Powershell: Text anhand seiner Position aus PDF-Dokumenten auslesen

function Get-PDFTextFromLocation {
    <#
    .Synopsis
       Extracts text from PDF documents
    .DESCRIPTION
       Extracts text from defined locations in PDF documents
    .LINK
       © @colinardo (https://administrator.de/tutorial/powershell-text-anhand-seiner-position-aus-pdf-dokumenten-auslesen-32779450015.html)
    .INPUTS
         System.IO.FileSystemInfo[] / You can pipe the output of Get-ChildItem to the function
    .OUTPUTS
         System.Management.Automation.PSCustomObject[]
    .EXAMPLE
        $locations = @{
            1 = [ordered]@{
                'Invoice No.' = 30,123,203,172    
                'Date' = 130,230,150,245    
            }
        }
        Get-ChildItem "E:\data" -File -Filter *.pdf | Get-PDFTextFromLocation -locations $locations  

       This example extracts 2 fields on page 1 from every PDF coming on the pipeline
    .NOTES
       Hashtable-Format for text location definition by page number:
        - first level of hashtable defines the page index from which to extract the field data
        - second level keys define a user defined unique names for the fields
        - the value in the second level defines the rectangle location as an  array with four items:
            - [LowerLeftX],[LowerLeftY],[UpperRightX],[UpperRightY])
            - coordinate origin is the non rotated lower left corner of the page.
        
        Example 1 (extracts 2 fields from page 1):
        ============
        @{
            1 = [ordered]@{
                'Invoice No.' = 30,123,203,172    
                'Date' = 130,230,150,245    
            }
        }

        Example 2 (extracts 2 fields from page 1, and 1 field from page 2):
        ============
        @{
            1 = [ordered]@{
                'Invoice No.' = 30,123,203,172    
                'Date' = 130,230,150,245    
            }
            2 = [ordered]@{
                'Sum' = 120,80,150,100    
            }
        }
        # ---------------------------------------------
        Hashtable-Format to select pages by regular expression (when the document contains multiple pages of the same format):
        - first level of hashtable defines a regular expression pattern wich selects the pages to be processes
        - second level keys define a user defined unique names for the fields
        - the value in the second level defines the rectangle location as an  array with four items:
            - [LowerLeftX],[LowerLeftY],[UpperRightX],[UpperRightY])
            - coordinate origin is the non rotated lower left corner of the page.
        
        Example 3 (extracts 2 fields from each page wich matches the regular expression pattern '(?ism)Invoice'):  
        ============
        @{
            '(?ism)Invoice' = [ordered]@{  
                'Invoice No.' = 30,123,203,172    
                'Date' = 130,230,150,245    
            }
        }

    #>
    param(
        # Filepath
        [Parameter(mandatory=$true,ValueFromPipeline=$true,ValueFromPipelineByPropertyName=$true)][Alias('Fullname')][string[]]$path,    
        # hashtable with definition of pages and locations
        [Parameter(mandatory=$true)][hashtable]$locations,
        # optional match page content by Regex instead of page number
        [Parameter(mandatory=$false)][switch]$UsePageSelectRegex,
        # output source fileinfo
        [Parameter(mandatory=$false)][switch]$passthru
    )

    begin{
        # function to load iTextSharp assembly
        function Load-iTextLibrary {
            if($psscriptroot -ne ''){    
                $localpath = join-path $psscriptroot 'itextsharp.dll'    
            }else{
                $localpath = join-path $env:TEMP 'itextsharp.dll'    
            }
            $zip = $null;$tmp = ''    
            $script:iTextDLL = $localpath
            try{
                if(!(Test-Path $localpath)){
                    Add-Type -A System.IO.Compression.FileSystem
                    $tmp = "$env:TEMP\$([IO.Path]::GetRandomFileName())"    
                    write-host "Downloading and extracting required 'iTextSharp.dll' ... " -F Green -NoNewline    

