Fix getting filename when extracting embedded files

[pietermarsman]

Description

The File Specification dictionary (PDF reference 1.7, chapter 3.10.2) contains a string that represents a uniform resource allocator (key F). It optionally contains a unicode text string for the same URL (key UF).

Pdfminer.six prefers the UF key, but does not fall back properly on the F key when UF does not exist.

Fixes #152

How Has This Been Tested?

• With the PDF in Incorrect parsing of filenames in file specifications #152.
• With all the other PDF's that I have in my downloads folder, but most dont contain embedded files.

Checklist

• I have added tests that prove my fix is effective or that my feature works
• I have updated the README.md and other documentation, or I am sure that this is not necessary
• I have added a consice human-readable description of the change to CHANGELOG.md
• I have added docstrings to newly created methods and classes
• I have optimized the code at least one time after creating the initial version

[pombredanne]

LGTM