赞
踩
最近opc 基金会将召开一个会议,主题是”OPC UA meets IT“。由此可见,工业自动化行业也开始研究和评估chatGPT带来的影响了。
本文谈谈本人对OPC UA 与chatGPT结合的初步实验和思考。
chatGPT 的确非常强大了,使用自然语言描述一个设备的信息模型,它能够直接生成该设备的OPC UA 信息模型,并且以OPCUA nodeset2.xml 格式输出:
-
- from openai import OpenAI
-
- client = OpenAI(
- api_key="sk-xxxxxx",
- base_url="https://api.chatanywhere.tech/v1"
- )
-
- completion = client.chat.completions.create(
- model="gpt-3.5-turbo",
- messages=[
- {"role": "system", "content": "creat opc ua information model ,output modeset2 xml format"},
- {"role": "user", "content": "typename motor ,attributes include current,voltage temperature,power data type float"}
- ],
- temperature=0.3,
- )
-
- print(completion.choices[0].message.content)
- runfile('E:/yao2024/python2024/chatgptSimple.py', wdir='E:/yao2024/python2024')
- ```xml
- <?xml version="1.0" encoding="UTF-8"?>
- <opc:Model xmlns:opc="http://opcfoundation.org/UA/2011/03/UANodeSet.xsd" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
- <opc:ObjectType NodeId="ns=1;i=1001" BrowseName="MotorType">
- <opc:DisplayName>Motor</opc:DisplayName>
- <opc:References>
- <opc:Reference ReferenceType="HasProperty" IsForward="false">i=47</opc:Reference>
- <opc:Reference ReferenceType="HasProperty" IsForward="false">i=48</opc:Reference>
- <opc:Reference ReferenceType="HasProperty" IsForward="false">i=49</opc:Reference>
- <opc:Reference ReferenceType="HasProperty" IsForward="false">i=50</opc:Reference>
- </opc:References>
- </opc:ObjectType>
- <opc:Variable NodeId="ns=1;i=47" BrowseName="Current" DataType="Float">
- <opc:DisplayName>Current</opc:DisplayName>
- </opc:Variable>
- <opc:Variable NodeId="ns=1;i=48" BrowseName="Voltage" DataType="Float">
- <opc:DisplayName>Voltage</opc:DisplayName>
- </opc:Variable>
- <opc:Variable NodeId="ns=1;i=49" BrowseName="Temperature" DataType="Float">
- <opc:DisplayName>Temperature</opc:DisplayName>
- </opc:Variable>
- <opc:Variable NodeId="ns=1;i=50" BrowseName="Power" DataType="Float">
- <opc:DisplayName>Power</opc:DisplayName>
- </opc:Variable>
- </opc:Model>
酷吧?关键是提示要写好。如何生成特定行业的DSL 语言,还需要进一步研究。例如生产线的工艺流程编排。这将极大提升系统设计和维护的工作效率,降低了工程成本。
下面的例子演示如何利用chatGPT按照JSON 模板,构建结构化数据。
- import json
- import os
- from langchain_openai import ChatOpenAI
- from langchain.agents import initialize_agent, Tool
- from langchain.agents.mrkl import prompt
- os.environ['OPENAI_API_KEY'] ="sk-xxxxxxxx"
- os.environ['OPENAI_BASE_URL'] ="https://api.chatanywhere.tech/v1"
- def get_template(productClass):
- #print(productClass)
-
- answer = [
- {"type": "product type",
- "brand": "product brand",
- "manufacture":"product manufacture",
- "color":"color of prodcts",
- "size":"product size"}
-
- ]
-
- return json.dumps(answer)
- def device_control(device_id):
- print(device_id)
- status=True
- answer = [
- {"状态": status}
- ]
- return json.dumps(answer)
- def lang_chain_agent(text):
- llm = ChatOpenAI(model_name="gpt-3.5-turbo",base_url="https://api.chatanywhere.tech/v1")
-
- tools = [
- Tool(
- name = "get_template",
- func=get_template,
- description="use this tool when you need to get product model tempplate ,To use the tool, you must provide chinese product class",
- )
-
- ]
-
- agent = initialize_agent(
- tools,
- llm,
- agent="zero-shot-react-description",
- agent_kwargs=dict(suffix='Answer should be json. ' + prompt.SUFFIX),
- verbose=True,
- return_intermediate_steps=True)
-
- response = agent({"input": text})
-
- return response
- lang_chain_agent("根据如下数据生成符合模型样板的json 产品数据, 类型 足球 品牌 小少年 制造商 鹰派运动用品公司 颜色 红色 尺寸 12 英寸")
- > Entering new AgentExecutor chain...
- I need to use the get_template tool to generate the product model template for a football product.
- Action: get_template
- Action Input: 足球
- Observation: [{"type": "product type", "brand": "product brand", "manufacture": "product manufacture", "color": "color of prodcts", "size": "product size"}]
- Thought:Now I can fill in the template with the provided data.
- Final Answer: {"type": "足球", "brand": "小少年", "manufacture": "鹰派运动用品公司", "color": "红色", "size": "12 英寸"}
-
- > Finished chain.
OPC UA是自动化行业广泛应用的工业标准,我们设想可以在chatGPT Agent 中增加一个OPCUA Client ,用它来获取现场设备的状态,并且实现chatGPT对物理设备的控制。其架构如下:
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。