Call ChatGPT

Integrate system functions with ChatGPT by calling the ChatGPT interface

Parameter Configuration

Parameter	Description
API Address	Interface address for calling ChatGPT, can be dynamically calculated and specified using expressions
API Key	Interface secret key used by the API interface, can be dynamically calculated and specified using expressions
Timeout (seconds)	Timeout time for calling the ChatGPT interface, 0 means no timeout, wait for interface return
temperature	Value between 0-2, larger values make answers more random
top_p	Nucleus sampling, value between 0-1, 0.1 means only consider tokens that make up the top 10% probability mass
Maximum token count returned	Each Chinese character occupies one token, the total number of input and output tokens cannot exceed 4096
presence_penalty	Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear so far in the text, increasing the likelihood that the model will talk about new topics
frequency_penalty	Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text, reducing the likelihood that the model will repeat the same line verbatim
Stop Words	Stop generation when encountering the following words, multiple stop words are separated by commas. Maximum 4, can be dynamically calculated and specified using expressions
Number of answers generated	Stop generation when encountering the following words, multiple stop words are separated by commas. Maximum 4, can be dynamically calculated and specified using expressions
Content sent	Prompt content sent to ChatGPT, can be dynamically calculated and specified using expressions

Return Values

Property	Type	Description
id	String	API call response ID
created	Integer	Creation time
object	String	Type of API interface called
model	String	Model used
choices	Array<ChatGptResultChoice>	OpenAI response result options
usage	ChatGptResultUsage	API interface usage

Property	Type	Description
index	Integer	Index
message	ChatGptResultMessage	Message
finish_reason	String	Reason for completion

Property	Type	Description
prompt_tokens	Integer	Number of prompt tokens
completion_tokens	Integer	Number of reply tokens
total_tokens	Integer	Total number of tokens

Property	Type	Description
role	String	Role
content	String	Content

Notes

• Api Key Can be generated through https://platform.openai.com/account/api-keys

  • Api Key cannot be viewed again after generation, needs to be deleted and recreated, please keep Api Key properly
  • Do not share your API key with others, or expose it in browser or other client-side code

• role field, which are system representing the system, user representing the user, and assistant representing the AI's answer

  - When role is ``system``
  
    The content represents an instruction we give to AI, that is, telling AI how to answer the user's questions. For example, if we want AI to answer in Chinese, we can write "You are an assistant who only answers questions in Chinese" in content, so that even if the user's questions are all in English, the AI's replies will be in Chinese
  - When role is ``user`` or ``assistant``
  
    The content represents the content of the conversation between the user and AI. Need to send the historical conversation together to OpenAI's interface so that it can understand the context of the entire conversation
  

Usage Scenarios

• Call OpenAI to parse data based on context content
• Call OpenAI to summarize and extract outlines based on context content
• Call OpenAI to generate formatted data for system entry based on context content