Functions/GenXdev.AI.Queries/Invoke-LLMBooleanEvaluation.ps1
|
<############################################################################## Part of PowerShell module : GenXdev.AI.Queries Original cmdlet filename : Invoke-LLMBooleanEvaluation.ps1 Original author : René Vaessen / GenXdev Version : 3.26.2026 ################################################################################ Copyright (c) 2026 René Vaessen / GenXdev Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. ################################################################################> ############################################################################### <# .SYNOPSIS Evaluates a statement using AI to determine if it's true or false. .DESCRIPTION This function uses AI models to evaluate statements and determine their truth value. It can accept input directly through parameters, from the pipeline, or from the system clipboard. The function returns a boolean result along with confidence level and reasoning from the AI model. .PARAMETER Text The statement to evaluate. If not provided, the function will read from the system clipboard. .PARAMETER Instructions Instructions to guide the AI model in evaluating the statement. By default, it will determine if the statement is true or false. .PARAMETER Attachments Array of file paths to attach to the AI query for additional context. .PARAMETER SetClipboard When specified, copies the result back to the system clipboard after processing. .PARAMETER Temperature Temperature for response randomness (0.0-1.0). Controls creativity vs determinism in AI responses. .PARAMETER ImageDetail Image detail level for visual processing. Valid values are low, medium, or high. .PARAMETER IncludeThoughts Include model's thoughts in output for debugging and transparency. .PARAMETER DontAddThoughtsToHistory Prevent model thoughts from being added to conversation history. .PARAMETER ContinueLast Continue from last conversation instead of starting fresh. .PARAMETER Functions Array of function definitions for AI tool use capabilities. .PARAMETER ExposedCmdLets Array of PowerShell command definitions to use as tools in AI operations. .PARAMETER NoConfirmationToolFunctionNames Array of command names that don't require confirmation before execution. .PARAMETER Speak Enable text-to-speech for AI responses. .PARAMETER SpeakThoughts Enable text-to-speech for AI thought responses. .PARAMETER NoSessionCaching Don't store session in session cache for privacy or debugging. .PARAMETER AllowDefaultTools Allow the AI to use default tools and capabilities. .PARAMETER LLMQueryType The type of LLM query to optimize for specific use cases. .PARAMETER Model The model identifier or pattern to use for AI operations. .PARAMETER ApiEndpoint The API endpoint URL for AI operations. .PARAMETER ApiKey The API key for authenticated AI operations. .PARAMETER SessionOnly Use alternative settings stored in session for AI preferences. .PARAMETER ClearSession Clear alternative settings stored in session for AI preferences. .PARAMETER PreferencesDatabasePath Database path for preference data files. .PARAMETER SkipSession Store settings only in persistent preferences without affecting session. .EXAMPLE Invoke-LLMBooleanEvaluation -Text "The Earth is flat" -Model "gpt-4" .EXAMPLE "Humans need oxygen to survive" | Invoke-LLMBooleanEvaluation .EXAMPLE equalstrue "2 + 2 = 4" ###############################################################################> function Invoke-LLMBooleanEvaluation { [CmdletBinding()] [OutputType([System.Boolean])] [Alias('equalstrue')] param ( ############################################################################### [Parameter( Position = 0, Mandatory = $false, ValueFromPipeline = $true, HelpMessage = 'The statement to evaluate' )] [string] $Text, ############################################################################### [Parameter( Position = 1, Mandatory = $false, HelpMessage = ('Instructions for the AI model on how to evaluate ' + 'the statement') )] [string] $Instructions = '', ############################################################################### [Parameter( Position = 2, Mandatory = $false, HelpMessage = 'Array of file paths to attach' )] [string[]] $Attachments = @(), ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Temperature for response randomness (0.0-1.0)' )] [ValidateRange(-1, 1.0)] [double] $Temperature = -1, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Image detail level' )] [ValidateSet('low', 'medium', 'high')] [string] $ImageDetail = 'low', ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Array of function definitions' )] [hashtable[]] $Functions = @(), ############################################################################### [Parameter( Mandatory = $false, HelpMessage = ('Array of PowerShell command definitions to use ' + 'as tools') )] [GenXdev.Helpers.ExposedCmdletDefinition[]] $ExposedCmdLets = @(), ############################################################################### [Parameter( Mandatory = $false, HelpMessage = ("Array of command names that don't require " + 'confirmation') )] [Alias('NoConfirmationFor')] [string[]] $NoConfirmationToolFunctionNames = @(), ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'The type of LLM query' )] [ValidateSet( 'SimpleIntelligence', 'Knowledge', 'Pictures', 'TextTranslation', 'Coding', 'ToolUse' )] [string] $LLMQueryType = 'Knowledge', ############################################################################### [Parameter( Mandatory = $false, HelpMessage = ('The model identifier or pattern to use for AI ' + 'operations') )] [string] $Model, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'The API endpoint URL for AI operations' )] [string] $ApiEndpoint, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'The API key for authenticated AI operations' )] [string] $ApiKey, ################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Indicates that LLM has no support for JSON schemas' )] [string] $NoSupportForJsonSchema, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Database path for preference data files' )] [Alias('DatabasePath')] [string] $PreferencesDatabasePath, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Copy the result to clipboard' )] [switch] $SetClipboard, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = "Include model's thoughts in output" )] [switch] $IncludeThoughts, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = "Don't add model thoughts to conversation history" )] [switch] $DontAddThoughtsToHistory, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Continue from last conversation' )] [switch] $ContinueLast, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Enable text-to-speech for AI responses' )] [switch] $Speak, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Enable text-to-speech for AI thought responses' )] [switch] $SpeakThoughts, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = "Don't store session in session cache" )] [switch] $NoSessionCaching, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Allow AI to use default tools and capabilities' )] [switch] $AllowDefaultTools, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = ('Use alternative settings stored in session for ' + 'AI preferences') )] [switch] $SessionOnly, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = ('Clear alternative settings stored in session ' + 'for AI preferences') )] [switch] $ClearSession, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = ('Store settings only in persistent preferences ' + 'without affecting session') )] [Alias('FromPreferences')] [switch] $SkipSession, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Return only markup blocks in the output' )] [switch] $OutputMarkdownBlocksOnly, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Filter for specific types of markup blocks' )] [string[]] $MarkupBlocksTypeFilter, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Temperature for audio response randomness' )] [double] $AudioTemperature, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Temperature for response generation' )] [double] $TemperatureResponse, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Language code or name for the response' )] [string] $Language, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Number of CPU threads to use for processing' )] [int] $CpuThreads, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Regular expression to suppress from output' )] [string] $SuppressRegex, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Audio context size for processing' )] [int] $AudioContextSize, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Silence threshold for audio detection' )] [double] $SilenceThreshold, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Length penalty for sequence generation' )] [double] $LengthPenalty, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Entropy threshold for output filtering' )] [double] $EntropyThreshold, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Log probability threshold for output filtering' )] [double] $LogProbThreshold, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'No speech threshold for audio detection' )] [double] $NoSpeechThreshold, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Disable speech output' )] [switch] $DontSpeak, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Disable speech output for