Class: ContextChatEngine
ContextChatEngine uses the Index to get the appropriate context for each query. The context is stored in the system prompt, and the chat history is preserved, ideally allowing the appropriate context to be surfaced for each query.
Implements
Constructors
constructor
• new ContextChatEngine(init)
Parameters
| Name | Type |
|---|---|
init | Object |
init.chatHistory? | ChatMessage[] |
init.chatModel? | OpenAI |
init.contextSystemPrompt? | (__namedParameters: Object) => string |
init.retriever | BaseRetriever |
Defined in
Properties
chatHistory
• chatHistory: ChatMessage[]
Defined in
chatModel
• chatModel: OpenAI
Defined in
contextSystemPrompt
• contextSystemPrompt: (__namedParameters: Object) => string
Type declaration
▸ («destructured»): string
Parameters
| Name | Type |
|---|---|
«destructured» | Object |
Returns
string
Defined in
retriever
• retriever: BaseRetriever
Defined in
Methods
chat
▸ chat(message, chatHistory?): Promise<Response>
Send message along with the class's current chat history to the LLM.
Parameters
| Name | Type | Description |
|---|---|---|
message | string | |
chatHistory? | ChatMessage[] | optional chat history if you want to customize the chat history |
Returns
Promise<Response>
Implementation of
Defined in
reset
▸ reset(): void
Resets the chat history so that it's empty.
Returns
void