Add ServiceFactory to support quick setup for tools model&agent.

[DavdGao]

Description

Target: Use ServiceFactory to help generate functions and JSON SCHEMA format dictionary that models can use.

Usage: Taking bing search as an example, the first argument to ServiceFactory.get is the function, following by keyword arguments that should be specified by developers. Then ServiceFactory.get returns

• a function for llms to use and
• a dict to introduce the function in JSON Schema format, which is compatible to OpenAI API

For open-source models, developers need to convert doc_dict to their required format.

_, doc_dict = ServiceFactory.get(
          bing_search,
          num_results=3,
          bing_api_key="xxx",
      )

print(doc_dict)

{
    "type": "function",
    "function": {
    "name": "bing_search",
    "description": "Search question in Bing Search API and return the searching results",
      "parameters": {
          "type": "object",
          "properties": {
              "question": {
                  "type": "string",
                  "description": "The search query string.",
              }
          },
          "required": ["question"],
      }
   }
}

Note

• The descriptions are extracted from docstring, which means for developer-defined functions, docstring is required.
• The types and default values of the arguments are extracted from function definition rather than docstring, which may be more accurate.
• Documents will be updated after the modifications are confirmed.

Besides the major modification, these changes are also included in this PR:

1. Add unit tests for service factory.
2. Rename service/code to service/code_ to avoid duplicated name during debug in pycharm.
3. Modify some service functions for llm to use.

Checklist

Please check the following items before code is ready to be reviewed.

• Code has passed all tests
• Docstrings have been added/updated in Google Style
• Documentation has been updated
• Code is ready for review

[pan-x-c]

LGTM