Creating unique identifiers for checks to enhance post-processing

[An-dir]

When using the generated JSON report for advanced post-processing, it is currently not possible to accurately identify and therefore track some of the checks over time. My suggestion is to assign a unique identifier to each individual test result. Based on your current default checks, they could look like this:

Short ID (generated from current default tests)

• CIS.1.2.1
• EIDSCA.AF01
• MS.AAD.1.1
• MS.SHAREPOINT.1.1
• MT.1004
• MT.1024.9e53e74a-30a5-4cae-932a-f4d7b81ab3cd_passwordHashSync

I have created a workaround that meets my needs for now, but it could break when updating Maester. Therefore I am interested in the extension of the default properties that are returned in the json report by the property "ID" or "MaesterID".

Just to share my workarround here a part of my script that i am currently testing with (not yet finished):

$MaesterVersion = "1.0.0"
$FileForJSONExport = ".\maester\psmodule\Maester\$MaesterVersion\internal\ConvertTo-MtMaesterResults.ps1"
$ExportToolRawData = Get-Content $FileForJSONExport
# DOING MODIFICATIONS TO GET A UNIQUE ID FOR TESTS
$ExportToolRawData = foreach ($line in $ExportToolRawData) {
    if ($line -match "^\s*\`$helpUrl = ''") {
        $line
        "        # Change01`r`n"+
        "        if (`$Name -like `"CIS *`") {`r`n"+
        # CIS Level may not be included as it can't be different within a chapter
        #"        `$TestID = `$Name -replace `"^CIS *([0-9.]*) \((.*?)\).*`",'CIS.`$1.`$2'`r`n"+
        "        `$TestID = `$Name -replace `"^CIS *([0-9.]*) \((.*?)\).*`",'CIS.`$1'`r`n"+
        "        }`r`n"+
        "        if (`$Name -notlike `"CIS *`") {`r`n"+
        "        `$TestID = `$Name -replace `"^(.*?)(:| ).*`$`",'`$1'`r`n"+
        "        }`r`n"+
        "        if (`$null -ne `$test.Block.Data.ID) {`r`n"+
        "        `$TestID += `".`" + `$test.Block.Data.ID`r`n"+
        "        }`r`n"+
        "        # Change01 END`r`n"
        } elseif ($line -match "^\s*helpUrl\s*=\s*\`$helpUrl") {
        $line
        "            #Change02 NewLine:";"$(" "*12)ID$(" "*14)= `$TestID"
        } else {
        $line
    }
}
$ExportToolRawData | Out-File -Encoding utf8BOM -Path $FileForJSONExport

I avoided to directly modify the file to be able to apply updates and eventually have my script updating the new "ConvertTo-MtMaesterResults.ps1" file again.

My workarround works well except with MT.1038 as there is a test with the same number in the "Name" in the default scripts. Of course I could validate that duplicate by checking the "ScriptBlockFile" property in the result, but using this as a secondary identifyer manually is relly not that great. As i have seen this is already addressed here: #604 and here: #583 (comment)

Making an offical requirement for checks to have "ID" as a required parameter and make it unique could be challenging. Maybe a combination of the scriptfoldername, test id and an ID generated from the query if required could work but should be coordinated. Here are some examples about that idea:

Long but duplicate preventing generated ID:

• "Maester/Entra/MT.1038" (Default: using the tests folder name adding the ID as it is in "Name")
• "Maester/Entra/MT.1033.60a8a48a-60cd-4069-b84c-99aaabbbdfa0" (Like default example above adding ID from API result)
• "Maester/Entra/MT.1024.9e53e74a-30a5-4cae-932a-f4d7b81ab3cd_passwordHashSync" (Default + ID from API result)
• "cis/CIS.1.1.3" (Default, but duplicate "CIS" because the test ist NOT just "CIS 1.1.3" but more "CIS Microsoft 365 1.1.3". A corrected ID would be "CIS/Microsoft.365.1.1.3" or something like that)
• "EIDSCA/EIDSCA.AF01" (Default but duplicate "EIDSCA" in the name like in "CIS")

With my ID workarround in the code i could almost finish the unique ID generation like that:
$CSAJsonContent.Tests | foreach {($_.ScriptBlockFile -replace [regex]::Escape($repl + "maester-tests\") -replace "test-.*\.ps1" -replace"\\",".") +$_.id}

I'd prefer the shorter version of the ID and I'd prefer not to generate it for myself. So (when) will there be a true "ID"?

And for the people asking "why not use the Name property as a Identifier?" I argue with the challanges of different languages, and sometimes correction of misspelled names.

P.S.: How about an additional parameter for the tables (aka "impactedResources") that are integrated in the "ResultDetail"? In Large reports this is a requirement to have it performant and readable in html and separated in json for postprocessing.