├── .gitignore
├── Initialize-Environment.ps1
├── README.md
├── Sort-Files.ps1
├── datelib.psm1
├── itextsharplib.psm1
└── tesseractlib.psm1


/.gitignore:
--------------------------------------------------------------------------------
1 | Input/
2 | Output/
3 | Lib/
4 | Test/
5 | *.zip
6 | *.jpg


--------------------------------------------------------------------------------
/Initialize-Environment.ps1:
--------------------------------------------------------------------------------
 1 | #
 2 | # Title:     Initialize-Environment.ps1
 3 | # Author:    Jourdan Templeton
 4 | # Email:     hello@jourdant.me
 5 | # Modified:  21/07/2015 04:35PM NZDT
 6 | #
 7 | 
 8 | #properties
 9 | #================================================================
10 | $tesseract_url = "https://nuget.org/api/v2/package/Tesseract"
11 | $tessdata_url = "https://nuget.org/api/v2/package/tesseract-ocr"
12 | $itextsharp_url = "https://nuget.org/api/v2/package/iTextSharp"
13 | 
14 | $tesseract_zip_name = "$PSScriptRoot\tesseract.zip"
15 | $tessdata_zip_name = "$PSScriptRoot\tessdata.zip"
16 | $itextsharp_zip_name = "$PSScriptRoot\itextsharp.zip"
17 | 
18 | $input_dir_name = "$PSScriptRoot\Input"
19 | $output_dir_name = "$PSScriptRoot\Output"
20 | $lib_dir_name = "$PSScriptRoot\Lib"
21 | #================================================================
22 | 
23 | #create dir structure
24 | If ((Test-Path $input_dir_name) -eq $False) { mkdir $input_dir_name | Out-Null }
25 | If ((Test-Path $output_dir_name) -eq $False) { mkdir $output_dir_name | Out-Null }
26 | If ((Test-Path $lib_dir_name) -eq $False) { mkdir $lib_dir_name | Out-Null }
27 | 
28 | #import assemblies into session
29 | Add-Type -Assembly "System.IO.Compression.FileSystem"
30 | 
31 | #download and extract tesseract libraries
32 | If ((Test-Path $tesseract_zip_name) -eq $False) {
33 | 	Write-Output "Downloading: '$tesseract_url'  To: '$tesseract_zip_name'"
34 | 	Invoke-WebRequest -Uri $tesseract_url -OutFile  $tesseract_zip_name
35 | }
36 | 
37 | If ((Test-Path $tesseract_zip_name) -eq $True)
38 | {
39 | 	$zip = [IO.Compression.ZipFile]::OpenRead($tesseract_zip_name)
40 | 	
41 | 	#extract tesseract libraries
42 | 	$zip.Entries | Where FullName -match "(x86|x64)|net451/tesseract\.dll" | % {
43 | 		$dir = (Get-Item $lib_dir_name).FullName
44 | 		If ($_.FullName.Contains("content")) { $dir +=  "\" + $matches[0] }
45 | 		If ((Test-Path $dir) -eq $False) { mkdir $dir | Out-Null }
46 | 
47 | 		$file = $dir + "\" + $_.Name
48 | 		[IO.Compression.ZipFileExtensions]::ExtractToFile($_, $file, $true) 
49 | 	}
50 | 
51 | 	$zip.Dispose()
52 | 	Remove-Item $tesseract_zip_name -Force
53 | }
54 | 
55 | #download and extract tesseract data files
56 | If ((Test-Path $tessdata_zip_name) -eq $False) {
57 | 	Write-Output "Downloading: '$tessdata_url'  To: '$tessdata_zip_name'"
58 | 	Invoke-WebRequest -Uri $tessdata_url -OutFile  $tessdata_zip_name
59 | }
60 | 
61 | If ((Test-Path $tessdata_zip_name) -eq $True)
62 | {
63 | 	$zip = [IO.Compression.ZipFile]::OpenRead($tessdata_zip_name)
64 | 	
65 | 	#extract tessdata libraries
66 | 	$zip.Entries | Where FullName -match "eng" | % {
67 | 		$dir = (Get-Item $lib_dir_name).FullName + "\tessdata"
68 | 		If ((Test-Path $dir) -eq $False) { mkdir $dir | Out-Null }
69 | 		
70 | 		$file = $dir + "\" + $_.Name
71 | 		[IO.Compression.ZipFileExtensions]::ExtractToFile($_, $file, $true) 
72 | 	}
73 | 
74 | 	$zip.Dispose()
