Om.Yaml.psm1
|
function Get-SplittedValue { param ( [switch] $onlyvalue, [Parameter(Mandatory = $true)] $value, [Parameter(Mandatory = $true)] $linenumber ) $returnvalue = $null if ($onlyvalue){ $processvalue = $value.trimend() Write-Debug "Onlyvalue" } else { $valueObj = ConvertFrom-StringData -Delimiter ':' -StringData $value $key = $($valueObj.Keys[0]) $processvalue = $($valueObj.values[0]) } # Cleaning key and value, removing - symbol $key = $key -replace '^[\s]*-[\s]+','' $processvalue = $processvalue -replace '^[\s]*-[\s]+','' $processvalue = $processvalue -replace '^[\s]*','' Write-Debug "Unprocessed value: [$value]" Write-Debug "Processed value: [$processvalue]" # If no value it is a key, return hash if (!($processvalue -eq "")) { switch -regex ($processvalue) { # Processing strings # Looks for quotes in the start, denotes a text value '^\s*("(.*)"|''(.*)'')\s*$' { $returnvalue = $Matches[1] $returnvalue = $returnvalue.Substring(1,$returnvalue.Length-2) break } #Processing multiline that preserves NEWLINEs. # Find '| some text in one line possibly with \n and paragraphs \n\n in the text' keep it all '^\|[\s]*NEWLINE([\S\s]*)' { $returnvalue = $($Matches[1].replace("`n",'\n')) $returnvalue = $($returnvalue.replace('NEWLINE',"`n")) break } #Processing multiline that folds NEWLINEs. '^>[\s]*NEWLINE([\S\s]*)$' { $returnvalue = $($Matches[1].replace("`n",'\n')) $returnvalue = $($returnvalue.replace('NEWLINE',"`n")) break } #Processing False. '^[\s]*(?i)(false)[\s]*$' { $returnvalue = $false break } #Processing True. '^[\s]*(?i)(true)[\s]*$' { $returnvalue = $true break } # Processing array '^[\s]*\[([\S\s]*)\][\s]*$' { try { $returnvalue = @(,$($Matches[0]) | convertfrom-json -Depth 100 -ErrorAction stop) } catch { throw "Error on line $linenumber in yaml. $_" } break } # Processing hash/json '^[\s]*\{([\S\s]*)\}[\s]*$' { try { $returnvalue = $($Matches[0] | convertfrom-json -Depth 100 -ErrorAction stop) } catch { throw "Error on line $linenumber in yaml. $_" } break } default{ # Processing int's try { $returnvalue = [int]$($processvalue) } # Processing strings catch { $returnvalue = $processvalue } } } } else { $returnvalue = $processvalue } if ($onlyvalue){ return $returnvalue } else { $valueObj = @{$key = $returnvalue} Write-Debug "Retruning key $($valueObj.keys[0]) and value $($valueObj.values[0])" return $valueObj } } function New-HelpIndex { param ( [Parameter(Mandatory = $true)][System.Collections.ArrayList]$data ) #$stringPattern = '^(((?!:(\s|$)).)+|\s*(\-\s+)?("([^"]*)"|''([^'']*)''))$' $valuePattern = '^(((?!:(\s|$)).)+|\s*(\-\s+)?("([^"]*)"|''([^'']*)''|{(.*)}|\[(.*)\]))$' $ignoredPattern = '(^[\s]*#.*$)|(^\s*$)' $containsCommentPattern = '(^.*\s#.*$)' $arrayPattern= '^\s*-\s+.*' $notMultilinePattern = '^((((?!:\s(\||>)).)+|\s*(\-\s+)?("([^"]*)"|''([^'']*)''))|\s*)$' $onlySpacesPattern = '^\s*$' $nonIndentedWithValuePattern = '^[^#\s][\s\S]*$' $keyPattern = '^[^:]+?:(\s+("([^"]*)"|''([^'']*)''|(?!:.*:\s.*)|[^"'']*\S+[^"'']*)\s*$|$)' [System.Collections.ArrayList]$helpIndex = @() # Looping through data and creating help index $foundNonIndentedLine = $false for($i =0; $i -lt $data.count; $i++ ) { $whitespace = $null $vartype = $null if($data[$i] -match $nonIndentedWithValuePattern) { $foundNonIndentedLine = $true } # Setting vartype to ignore if starting with # or empty line (will be removed from the list). if($data[$i] -match $ignoredPattern) { Write-Debug "On line $i in ignore match" $vartype = "ignore" } # Setting the whitespace count for lines that will be kept $whitespace = $($($data[$i]).Length - $($($data[$i]) -replace "^\s*","").Length) $indent = $($($data[$i]).Length - $($($data[$i]) -replace "^\s*(-\s+)?","").Length) # Creating and adding the item to help index list. Setting the original line number (Needed because some lines will be