promptdojo_

The voice fingerprint — a client's voice, measured from their own copy — step 8 of 8

Checkpoint

One last thing before we move on. Same surface as a write step — but the lesson doesn't complete until this passes.

Checkpoint: all three Windrow samples are fingerprinted. Write derive_pack(prints) that turns the measurements into the rules section of the voice pack:

  • house_avg: mean of the samples' avg_words, rounded to 1 decimal.
  • max_sentence_words: round(house_avg * 1.5) — using the rounded house_avg.
  • exclamations_allowed: True if ANY sample has exclaims > 0.
  • contractions_required: True if the pooled contraction rate — total contractions across the samples divided by total sentences — is at least 0.2. Same threshold as step 5, now computed from the raw counts.

Expected output:

house avg: 10.2 words/sentence
max_sentence_words: 15
exclamations_allowed: False
contractions_required: True

full-screen editor opens — close anytime to keep reading.