                    (New-Object System.Net.WebClient).DownloadFile('https://www.nuget.org/api/v2/package/iTextSharp/5.5.13.1', $tmp)    
                    $zip = [System.IO.Compression.ZipFile]::OpenRead($tmp)
                    $zip.Entries | ?{$_.Fullname -eq 'lib/itextsharp.dll'} | %{    
                        [System.IO.Compression.ZipFileExtensions]::ExtractToFile($_,$localpath)
                    }
                    write-host "OK" -F Green     
                }
                if (Get-Item $localpath -Stream zone.identifier -ea SilentlyContinue){
                    Unblock-File -Path $localpath
                }
                Add-Type -Path $localpath
            }catch{
                throw "Error: $($_.Exception.Message)"    
            }finally{
                if ($zip){$zip.Dispose()}
                if ($tmp -ne ''){del $tmp -Force -EA SilentlyContinue}    
            }
        }

        # load iText library
        Load-iTextLibrary

        # also read protected files
        [iTextSharp.text.pdf.PdfReader]::unethicalreading = $true

        if (!("iText.GetTextFromLocationStrategy" -as [type])){    
            # Create GetTextFromLocationStrategy class
            Add-Type -ReferencedAssemblies $script:iTextDLL -TypeDefinition '    
                using System.Text;
                using System.Collections;
                using System.Collections.Generic;
                using System.Collections.Specialized;
                using iTextSharp.text;
                using iTextSharp.text.pdf.parser;
                using iTextSharp.text.pdf;

                namespace iText {
                    public class GetTextFromLocationStrategy : iTextSharp.text.pdf.parser.ITextExtractionStrategy {
                        StringBuilder result = new StringBuilder();
                        OrderedDictionary resultfields = new OrderedDictionary();
                        Vector lastBaseline = new Vector(0, 0, 0);
                        OrderedDictionary rects;

                        public GetTextFromLocationStrategy(OrderedDictionary locations) {
                            rects = locations;
                        }

                        public void RenderText(TextRenderInfo renderinfo) {
                            Vector curBaseline = renderinfo.GetBaseline().GetStartPoint();
                            Vector topRight = renderinfo.GetAscentLine().GetStartPoint();
                            iTextSharp.text.Rectangle rect = new iTextSharp.text.Rectangle(curBaseline[Vector.I1], curBaseline[Vector.I2], topRight[Vector.I1], topRight[Vector.I2]);
                            foreach (DictionaryEntry field in rects) {
                                Rectangle fieldRect = (Rectangle)field.Value;
                                if (!resultfields.Contains(field.Key)) {
                                    resultfields.Add(field.Key, new StringBuilder(""));    
                                }
                                if (rect.Left >= fieldRect.Left && rect.Top <= fieldRect.Top && rect.Bottom >= fieldRect.Bottom && rect.Right <= fieldRect.Right) {
                                    if (lastBaseline[Vector.I2] > curBaseline[Vector.I2] && ((StringBuilder)resultfields[field.Key]).ToString() != "") {    
                                        ((StringBuilder)resultfields[field.Key]).Append("\r\n");  
                                        result.Append("\r\n");    
                                    }
                                    string txt = renderinfo.GetText();
                                    ((StringBuilder)resultfields[field.Key]).Append(txt);
                                    result.Append(txt);
                                }
                            }
                            lastBaseline = curBaseline;
                        }

                        public string GetResultantText() {
                            return result.ToString();
                        }
                        public OrderedDictionary GetResultTable() {
                            OrderedDictionary dic = new OrderedDictionary();
                            foreach (DictionaryEntry field in resultfields) {
                                dic.Add(field.Key, field.Value.ToString());
                            }
                            return dic;
                        }
                        public void BeginTextBlock() { }
                        public void EndTextBlock() { }
                        public void RenderImage(ImageRenderInfo renderinfo) { }
                    }
                }
            '    
        }
        
        
        # function to convert mm to points
        function M2P($mm){$mm * 2.83465}

        # convert user supplied location values to rectangles with point values
        foreach($page in $locations.Keys){
            [string[]]$locations.$page.Keys | %{$locations.$page.$_ = [iTextSharp.text.Rectangle]::new((M2P $locations.$page.$_[0]),(M2P $locations.$page.$_[1]),(M2P $locations.$page.$_[2]),(M2P $locations.$page.$_[3]))}
        }
    }

    process{
        # Process files
        foreach($file in $path) {
            $reader = $null
            try {
                write-verbose "Processing file '$($file)' ..."    
                # create pdf reader object
                $reader = New-Object iTextSharp.text.pdf.PdfReader $file
                # parse pages defined
                
