Introduction

'A2A' (API to AI) is a framework designed to facilitate programming for Large Language Models (LLMs).

• It provides a consistent interface for common programming tasks, making it easier for LLMs to write correct code.
• For code written by LLMs:
  • It can be executed directly, without the need for additional environment installation or configuration.
  • It can also be directly deployed as a service.
  • Furthermore, it can generate scheduled tasks and other automated tasks.

This project compiles a collection of common programming task use cases, allowing different LLMs to write the code for these tasks. This process allows for the evaluation of the programming capabilities of various LLMs.

All code in this project, including the project's own code and the code for various programming tasks, is written by LLMs :)

Directory Structure

• models.txt: A list of the Large Language Models participating in the programming.
• app/: The project's control code, managing the writing and execution of each use case, etc.
• mock/: Mock API services.
• case/: Code for various scenario use cases.
  • data_process/: Data processing use cases.
    • import_json_to_db/: Use case for importing data from a JSON file into a database.

Use Cases

For each use case, the following is included:

• task.md: A description of the use case, which the LLM will use to write the code.
• conf/*: Configuration files related to the use case.
• datas/*: Data files related to the use case, if any.
• usage.txt: Statistical information such as tokens and time used by the LLM when writing the code.
• task.*.js: Code written by different LLMs.