The built-in local Parakeet transcription engine used by the gateway.
The built-in local Parakeet transcription engine used by the gateway.
(clean-transcript text)Deterministically remove ASR hesitation tokens and adjacent stutter runs from plain
TEXT — the same pass clean-words makes over timed words, for a transcript that
arrived without any times.
Deterministically remove ASR hesitation tokens and adjacent stutter runs from plain TEXT — the same pass [[clean-words]] makes over timed words, for a transcript that arrived without any times.
(clean-words words)Deterministically remove ASR hesitation words and adjacent stutter runs, keeping every surviving word exactly where it fell in the recording.
The cleaning happens on WORDS and not on finished text, because the times are the point: strip a stutter out of a string and every timestamp after it belongs to a word that is no longer there.
Deterministically remove ASR hesitation words and adjacent stutter runs, keeping every surviving word exactly where it fell in the recording. The cleaning happens on WORDS and not on finished text, because the times are the point: strip a stutter out of a string and every timestamp after it belongs to a word that is no longer there.
The gateway's fixed local transcription engine descriptor.
The gateway's fixed local transcription engine descriptor.
(transcribe {:keys [audio-path on-progress]})Transcribe one recording through the gateway's Parakeet engine: the words, and the timed lines they fall into so a player can follow them.
Transcribe one recording through the gateway's Parakeet engine: the words, and the timed lines they fall into so a player can follow them.
(words->segments words)Timed words grouped into the LINES a player can highlight: {:start :end :text},
in order, each one a piece of the same transcript.
A line ends at sentence punctuation, at a pause of [[segment-pause-seconds]], or once it has grown past what one highlighted line usefully holds — so a memo spoken without a single full stop still follows the audio in readable pieces.
Timed words grouped into the LINES a player can highlight: `{:start :end :text}`,
in order, each one a piece of the same transcript.
A line ends at sentence punctuation, at a pause of [[segment-pause-seconds]], or
once it has grown past what one highlighted line usefully holds — so a memo spoken
without a single full stop still follows the audio in readable pieces.(words->text words)The words as one line of prose.
The words as one line of prose.
cljdoc builds & hosts documentation for Clojure/Script libraries
| Ctrl+k | Jump to recent docs |
| ← | Move to previous article |
| → | Move to next article |
| Ctrl+/ | Jump to the search field |