tatt为多个语音到文本（stt）服务创建了一个统一的api。
tatt的Python项目详细描述
转录所有内容™
tatt是一个用于创建和管理语音文本转换的cli。
安装
pip install tatt
依赖关系
- 要转录的录音。
- a）一个aws帐户或b）一个带有语音对文本api和 已启用云存储。
ffmpeg对于某些操作，但是tatt将在需要时通知您。
用法
列出所有命令
$ transcribe --help
Usage: transcribe [OPTIONS] COMMAND [ARGS]...
Options:
--help Show this message and exit.
Commands:
get Downloads and/or saves completed transcript.
languages
list Lists available STT services.
services Lists available speech-to-text services.
status Check the status of a transcription job.
this Sends a media file to be transcribed.
列出所有STT服务
$ transcribe services
Here are all the available speech-to-text services:
amazon $0.006 per 15 seconds
google $0.009 per 15 seconds
获取转录的内容
$ transcribe this <path_to_media_file> <service_name>
Okay, transcribing <path_to_media_file> using <service_name>...
Okay, job <job_name> is being transcribed. Use "get" command to download it.
$ transcribe this --help
Usage: transcribe this [OPTIONS] MEDIA_FILEPATH SERVICE_NAME
Sends a media file to be transcribed.
Options:
--punctuation only for Google Speech, defaults to True
--speaker-id / --no-speaker-id only for google and amazon, defaults to True
--num_speakers INTEGER only for google and amazon, defaults to 2
--model TEXT only for Google Speech, defaults to
"phone_call"
--use-enhanced only for Google Speech, defaults to True
--language-code TEXT only for google and amazon, defaults to en-
US
--help Show this message and exit.
列出成绩单
$ transcribe list
Service Job Name Status
------- -------- ------
amazon tester.mp3 IN_PROGRESS
amazon messed_up.mp4 FAILED
amazon done_test.mp3 COMPLETED
amazon also_done.MP3 COMPLETED
google hey_there.mp3 COMPLETED
$ transcribe list <job_name>
Service Job Name Status
------- -------- ------
amazon <job_name> IN_PROGRESS
获取完整的成绩单
$ transcribe get <job_name> # prints to stdout
{'accountId': '416321668733',
'jobName': 'a1bace5e-8b08-4ce4-b08c-834a23aafcf1',
'results': {'items': [{'alternatives': [{'confidence': '0.9774',
'content': 'Hi'}],
'end_time': '1.5',
'start_time': '1.23',
'type': 'pronunciation'},
{'alternatives': [{'confidence': '0.9429',
'content': 'is'}],
'end_time': '1.71',
'start_time': '1.5',
'type': 'pronunciation'},
...
$ transcribe get --save <job_name>
Saved transcript to <job_name>.json
支持的服务
计划
欢迎加入QQ群-->： 979659372
