创建身份&读写公私有数据

Github

https://github.com/auliwenjiang/agentcp/blob/master/create_profile.py

1、使用方法

1)、安装agentcp库

pip install agentcp

2)、修改create_profile.py中修改加密种子

seed_password修改为自定义的加密种子，agentcp库会根据加密种子对后续用户私钥进行加密

3)、修改create_profile.py中"profile_json_data"智能体描述信息

包括但不限于name(agent名称)、avaUrl(agent头像)、description(agent描述)、version(agent版本)、capabilities(agent能力)

4)、执行create_profile.py文件

python create_profile.py

2、详细介绍

1)、创建身份(aid)

a、create_profile.py

介绍：create_profile.py包含身份创建，身份管理，agentprofile.json自动生成

from datetime import datetime, timezone
import agentcp
from pathlib import Path
import json

def create_financial_analyzer_json(publisherInfo):
    """创建智能体能力、权限描述"""
    profile_json_data = {
        "publisherInfo": publisherInfo,
        "avaUrl": "https://example.com/avatar.jpg",
        "version": "1.0.0",
        "lastUpdated": datetime.now(timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ"),
        "name": "智能体搜索",
        "description": "搜索网络中存在的各种能力的智能体",
        "capabilities": {
            "core": ["搜索智能体"],
            "extended": []
        },
        "llm": {
            "model": "qwen-plus",
            "num_parameters": "",
            "quantization_bits": "",
            "context_length": "",
        },
        "references": {
            "knowledgeBases": [""],
            "tools": [""],
            "companyInfo": [""],
            "productInfo": [""]
        },
        "authorization": {
            "modes": ["free"],
            "fee": {},
            "description": "当前智能体免费使用，无费用",
            "sla": {}
        },
        "input": {
            "types": ["content"],
            "formats": ["json"],
            "examples": {
                "type": "content",
                "format": "text",
                "content": "搜索智能体：xxx"
            },
            "semantics": [""],
            "compatibleAids": ["*"]
        },
        "output": {
            "types": ["content"],
            "formats": ["markdown"],
            "examples": {
                "type": "content",
                "format": "markdown",
                "content": ""
            },
            "semantics": [""],
            "compatibleAids": [""]
        },
        "supportStream": True,
        "supportAsync": True,
        "permission": ["*"]
    }
    return profile_json_data

if __name__ == "__main__":
    acp = agentcp.AgentCP(".", seed_password="888777", debug=False)
    agentid_list = acp.get_aid_list()
    # ... 身份选择和创建逻辑

b、使用create_profile.py

python create_profile.py

2)、读写公有数据

import agentcp
acp = agentcp.AgentCP(".")
aid = acp.create_aid("aid.pub", "guest")
if aid == None:
    print("aid创建失败,请打开日志查看")
    exit(1)
public_data_path = aid.get_agent_public_path()
# todo

3)、读写私有数据

import agentcp
acp = agentcp.AgentCP(".")
aid = acp.create_aid("aid.pub", "guest")
if aid == None:
    print("aid创建失败,请打开日志查看")
    exit(1)
private_data_path = aid.get_agent_private_path()
# todo

4)、同步公有数据到ap

import agentcp
acp = agentcp.AgentCP(".")
aid = acp.create_aid("aid.pub", "guest")
if aid == None:
    print("aid创建失败,请打开日志查看")
    exit(1)
aid.sync_public_files()