removed). $indexItem = [PSCustomObject]@{ Indent = $indent whitespace = $whitespace vartype = $vartype lineNumber = $i } [void]$helpIndex.add($indexItem) } if( -not $foundNonIndentedLine){ throw "Did not find any non indented lines. Yaml is not valid" } # Finding multilines. Empty multilines should not be removed for($i =0; $i -lt $data.count; $i++ ) { # If multiline: if (-not ($data[$i] -match $notMultilinePattern) ) { $helpIndex[$i].vartype = "dictionary" $firstMline = $i+1 # Getting all the multilines for ($mline = $firstMline; $mline -lt $data.count; $mline++) { if ( ($helpIndex[$mline].whitespace -ge $helpIndex[$firstMline].whitespace) -or (($data[$mline] -match $onlySpacesPattern) -and $helpIndex[$mline+1].whitespace -ge $helpIndex[$firstMline].whitespace) ) { $helpIndex[$mline].vartype = "multiline" $helpIndex[$mline].indent = $helpIndex[$firstMline].indent } else { # Not part of multiline. Setting $i to $mline -1 to skip multilines in outer loop (need to check this again) $i = $mline -1 break } } } } # Removing lines that should be ignored for($i = $data.count -1; $i -ge 0; $i-- ) { if ($helpIndex[$i].vartype -eq "ignore") { [void]$data.RemoveAt($i) [void]$helpIndex.RemoveAt($i) } } # Setting the vartype on every item in the list for($i =0; $i -lt $data.count; $i++ ) { # If dictionary has multiple lines if ($null -eq $helpIndex[$i].vartype) { if ( -not ($data[$i] -match $valuePattern) ) { # Checking if array or dictionary if ($data[$i+1] -match $arrayPattern ) { # Indent if ($helpIndex[$i].indent -lt $helpIndex[$i+1].indent) { $helpIndex[$i].vartype = "array" } else { $helpIndex[$i].vartype = "dictionary" } } else { $helpIndex[$i].vartype = "dictionary" } } elseif ($data[$i] -match $valuePattern ) { $helpIndex[$i].vartype = "value" } } if ($helpIndex[$i].vartype -ne "multiline" ) { # If not multiline, looking for '#' in the middle of the line if ($allowInlineComments) { Write-Debug "On line $i in inline comment. Substring after '#' will be removed" $data[$i] = Remove-InlineComment -inputLine $data[$i] } else { $returnedString = $null $returnedString = Remove-InlineComment -inputLine $data[$i] if ($returnedString -ne $data[$i]) { Write-Debug "$returnedString -ne $($data[$i])" Throw "Error on line $i. Found inline comment. This is not allowed, but can be enabled by adding the allowInlineComments (in beta)" } } } } return [PSCustomObject]@{ inputDataDirty = $data helpIndexDirty = $helpIndex } } function Get-CleanData { param ( [Parameter(Mandatory = $true)][System.Collections.ArrayList]$helpIndex, [Parameter(Mandatory = $true)][System.Collections.ArrayList]$data ) $mValue = $null $onlySpacesPattern = '^\s*$' $stringPattern = '^(((?!:(\s|$)).)+|\s*(\-\s+)?("([^"]*)"|''([^'']*)''))$' $inArrayPattern = '^\s*-\s+.*$' #$keyPattern = '^[^:]+?:(\s+("([^"]*)"|''([^'']*)''|(?!:.*:\s.*)|[^"'']*\S+[^"'']*)\s*$|$)' $keyOnlyPattern = '^[^:]+?:\s*$' for($i = $data.count -1; $i -ge 0; $i-- ) { if ($helpIndex[$i].vartype -eq "multiline") { if ($helpIndex[$i-1].vartype -eq "multiline") { if ($helpIndex[$i-1].whitespace -ge $helpIndex[$i-1].indent) { $currentMvalue = $($data[$i-1].substring($helpIndex[$i-1].indent)) } else { $currentMvalue = $($data[$i-1].substring($helpIndex[$i-1].whitespace)) } } else { $currentMvalue = $($data[$i-1]) } if ($mValue) { $mValue = "$currentMvalue" + 'NEWLINE' + "$mValue" } else { $mValue = "$currentMvalue" + 'NEWLINE' + "$($data[$i].substring($helpIndex[$i].indent))" } [void]$data.RemoveAt($i) [void]$helpIndex.RemoveAt($i) } else { if($mValue) { $data[$i] = $mValue + 'NEWLINE' $mValue = $null } elseif ($sValue) { $data[$i] = $sValue $sValue = $null } } } # Loop and check for errors in yaml (syntax error) for($i =0; $i -lt $data.count; $i++ ) { #elseif ($data[$i] -Match "^[\s]*[\S]+.