More precise file detection

[fan-tom]

Currently plugin uses first few characters to detect file type (Github or Bitbucket). But if it doesn't find Bitbucket-specific char sequencief, it asssumes it is Github file. And it becomes wrong assumption if there is long documentation comment header in file, so that first codeowners rule (which is of interest for file type detector) is not in that few first characters.

Actual behaviour: Bitbucket file is detected as being Github one which gives syntactic error

Expected Behaviour: Bitbucket file detected correctly no matter how far first rule is from beginning of a file