75 | 	Remove-Item $tessdata_zip_name -Force
76 | }
77 | 
78 | #download and extract iTextSharp library
79 | If ((Test-Path $itextsharp_zip_name) -eq $False) { 
80 | 	Write-Output "Downloading: '$itextsharp_url'  To: '$itextsharp_zip_name'"
81 | 	Invoke-WebRequest -Uri $itextsharp_url -OutFile  $itextsharp_zip_name
82 | }
83 | 
84 | If ((Test-Path $itextsharp_zip_name) -eq $True)
85 | {
86 | 	$zip = [IO.Compression.ZipFile]::OpenRead($itextsharp_zip_name)
87 | 	
88 | 	#extract itextsharp libraries
89 | 	$zip.Entries | Where FullName -match "itextsharp.dll" | % {
90 | 		$dir = (Get-Item $lib_dir_name).FullName
91 | 		$file = $dir + "\" + $_.Name
92 | 		[IO.Compression.ZipFileExtensions]::ExtractToFile($_, $file, $true) 
93 | 	}
94 | 
95 | 	$zip.Dispose()
96 | 	Remove-Item $itextsharp_zip_name -Force
97 | }
98 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | #powershell-paperless
 2 | _Going paperless in a world filled with paper._
 3 | 
 4 | > **If you have just arrived here, check out my blog post for a better explanation: [http://blog.jourdant.me/powershell-and-tesseract-going-paperless-with-ocr/](http://blog.jourdant.me/powershell-and-tesseract-going-paperless-with-ocr/)**
 5 | 
 6 | These scripts will allow you to harness the power of Tesseract and PowerShell to sort scanned copies of documents. Or use OCR on any image in general.
 7 | 
 8 | 
 9 | ##Usage
10 | 1. Clone repo to your machine
11 | 2. Run Initialize-Environment.ps1 - this will download the necessary libraries and extract the files to the correct locations
12 | 3. Run Sort-Images.ps1 with your directories
13 | 4. High five!
14 | 
15 | <br/>
16 | <br/>
17 | <br/>
18 | 
19 | ######Jourdan Templeton - 2015
20 | 


--------------------------------------------------------------------------------
/Sort-Files.ps1:
--------------------------------------------------------------------------------
 1 | #
 2 | # Title:     Sort-Files.ps1
 3 | # Author:    Jourdan Templeton
 4 | # Email:     hello@jourdant.me
 5 | # Modified:  10/01/2015 21:52PM NZDT
 6 | #
 7 | 
 8 | [CmdletBinding(SupportsShouldProcess=$true)]Param(
 9 | 	[Parameter(Mandatory=$true)][string]$InputDirectory,
10 | 	[Parameter(Mandatory=$true)][string]$OutputDirectory
11 | )
12 | 
13 | #check paths
14 | If ((Test-Path $InputDirectory) -eq $False) { Throw "Input directory does not exist." } Else { $InputDirectory = (Get-Item $InputDirectory).FullName }
15 | "Input Directory:" + $InputDirectory
16 | If ((Test-Path $OutputDirectory) -eq $False) { Throw "Output directory does not exist." } Else { $OutputDirectory = (Get-Item $OutputDirectory).FullName }
17 | "Output Directory: " + $OutputDirectory + " `r`n"
18 | 
19 | #import libraries
20 | Remove-Module *lib
21 | Get-ChildItem -Filter "*lib.psm1" | % { Import-Module $_.FullName }
22 | 
23 | #get files
24 | $files = Get-ChildItem -Path $InputDirectory  -Recurse -Filter *.pdf | Where-Object { !$_.PSIsContainer }
25 | "Total files to process: " + $files.count
26 | 
27 | #process files
28 | If ($files.count -lt 1) { "No files to process. Closing..."; return }
29 | ForEach ($file in $files)
30 | {	
31 | 	"Processing: " + $file.FullName
32 | 	#ocr image with tesseract
33 | 	If ($file.Name.Split('.')[-1] -notmatch "pdf")
34 | 	{
35 | 		$image = New-Object System.Drawing.Bitmap($file.FullName)
36 | 		$ocr = Get-TessTextFromImage -Image $image
37 | 		"Confidence: " + ($ocr.Confidence * 100).ToString("##") + "%"
38 | 		
