Cleaning actual transcriptions produces the same output as cleaning expected transcriptions.

[KatieProchilo]

Description:

• The expected transcription currently:
  • is lowercase.
  • contains apostrophes in words that are contractions.
  • replaces "okay" with "ok".
  • replaces hyphens with a space.
  • errors out if any other special characters are present.
  • allows digits (which we may want to change since we expect lexical responses from the STT service).

Acceptance criteria:

• Determine what special characters are returned from the STT service, and identify any patterns.
• Account for these patterns when cleaning the actual transcription.
• The actual transcription should match the expected transcription. Right now it:
  • is lowercase.
  • is lexical (e.g. the number 2 is returned as "two").

[KatieProchilo]

@HobbyProjects this is first on my list to get to today - I know this is of interest to you. I'll add in whatever oddities you've found so far, but let me know if there are new ones.

[HobbyProjects]

Sent you the list.