public/Invoke-TuneChat.ps1
|
function Invoke-TuneChat { <# .SYNOPSIS Invokes a chat with OpenAI using the specified model and messages. .DESCRIPTION The Invoke-TuneChat function sends a series of messages to the OpenAI API using the specified model. The messages are processed in the order they are received, and the response from the API is returned. .PARAMETER Message An array of messages to be sent to the OpenAI API. Each message should be a hashtable with 'role' and 'content' keys. .PARAMETER Model The model to be used for the chat. Valid options are 'gpt-3.5-turbo-0613', 'babbage-002', and 'davinci-002'. .EXAMPLE Invoke-TuneChat -Message @(@{role='system'; content='You are a helpful assistant.'}, @{role='user'; content='Who won the world series in 2020?'}) -Model 'gpt-3.5-turbo-0613' This example sends two messages to the OpenAI API using the 'gpt-3.5-turbo-0613' model and returns the response. #> [CmdletBinding()] param ( [Parameter(ValueFromPipeline)] [string[]]$Message = @("how's Potato?"), [string]$Model = $script:currrentmodel ) begin { if (-not $Model) { $Model = $script:currrentmodel = (Get-TuneJob | Select-Object -First 1).fine_tuned_model } else { $Model = "gpt-3.5-turbo-0613" } $jsonmsg = @() foreach ($msg in $Message) { $jsonmsg += @{ role = "user" content = $msg } } } process { $body = @{ model = $Model messages = $jsonmsg } | ConvertTo-Json # Create a hashtable containing the parameters $params = @{ Uri = "https://api.openai.com/v1/chat/completions" Method = "POST" Body = $body } Write-Verbose "Asking $Message" (Invoke-OpenAIAPI @params).choices.message.content } } |