As is known to programmers, it is relatively tedious and labourious to write codes on Computer Graphics. For exemple, it may take hundreds of lines of C++ codes to generate VBO, VAO, shader, BEFORE a single triangle can be rendered. Combined with the complicated OpenGL API, things become even worse.
That's why I've decided to code the main functionality of Computer Graphics into Python as a practice. The advanced feature provided by python and numerous handy libs(numpy, etc) can greatly reduce the trivial but tedious parts of coding, and help we programmers focus on the core logic of computer graphics.
So far, I've implemented 6 stages of codes, each containing a basic but key feature in computer graphics. The six feature is directly in the root directory, not wrapped into folders yet.
Despite the convenience provided by Phthon, there's no denying that Python is a poor at perfomance, which is very import in Computer Graphics. I've tried numba, but it seems to work only for Python programs that is dense in numerical computing. Pypy maybe a choice in boosting the perfomance of these codes, which I've planned to try in the future.
Considering the limited speed of current code in this repository, it is recommended to use pyopengl as a learning platform to help you understand the fundamental logic and basic workflow of opengl, and then shift to C++ afterwards. I've found such switch interesting and beneficial.
The part below illustrate the first five stages.
修改了代码，成功封装到了app.py: class Application中
应用了VAO，VBO，改用shader进行渲染，并封装到了python类中
采用numpy传递数据。gl接口中原本填sizeof(data)的地方改为arr.nbytes即可
用RenderData封装VAO和VBO，没有封装EBO
智能识别顶点格式，shader智能传递uniform
see: shader.py, render_data.py, vert.glsl, frag.glsl
加入了纹理，并封装入texture.py: class Texture中。
- 现在会渲染两个纹理混合的一个立方体
纹理图像的加载采用pillow库，见import PIL
将变换、摄像机相关代码抽离出来，封装成了camera.py: class Camera并优化了代码逻辑，便于日后修改摄像机模式
会绘制xyz轴了
使用DearPyGui包添加了异步（多线程）的gui，解决了一些多线程的问题
- 现在只有一个选项
- 但基础框架已经建立了
完成了phong光照模型，并配上了可以动态修改参数的gui