Description:
------------
When checking PDF, in some case the mime type returned by this function will be wrong.
The thing is that some editors add stuff before the %PDF string in the file.
And this is exactly the issue. The %PDF string can be found in the first 1024 bytes andkeep the PDF valid.

I tried to remove the specials chars from the file and the function detect the correct mime type.



Test script:
---------------
- Create a PDF
- test the function
- it swill return application/pdf
- edit the file and add some chars at the beginning before %PDF 
- retry and it will return text/plain



Expected result:
----------------
application/pdf

Actual result:
--------------
it returns text/plain