Class OpenAiChat

Namespace

BotManager.OpenAi.Expressions

Assembly

BotManager.OpenAi.dll

Executes a char request to the OpenAI api. The OpenAiInit must be initialized.

Return type is string.

This json example sends an OpenAI chat request and returns the result:

{ "$OpenAiChat": { "Messages": [
  { "Role": "System", "Text": "Create a recipe suggestion for the following ingredient list:" },
  { "Role": "User", "Text": "Cheesecake" }
] } }

public sealed class OpenAiChat : IExpression

Inheritance

object

OpenAiChat

Implements

IExpression

Inherited Members

object.GetType()

object.ToString()

object.Equals(object)

object.Equals(object, object)

object.ReferenceEquals(object, object)

object.GetHashCode()

Constructors

OpenAiChat()

public OpenAiChat()

Properties

Messages

The chat messages to respond to.

You should provide your ruleset as a separate OpenAI.GPT3.ObjectModels.StaticValues.ChatMessageRoles.System message and not concat it with user input. Use a message with OpenAI.GPT3.ObjectModels.StaticValues.ChatMessageRoles.User role for user input. This prioritizes your ruleset over the user input and protects against prompt injection.

public OpenAiChatMessage[]? Messages { get; set; }

Property Value

OpenAiChatMessage[]

Methods

ExecuteAsync(RuntimeContext, Type?)

Executes the expression in the given RuntimeContext. Please do not use this method to run an expression. Use ExecuteAsync(IExpression?) instead.

public Task<object?> ExecuteAsync(RuntimeContext context, Type? returnType)

Parameters

context RuntimeContext

The current runtime context.

returnType Type

The expected return type. Use typeof(object) for a generic result. Use null if you don't need the return value.

Returns

Task<object>

Returns the result of the expression.