                if ($UsePageSelectRegex.IsPresent){
                    $pages = 1..($reader.NumberOfPages) | ?{[iTextSharp.text.pdf.parser.PdfTextExtractor]::GetTextFromPage($reader,$_) -match $locations.Keys[0]}
                    $data = foreach($page in $pages){
                         # create instance from custom Text-Extraction-Strategy
                         $strategy = New-Object iText.GetTextFromLocationStrategy ([System.Collections.Specialized.OrderedDictionary]($locations.([string]$locations.Keys[0])))
                         # execute extraction
                         [void][iTextSharp.text.pdf.parser.PdfTextExtractor]::GetTextFromPage($reader,$page,$strategy)
                         # output result object
                         [pscustomobject]$strategy.GetResultTable()
                    }
                }else{
                    $data = [ordered]@{}
                    $pages = $locations.Keys | ?{$_ -in (1..$reader.NumberOfPages)}
                    foreach($page in $pages){
                         # create instance from custom Text-Extraction-Strategy
                         $strategy = New-Object iText.GetTextFromLocationStrategy ([System.Collections.Specialized.OrderedDictionary]$locations.$page)
                         # execute extraction
                         [void][iTextSharp.text.pdf.parser.PdfTextExtractor]::GetTextFromPage($reader,$page,$strategy)
                         # append to result object
                         $data += $strategy.GetResultTable()
                    }
                    if ($passthru.IsPresent){
                        # append file info
                        $data.FileInfo = Get-Item $file
                    }
                    $data = [pscustomobject]$data
                }
                # output data
                $data
            }catch{
                write-host $_.Exception.Message -F Red
            }finally{
                if($reader){$reader.Dispose()}
            }
        }
    }
}

[LOWER-LEFT-X] , [LOWER-LEFT-Y] , [UPPER-RIGHT-X] , [UPPER-RIGHT-Y]

$locations = @{
    1 = [ordered]@{
        'Geschäftspartnernummer' = 167,223,188,228  
        'Empfänger' = 23,211,73,233  
    }
}

$locations = @{
    1 = [ordered]@{
        'Geschäftspartnernummer' = 167,223,188,228  
        'Empfänger' = 23,211,73,233  
    }
    3 = [ordered]@{
        'Bankleitzahl' = 10,20,40,30  
    }
}

$locations = @{
    'Abschlagsänderung' = [ordered]@{  
        'Geschäftspartnernummer' = 167,223,188,228  
        'Empfänger' = 23,211,73,233  
    }
}

# Positionen der zu extrahierenden Textfelder definieren
$locations = @{
    1 = [ordered]@{
        'Geschäftspartnernummer' = 167,223,188,228  
        'Empfänger' = 23,211,73,233  
    }
}
# Ein einzelnes PDF-Dokumente auslesen und das Ergebnis auf der Konsole als Liste ausgeben
Get-PDFTextFromLocation -Path "E:\dokument.pdf" -locations $locations | format-list *  

# Positionen der zu extrahierenden Textfelder definieren
$locations = @{
    1 = [ordered]@{
        'Geschäftspartnernummer' = 167,223,188,228  
        'Empfänger' = 23,211,73,233  
    }
}
# Alle gleich aufgebauten PDF-Dokumente eines Ordners verarbeiten und das Ergebnis in eine CSV-Datei ausgeben
Get-ChildItem "E:\quelldaten" -File -Filter *.pdf | Get-PDFTextFromLocation -locations $locations | export-csv "E:\output.csv" -Delimiter ";" -NoTypeInformation -Encoding UTF8 -Force  

"Geschäftspartnernummer";"Empfänger"
"1222334455";"Herrn
Max Mustermann
Musterstr. 99
99999 Musterhausen"
"22342344";"Frau
Monika Musterfrau
Musterweg. 1
88888 Musterdorf"

# Regular Expression Pattern der auszuwählenden Seiten und Positionen der zu extrahierenden Textfelder definieren
$locations = @{
    '(?i)Abschlagsänderung' = [ordered]@{  
        'Geschäftspartnernummer' = 167,223,188,228  
        'Empfänger' = 23,211,73,233  
    }
}
# PDF-Dokument auslesen und das Ergebnis in eine CSV-Datei ausgeben
Get-PDFTextFromLocation -Path "E:\dokument.pdf" -locations $locations -UsePageSelectRegex | export-csv "E:\output.csv" -Delimiter ";" -NoTypeInformation -Encoding UTF8 -Force