Stop relying on label to decide the file is VTT

[JackLewis-digirati]

What does this change?

Previously, label was used to figure out if the file was in the VTT format, along with checking for various strings in it. This caused issues as it would decide this label is VTT:

{
"label": {
      "en": [
          "METS-ALTO transcription"
      ]
  }
}

This tightens up that validation to only rely on profile and label, as well as remove some redundant checks on text/vtt, as it already checks these same fields for vtt