*:[\s]+[\S]+") { if (-not ($data[$i] -Match $keyOnlyPattern)) { # If key contains value on line, do not allow value on next line ( no indent) if ($($helpIndex[$i+1].Indent) -gt $($helpIndex[$i].Indent) ) { throw "Error on line $($helpIndex[$i].lineNumber), $($inputDataDirty[$($helpIndex[$i].lineNumber)]). While parsing a block mapping, found multipe types" } } elseif ( $data[$i] -Match $stringPattern -and (-not ($data[$i] -Match $inArrayPattern) ) ) { # If string/value is no in an array, throw error (we do not support it) throw "Error on line $($helpIndex[$i].lineNumber), $($inputDataDirty[$($helpIndex[$i].lineNumber)]). Multiline string is only supported by using | or >" } else { $blockLevel =$i+1 for($x = $i+2; $x -lt $data.count; $x++ ) { if ($helpIndex[$x].Indent -le $helpIndex[$i].Indent) { break } if ( ($helpIndex[$x].Indent -gt $helpIndex[$i].Indent) -and ($helpIndex[$x].Indent -lt $($helpIndex[$blockLevel].Indent)) ) { throw "Error on line $($helpIndex[$x].lineNumber), $($inputDataDirty[$($helpIndex[$x].lineNumber)]). While sparsing a block mapping, indent error " } } } } return [PSCustomObject]@{ inputDataClean = $data helpIndexClean = $helpIndex } } function Get-ParsedTree { param ( [Parameter(Mandatory = $true)]$helpIndex, [Parameter(Mandatory = $true)]$inputData, [Parameter(Mandatory = $true)]$startRow, [Parameter(Mandatory = $true)]$endRow ) $hash = [ordered]@{} for($i = $startRow; $i -le $endRow; $i++ ) { Write-Debug "On top. i is $i" # Checking if current row contains an array or not (block of arrays) if ($($helpIndex[$i]).vartype -eq "array") { [System.Collections.ArrayList] $array = @() $hashName = Get-SplittedValue -value $($inputData[$i]) -lineNumber $helpIndex[$i].lineNumber # Should find block to parse. Looping on every array elemen # Should call Get-ParsedTree to parse underlying block. Array should be retrived which will be added to the the key on cunnect line $i # Finding array space/block [array]$arrayList = $null [array]$arrayList += for ($z = $i+1; $z -le $endRow; $z++) { if ($($helpIndex[$z]).whitespace -le $($helpIndex[$i+1].whitespace)) { $z if ($($helpIndex[$z]).indent -lt $($helpIndex[$i+1].indent)) { #End of Arrayblock break } } if ($z -eq $endRow) { # Need extra stop row to measure distance (find stoprow) $($z+1) } } # Lopping over the array elements (parsing every element using get-parsedTree) for ($a = 0; $a -lt $arrayList.Length -1; $a++) { $stopRow = $($arrayList[$a+1]-1) Write-Debug "From Get-ParsedTree array element/block ----------------->>>> StartRow:$($arrayList[$a]) EndRow: $stopRow ----------------------------------------------------------------" $returnedData = Get-ParsedTree -inputData $inputDataClean -helpIndex $helpIndexClean -startRow $($arrayList[$a]) -endRow $stopRow Write-Debug "From Get-ParsedTree array element/block -----------------<<<< StartRow:$($arrayList[$a]) EndRow: $stopRow ----------------------------------------------------------------" [void]$array.Add($returnedData) } # Finished retrieving all array members. Retruning the array Write-Debug "Finished retrieving all array members. Retruning the array. Array looks like this: : $(@{$($hashName.keys[0]) = $array} | convertto-json -Depth 100)" $hash.add($($hashName.keys[0]), $array.ToArray()) Write-Debug "Setting i to $stopRow. arrayblock is already processed" $i = $stopRow } elseif ($($helpIndex[$i]).vartype -eq "dictionary") { # Should find block to parse # Should call Get-ParedTree to parse underlying block. Dictionary should be retived witch wil be added to the key $hashData = Get-SplittedValue -value $($inputData[$i]) -lineNumber $helpIndex[$i].lineNumber # Finding dictionary space/block $count = $null for ($n = $i+1; $n -le $endRow; $n++) { if ($($helpIndex[$n]).indent -gt $($helpIndex[$i].indent)) { $count++ } else { break } } if ($count -gt 0) { Write-Debug "Hash key name containing the dictionary is $($hashData.keys[0]). Sub