an programmatic approach to positively identify fillable PDF form(s) among non-PDF form files.

The options that I believe are available are:

• Parse the PDF code and content
• Parse the file for signature identification with a hex capable language such as Python
• Parse the file with a hex capable language such as Python to flag telltale signs

I'm looking for a solution not involving a human or visual confirmation, but something in the file or code that is a dead giveaway.  What is the difference between form pdfs and their non counterpart?