bilive: Unsupervised live recording and automatic slicing and uploading tools for B station

General Introduction

bilive is a tool designed specifically for live streaming recording on B.com, providing extremely fast live streaming recording, auto-slicing, pop-up rendering and subtitle generation. bilive is compatible with ultra-configurable machines and supports 7x24 hours of unmanned video recording. The tool is compatible with ultra-low-configuration machines, supports 7x24 hours of unattended recording, automatically recognizes and renders pop-ups and subtitles, and automatically slices and uploads them to B. bilive uses pipelining to process the video, ensuring that the recording is almost synchronized with the live broadcast, making it suitable for multi-room recording and low-hardware-requirement scenarios.

Function List

• Extremely fast recording: A pipeline process is used, and the recording is within half an hour of the live broadcast.
• Multi-Room Recording: Record multiple live streams at the same time, including regular pop-ups, paid pop-ups and gift messages.
• Automatic slicing: Slices and dices highlights based on pop-up density and automatically generates interesting captions.
• pop-up rendering (computing): Automatically converts XML to ASS pop-up files and renders them to video.
• Subtitle Generation: Automatically recognizes speech and generates subtitles using OpenAI's Whisper model.
• Automatic uploading: Automatically uploads recorded and sliced videos to the B site.
• Low Configuration Compatibility: No GPU required, minimum configuration of single core CPU and minimum RAM to run.

Using Help

Installation process

1. environmental preparation::
   • Operating System: Ubuntu 22.04 and above is recommended.
   • Hardware Requirements: Minimum configuration is 2-core CPU, 2G RAM, 40G hard disk and 3Mbps bandwidth.
2. Installation steps::
   • Cloning Project:git clone https://github.com/timerring/bilive.git
   • Go to the project catalog:cd bilive
   • Install the dependencies:pip install -r requirements.txt
   • Configuration file: modify as neededsettings.tomlDocumentation.

Instructions for use

1. Start recording::
   • (of a computer) runrecord.shThe script initiates recording:. /record.sh
   • Video and pop-up files are automatically saved during the recording process.
2. Automatic slicing and uploading::
   • (of a computer) runscan.shThe script scans the video folder and initiates slicing and uploading:. /scan.sh
   • The system automatically slices the pop-up density and generates interesting titles and content.
3. Pop-up and subtitle rendering::
   • The system automatically converts the XML pop-up file to ASS format and renders it into the video.
   • Use Whisper models to automatically recognize speech and generate subtitles to be rendered into the video.
4. Multi-Room Recording::
   • Support recording multiple live room contents at the same time, multiple live room IDs can be set in the configuration file.

Detailed Operation Procedure

1. Recording Settings::
   • existsettings.tomlfile to configure the live room ID and recording parameters.
   • After starting the recording, the system will automatically listen and record the content of the specified live room.
2. Slicing and uploading settings::
   • Configure slicing parameters, such as pop-up density threshold and slice duration.
   • The system automatically slices according to the configuration and uploads the video through the B-site API.
3. Pop-up and subtitle settings::
   • Configure pop-up and subtitle rendering parameters such as font, color, and position.
   • The system automatically renders the pop-ups and subtitles and generates the final video file.