Difficulty in Processing "Verified Commentary"

[eoppe1022]

Check out the output from this:

https://genius.com/Fleet-foxes-third-of-may-odaigahara-lyrics

# A tibble: 55 x 3
   track_title               lyric                                         line
   <chr>                     <chr>                                        <int>
 1 Third of May / Ōdaigahara Light ended the night, but the song remained     1
 2 Third of May / Ōdaigahara "And I was "                                     2
 3 Third of May / Ōdaigahara hiding by the stair                              3
 4 Third of May / Ōdaigahara half here                                        4
 5 Third of May / Ōdaigahara Half there                                       5
 6 Third of May / Ōdaigahara ", past the "                                    6
 7 Third of May / Ōdaigahara lashing rain                                     7
 8 Third of May / Ōdaigahara "And as the "                                    8
 9 Third of May / Ōdaigahara " would "                                        9
10 Third of May / Ōdaigahara petal white                                     10
# ... with 45 more rows

I'm not sure if this is consistent, but I thought I'd bring it up.

[JosiahParry]

Shoot. This is a problem causes from the annotation. The annotations are in a <p> tag, not a <span> tag. This makes it genius_url() think it's a new line.

On line 8 of the tibble, sky[e] is omitted because of the bracket. I currently have the function written so that lines with brackets are omitted. This is because they tend to be things like guest vocals or instrumentation etc.

[eoppe1022]

I'm gonna do some investigating, but I'm guessing if the set of brackets is on it's own line, it should always be omitted.

There's probably some regex for this.
Also, this is less robust but you could do something like:

Omit and skip line if characters between [] > 3 or includes ?, !, or other punct.
else delete brackets and contents but maintain line structure

EDIT: I realize this probably wouldn't solve anything. I'm gonna think hard about this though

[JosiahParry]

fixed in pr #24