Fixes #66 handle pdf specs

[MaximeLaffaire]

As described in #66 , some specs were missing from the PDF detection :

• the header can be anywhere in the 1024 first bytes of a file
• the header can also be %!PS−Adobe−N.n PDF−M.m

Since I override the base implementation of FileFormat.IsMatch, I added some unit tests to cover that. I also chose to not call base.IsMatch in the overriden method, it felt redundant and unoptimized 🤔

Also, I added two tests on real pdf files, you can check the headers by opening the files with a hex editor like Hex Fiend

[neilharvey]

Thanks, looks good - I'll merge the PR and push a new release when I get a chance. Cheers!