Skip to content

dictation_dictionary

diapason.speech.dictation_dictionary

User dictation dictionary — STT replacement words (Diapason-style).

Stored as JSON (default ~/.diapason/dictation_dictionary.json). Not a filler lexicon — fillers live in dictate_polish.py.

apply_dictionary(text, entries=None, *, path=None, bump_usage=True)

Replace STT mistakes using case-insensitive, word-boundary matches.

learn_from_correction(original, corrected, *, path=None, locale='')

Infer dictionary entries from an STT→user-corrected pair.

Uses difflib opcodes to find short replace spans (1–3 tokens). Skips filler-only and tiny tokens. Upserts into the dictionary file.

transcription_hints(entries=None, *, path=None, limit=40)

Canonical words for STT boosting prompts.