Chinese language recognition on Python and Baidu API

The script on the entrance receives an audio or video file containing Chinese conversation.
Further, he extracts the sound path and recognizes through the official API of the Chinese search giant Baidu.
Duration limits and technical audio requirements are subject to various file manipulations.
In particular, there is a limit to recognition not more than 1 minute per request.
To avoid this, the soundtrack is cut into pieces up to 60 seconds, provided that in each of them there are completed offers to the point.