dictionary() should automatically recognise formats by file ending

[kbenoit]

From @conjugateprior:

While your complete knowledge is swapped in: should 'dictionary' work by file ending inference and convention like textfile, or explicitly as it currently does?

Good point, it would be sensible to include:

• .dic = LIWC
• .cat = Wordstat
• .ykd = Yoshikoder
• .lcd = Lexicoder

[adamobeng]

We don't currently support Lexicoder at all, right?

[adamobeng]

Except for Lexicoder, this is now implemented in PR 230.