regex101: Distinguish torrent files (series vs movies)

Regular Expression

^
# get the title of this movie or series
(?P<title> 
  [-\w'"]+
  # match separator to later replace into correct title
  (?P<separator> [ .] ) 
  # note this *must* be lazy for the engine to work ltr not rtl
  (?: [-\w'"]+\2 )*?
)
# start of movie vs serie check
(?:
  # if this is an episode, lets match the season 
  # number one way or another. if not, the year 
  # of the movie
  (?:
    # series. can be a lot prettier if we used perl regex...
    # make sure this is not just a number in the title followed by our separator. 
    # like, iron man 3 2013 or my.fictional.24.series
    (?! \d+ \2 )
    # now try to match the season number
    (?: s (?: eason \2? )? )?
    (?P<season> \d\d? )
    # needed to validate the last token is a dot, or whatever.
    (?: e\d\d? (?:-e?\d\d?)? | x\d\d? )? |
    # this is likely a movie, match the year
    (?P<year> [(\]]?\d{4}[)\]]? ) 
  )
  # make sure this ends with the separator, otherwise we 
  # might be in the middle of something like "1080p"
  (?=\2) |
  # if we get here, this is likely still a movie.
  # match until one of the keywords
  (?= BOXSET  | XVID   | DIVX | LIMITED   | 
      UNRATED | PROPER | DTS  | AC3 | AAC | BLU[ -]?RAY | 
      HD(?:TV|DVD) | (?:DVD|B[DR]|WEB)RIP | \d+p |Â [hx]\.?264
  )
)

Unit Tests

There are no unit tests, click to add some.