deblan-mirror/buzz
1
0
Fork 0
mirror of https://github.com/chidiwilliams/buzz.git synced 2024-06-03 08:22:12 +02:00
Code Issues Packages Projects Releases Wiki Activity

fix: openai transcription file and language

Browse source
This commit is contained in:
Chidi Williams 2024-01-11 09:06:49 +00:00
parent ad784c1be8
commit 43c3f66aa7
1 changed files with 21 additions and 19 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

40
buzz/transcriber/openai_whisper_api_file_transcriber.py
View file

@ -95,26 +95,28 @@ class OpenAIWhisperAPIFileTranscriber(FileTranscriber):
return segments
def get_segments_for_file(self, file: str, offset_ms: int = 0):
kwargs = {
"model": "whisper-1",
"file": file,
"response_format": "verbose_json",
"language": self.transcription_task.transcription_options.language,
}
transcript = (
self.openai_client.audio.transcriptions.create(**kwargs)
if self.transcription_task.transcription_options.task == Task.TRANSLATE
else self.openai_client.audio.translations.create(**kwargs)
)
return [
Segment(
int(segment["start"] * 1000 + offset_ms),
int(segment["end"] * 1000 + offset_ms),
segment["text"],
with open(file, "rb") as file:
transcript = (
self.openai_client.audio.transcriptions.create(
model="whisper-1",
file=file,
response_format="verbose_json",
language=self.transcription_task.transcription_options.language,
)
if self.transcription_task.transcription_options.task == Task.TRANSLATE
else self.openai_client.audio.translations.create(
model="whisper-1", file=file, response_format="verbose_json"
)
)
for segment in transcript["segments"]
]
return [
Segment(
int(segment["start"] * 1000 + offset_ms),
int(segment["end"] * 1000 + offset_ms),
segment["text"],
)
for segment in transcript["segments"]
]
def stop(self):
pass