feat: refine the code in model description and fix some bugs in feedback.py

[WinstonLiyt]

Description

• Fixed the error in the feedback, allowing the code to run again.
• Improved the code logic, especially in how data_description and model_description are used in the feedback.
• Added the implementation for model_description (initialization and subsequent assignment).
• Added logic to parse model details such as name, description, formulation, architecture, variables, hyperparameters, and model type.

Motivation and Context

How Has This Been Tested?

• Pass the test by running: pytest qlib/tests/test_all_pipeline.py under upper directory of qlib.
• If you are adding a new feature, test on your own test scripts.

Screenshots of Test Results (if appropriate):

1. Pipeline test:
2. Your own tests:

Types of changes

• Fix bugs
• Add new feature
• Update documentation

---

📚 Documentation preview 📚: https://RDAgent--288.org.readthedocs.build/en/288/