Function model_docs

bmtools/tools/hugging_tools/api.py:318–337

returns a document about the usage, examples, and available models of existing API. Every time before you use the API, you should get the document of EXISTING API ONLY and ONLY use the available models in the document. task: the name of API, includes %s retur

(task: str)

    @tool.get("/model_docs")
    @format_docs(str(task_list))
    def model_docs(task: str) -> str:
        '''returns a document about the usage, examples, and available models of existing API.

        Every time before you use the API, you should get the document of EXISTING API ONLY and ONLY use the available models in the document.

        task: the name of API, includes %s
        
        return the document of the API.

        example:
        Action: model_docs
        Action Input: {"task" : "question_answering"}
        Observation: "question_answering is a function that uses a pre-trained language model to answer questions based on a given context.\n        You can choose one model from: 'distilbert-base-uncased-distilled-squad', 'deepset/minilm-uncased-squad2', 'etalab-ia/camembert-base-squadFR-fquad-piaf'.\n\n        Action Input: {\"model_id\" : \"distilbert-base-uncased-distilled-squad\", \"question\" : \"When did the first moon landing occur?\", \"context\" : \"The first manned moon landing was achieved by the United States on July 20, 1969, in the Apollo 11 mission.\"}\n        "
        '''
        with open(f'{DIRPATH}/{DOCS_PATH}', 'r') as f:
            docs = json.load(f)
        if task in docs.keys():
            return docs[task]
        else:
            return "The function doesn't exist. Please input the valid function name."
        
    return tool