thoughts' )] [switch] $DontSpeakThoughts, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Disable VOX (voice activation)' )] [switch] $NoVOX, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Use desktop audio capture for input' )] [switch] $UseDesktopAudioCapture, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Do not use context in the query' )] [switch] $NoContext, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Use beam search sampling strategy' )] [switch] $WithBeamSearchSamplingStrategy, ############################################################################### [Parameter( Mandatory = $false, HelpMessage = 'Return only responses from the model' )] [switch] $OnlyResponses, ############################################################################### [int] $MaxToolcallBackLength ############################################################################### ) begin { # prepend evaluation instructions to user instructions $Instructions = @" Evaluate if the following statement is true or false. Respond with a JSON object containing 'result' (boolean), 'confidence' (0.0 - 1.0) and 'reason' (string). The returned 'reason' should explain as short as possible, why the statement was evaluated as true or false. Only pure facts should have high confidence. $Instructions "@ # define response format schema for boolean evaluation $responseSchema = @{ type = 'json_schema' json_schema = @{ name = 'boolean_evaluation_response' strict = 'true' schema = @{ type = 'object' properties = @{ result = @{ type = 'boolean' description = ('The evaluation result: true if ' + "the statement is true, false if it's false") } confidence = @{ type = 'number' minimum = 0 maximum = 1 description = ('Confidence level in the ' + 'evaluation (0-1)') } reason = @{ type = 'string' description = ('Explanation for why the statement ' + 'was evaluated as true or false') } } required = @('result', 'confidence', 'reason') } } } | Microsoft.PowerShell.Utility\ConvertTo-Json -Depth 10 # log initialization information Microsoft.PowerShell.Utility\Write-Verbose ( "Starting boolean evaluation with model: $Model" ) # initialize result variables $script:result = $false $response = $null } process { # check if we should read from clipboard $isClipboardSource = [string]::IsNullOrWhiteSpace($Text) if ($isClipboardSource) { # log clipboard reading operation Microsoft.PowerShell.Utility\Write-Verbose ( 'No direct text input, reading from clipboard' ) # get text from system clipboard $Text = Microsoft.PowerShell.Management\Get-Clipboard if ([string]::IsNullOrWhiteSpace($Text)) { Microsoft.PowerShell.Utility\Write-Verbose ( 'No text found in the clipboard.' ) return } } try { # log processing start Microsoft.PowerShell.Utility\Write-Verbose ( 'Processing statement for evaluation' ) # copy parameters from current function to invoke-llmquery $invocationParams = GenXdev\Copy-IdenticalParamValues ` -BoundParameters $PSBoundParameters ` -FunctionName 'GenXdev\Invoke-LLMQuery' # configure query parameters for boolean evaluation $invocationParams.Query = $Text $invocationParams.Instructions = $Instructions $invocationParams.IncludeThoughts = $false $invocationParams.ResponseFormat = $responseSchema $invocationParams.Temperature = $Temperature if ($AllowDefaultTools) { # enable text prompt mode for tool usage $invocationParams.ChatMode = 'textprompt' $invocationParams.ChatOnce = $true } # get evaluation result from ai model $response = GenXdev\Invoke-LLMQuery @invocationParams | Microsoft.PowerShell.Utility\ConvertFrom-Json # extract boolean result from response $script:result = [bool]$response.result # create summary for verbose output $summary = ("`r`n`"$Text`"`r`n`r`nevaluates to be: " + "$($response.result)`r`nconfidence: $($response.confidence) " + "confidence.`r`nReason: $($response.reason)`r`n") # log evaluation summary Microsoft.PowerShell.Utility\Write-Verbose $summary } catch { Microsoft.PowerShell.Utility\Write-Error ( "Failed to evaluate statement with AI model: $_" ) } } end { if ($null -ne $response) { if ($SetClipboard) { # log clipboard operation Microsoft.PowerShell.Utility\Write-Verbose ( 'Copying result to clipboard' ) if ($IncludeThoughts) { # create detailed summary with thoughts $summary = ("`r`n`"$Text`"`r`n`r`nevaluates to be: " + "$($response.result)`r`nconfidence: " + "$($response.confidence) confidence.`r`nReason: " + "$($response.reason)`r`n") # copy summary to clipboard instead of piping $null = Microsoft.PowerShell.Management\Set-Clipboard ` -Value $summary } else { # copy boolean result to clipboard $null = Microsoft.PowerShell.Management\Set-Clipboard ` -Value $script:result } } } # return the boolean result $script:result } } ############################################################################### |