[FEATURE] Adjust the project structure, and allow installation via pip

[try-agaaain]

Description

The structure of the project was reorganized in this PR. The main files of the current project are as follows:

.
├── LICENSE
├── README.md
├── data_save
│   ├── funcs.json
│   └── vector_database
├── gpttrace
│   ├── GPTtrace.py
│   ├── __init__.py
│   ├── __main__.py
│   ├── cmd.py
│   ├── config.py
│   ├── execute.py
│   ├── generate.py
│   └── utils
├── pyproject.toml
└── requirements.txt

• The data_save folder is used to store additional data, such as function call information (funcs.json) and vector database files.
• The gpttrace folder contains the main code of the project, where GPTtrace.py defines the entry function for GPTtrace.
• cmd.py, execute.py, and generate.py correspond to the three options of the gpttrace command: --cmd, --execute, and --generate, respectively.
• config.py is the project's configuration file.
• The utils folder contains other miscellaneous functions.
• pyproject.toml defines the project's build environment and contents.

The reorganized code structure has become clearer, and any further suggestions for improvement are welcomed.

Using the pyproject.toml file, gpttrace can be built using python -m build and uploaded to PyPI. Refer to https://packaging.python.org/en/latest/tutorials/packaging-projects/ for more information. #9

Currently, gpttrace can be installed using a more convenient way:

pip install gpttrace==0.1.2

I have tested the above command and it currently installs and runs the gpttrace command properly.

I have registered a PyPI account, and currently, gpttrace is hosted under my account. Additionally, I have applied for an organization named eunomia-bpf. Once this application is approved, I will transfer the gpttrace project to the eunomia-bpf organization.