block count is $count" Write-Debug "From Get-ParsedTree dictionary element/block----------------->>>> StartRow:$($i+1) EndRow: $($i + $count ) ----------------------------------------------------------------" $returnedData = Get-ParsedTree -inputData $inputDataClean -helpIndex $helpIndexClean -startRow $($i+1) -endRow $($i + $count ) Write-Debug "From Get-ParsedTree dictionary element/block-----------------<<<< StartRow:$($i+1) EndRow: $($i + $count ) ----------------------------------------------------------------" # Finished retrieving the hashtable members. Retruning the hashtable Write-Debug "Finished retrieving all hashtable members. Retruning the hashtable. Hashtable looks like this: : $(@{$($hashData.keys[0]) = $returnedData} | convertto-json -depth 100)" $hash.add("$($hashData.keys[0])", $returnedData) } else { write-Debug "No need to parse. No sub block. Hashtable looks like this: : $($hashData | convertto-json -depth 100)" $hash.add("$($hashData.keys[0])", $($hashData.values[0])) } Write-Debug "Setting i to $($i + $count ). Dictionary block is already processed" $i = $i + $count #Get-ParsedTree -inputData $inputDataClean -helpIndex $helpIndexClean -startRow 0 -endRow $($inputDataClean.Length -1) } elseif ($($helpIndex[$i]).vartype -eq "value") { # Should call split to get value. How will this be added as value to the dictionary? This cunnrent session should output only this value. Should only happen if startrow -eq endrow. # If multipe lines, should probably add all value rows ($value+$value) $value = Get-SplittedValue -value $($inputData[$i]) -lineNumber $helpIndex[$i].lineNumber -onlyvalue $value } else { Throw "Error, unknow vartype: $($($helpIndex[$i]).vartype)" } } if ($hash.Count -ne 0) { $hash } } function Remove-InlineComment { param ( [Parameter(Mandatory = $true)]$inputLine ) # Creating an index the quoted text (hashtags found in this space should be ignored) $quoted = Select-String -InputObject $inputLine -Pattern "(""[^""]*""|'[^']*')" -AllMatches $inQuotesIndex = for ($i = 0; $i -lt $quoted.Matches.Count; $i++) { $firstOrSpace = $false $lastOrSpace = $false $startIndex = $($quoted.Matches[$i].Index) $endIndex = $startIndex+$quoted.Matches[$i].Length-1 # if the quoted string is first or space before quote if ($startIndex -eq 0) { $firstOrSpace = $true } elseif ($inputLine[$startIndex-1] -eq " ") { $firstOrSpace = $true } # if the quoted string is last or space after quote if ($endIndex -eq $($inputLine.Length -1)) { $lastOrSpace = $true } elseif ($inputLine[$endIndex+1] -eq " ") { $lastOrSpace = $true } if ($firstOrSpace -and $lastOrSpace) { for ($z = $startIndex; $z -le ($endIndex); $z++) { $z } } } # Checking for the first unquoted hashtag (if any). Must start with space for ($i = 0; $i -lt $inputLine.Length; $i++) { if (($inputLine[$i] -eq "#") -and ($inputLine[$i-1] -eq " ")) { if ($i -notin $inQuotesIndex) { return $inputLine.SubString(0,$i) } } } # If comment not found. retrun string $inputLine } function ConvertFrom-OmYaml { [CmdletBinding()] param ( [Parameter(Mandatory = $true)]$Path, [switch]$allowInlineComments ) [array]$content = Get-Content -Path $path # Adding data from Yaml to a hash with key "processedData". This way we know that the Get-ParsedTree will start with a dictionary and don't need logic to check if array. for ($i = 0; $i -lt $content.Count; $i++) { $content[$i] = " $($content[$i])" } $content = @("processedData:") + $content $dataDirty = New-HelpIndex -data $content $dataClean = Get-CleanData -data $dataDirty.inputDataDirty -helpIndex $dataDirty.helpIndexDirty [array]$inputDataClean = $dataClean.inputDataClean [array]$helpIndexClean = $dataClean.helpIndexClean $tree = Get-ParsedTree -inputData $inputDataClean -helpIndex $helpIndexClean -startRow 0 -endRow $($inputDataClean.Length -1) $tree.processedData } |