Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/qin2dim/armour-geetest
一种针对GeeTest滑动验证的高识别率解决方案
https://github.com/qin2dim/armour-geetest
geetest python3 recaptcha selenium sspanel sspanel-uim
Last synced: 6 days ago
JSON representation
一种针对GeeTest滑动验证的高识别率解决方案
- Host: GitHub
- URL: https://github.com/qin2dim/armour-geetest
- Owner: QIN2DIM
- License: gpl-3.0
- Created: 2021-07-21T16:33:08.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2022-04-27T14:24:04.000Z (over 2 years ago)
- Last Synced: 2024-10-18T08:15:50.774Z (3 months ago)
- Topics: geetest, python3, recaptcha, selenium, sspanel, sspanel-uim
- Language: Python
- Homepage:
- Size: 33.2 KB
- Stars: 47
- Watchers: 4
- Forks: 18
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# GeeTest Challenger
通过观看 [QIN2DIM's Fingers](https://www.wolai.com/pbhA77godFuXjofA1QC7JG?theme=light) 了解本项目的工作范围
https://user-images.githubusercontent.com/62018067/165540682-9321ebee-a02f-4988-8616-b14e8acb5113.mp4
## Introduction
[ArmourGeeTest](https://github.com/QIN2DIM/armour-geetest) 是一种针对 [GeeTest滑动验证 ](https://www.geetest.com/) 的高通过率解决方案;引入`动量收敛` 以及 `震荡` 等仿生算法解决二维空间中的像素对齐问题。当这个难倒了大批爬虫玩家的问题被抽象成`缺口识别`以及`像素对齐`两个指标时使用本方案进行百次实验:
- 当`缺口识别率`为100%时,`gt3`通过率为92%。失败案例中超半数由收敛超时引发,剩下的被怪兽吃掉了;
- 当`缺口识别率`为100%时,`gt2`通过率100%。仅在缺口被遮挡时失败,但此时倾向认为`缺口识别率`<100%;## Preview
1. 本项目依赖 `google-chrome` 完成挑战,请确保您的设备中已装有 **最新版谷歌浏览器**!
2. 本项目由 `webdriver-manager` 实现驱动托管。因此,你不需要关心浏览器驱动的存放路径问题,只要你的设备上装了谷歌浏览器,`webdriver-manager` 可以自动下载版本匹配的驱动并放置到绝对索引路径。## Usage
1. Clone project
```bash
git clone https://github.com/QIN2DIM/armour-geetest.git armor-geetest
```2. Initialize workspace
```bash
cd armor-geetest/ && pip install -r requirements.txt && cd src
```3. Run the demo
```bash
# armor-geetest/src/
python main.py demo --version=3 --debug=True --silence=False
```