39 | 		$text = $ocr.Text
40 | 		$ocr = $null
41 | 
42 | 	#process with itextsharp
43 | 	} Else {
44 | 		$text = Get-ItsTextFromPdf -Path $file.FullName
45 | 	}
46 | 
47 | 	#get date
48 | 	$path = $OutputDirectory
49 | 	$date = ''
50 | 	$output = ''
51 | 
52 | 	Try 
53 | 	{
54 | 		#parse date and sort
55 | 		$date = Find-Date -InputText $text
56 | 		$path += "/" + $date.ToString("yyyy/MM-MMM")
57 | 		$output = $path + "/" + $date.ToString("yyyy-MM-dd_") + $file.Name.Replace(" ", "")
58 | 	}
59 | 	Catch 
60 | 	{
61 | 		$path += "/Unknown"
62 | 		$output = $path + "/" + $file.Name
63 | 	}
64 | 	Finally 
65 | 	{
66 | 		#create output dir
67 | 		If ((Test-Path $path) -eq $false) { mkdir $path | Out-Null }
68 | 		"Copying to: '" + $output + "'"
69 | 		Copy-Item -Path $file.FullName -Destination $output -Force
70 | 		
71 | 		$text = $null
72 | 	}
73 | 
74 | 	"`r`n"
75 | }
76 | 


--------------------------------------------------------------------------------
/datelib.psm1:
--------------------------------------------------------------------------------
 1 | #
 2 | # Title:     itextsharplib.psm1
 3 | # Author:    Jourdan Templeton
 4 | # Email:     hello@jourdant.me
 5 | # Modified:  10/01/2015 21:49PM NZDT
 6 | #
 7 | 
 8 | ###regex breakdown
 9 | $r_day	=				'(?<!\d)(3[0-1]|[1-2]\d|0?\d)(?:\s?[SsTt][TtHh])?'	
10 | $r_month_0 =			'(1[0-2]|0?[1-9]|'																														#Captures 01-31 and matches with the optional st or th
11 | $r_month_1 =			'A[UP][RG][A-z]{0,3}|[J][AU][NL][A-z]{0,4}|FEB[A-z]{0,5}|MA[RY][A-z]{0,2}|SEP[A-z]{0,6}|OCT[A-z]{0,4}|NOV[A-z]{0,5}|DEC[A-z]{0,5})'		#Captures Jan-Dec
12 | $r_month =				$r_month_0 + $r_month_1																													#Captures 01-12 or Jan-Dec
13 | $r_month_name =			'(' + $r_month_1																														#Captures 
14 | $r_year =				'((?:19)?9\d|(?:20)?[01]\d)'																											#Captures 1991-1999 2000-2016 (91-99 00-16)
15 | $r_seperator =			'(?:(?<!\n|\r)[\W-_]{0,2})?'																											#Captures ' ' \ / - , .
16 | 
17 | $patterns =				@(($r_day + $r_seperator + $r_month + $r_seperator + $r_year),				# 20/10/1999
18 | 						  ('(?<!\d)' + $r_month + $r_seperator + $r_day + $r_seperator + $r_year),	# September 20th 2013
19 | 						  ($r_year + $r_seperator + $r_month + $r_seperator + $r_day),				# 2014-2-31
20 | 						  ($r_year + $r_seperator + $r_day + $r_seperator + $r_month),				# 2014/31/may
21 | 						  ($r_year + $r_seperator + $r_month_name),									# 2014-May
22 | 						  ($r_month_name + $r_seperator + $r_year))									# Oct 2013
23 | 
24 | $patterns | % { Write-Host ($_ + " `r`n") }
25 | ###/regex breakdown
26 | 
27 | $min_year = 1991
28 | $max_year = 2020
29 | 
30 | <#
31 | .SYNOPSIS
32 | 
33 | Returns a date from the given text.
34 | .DESCRIPTION
35 | 
36 | This cmdlet agressively parses the text to find a date matching multiple formats. These formats can be adjusted in the module code.
37 | .PARAMETER InputText
38 | 
39 | The text to be parsed.
40 | 
41 | .EXAMPLE
42 | 
43 | $text = Get-TessTextFromImage -Path C:\Temp\temp.jpg
44 | $date = Find-Date -InputText $text
45 | #>
46 | Function Find-Date([Parameter(Mandatory=$true)][string]$InputText) 
47 | {
48 | 	If ($InputText -eq $null) { throw "$InputText cannot be null" }
49 | 
50 | 	$text = $InputText.ToUpper()
51 | 	Write-Verbose $text
52 | 
53 | 	#capture all dates within the image
54 | 	$dates = @()
55 | 	ForEach ($regex in $script:patterns) 
56 | 	{
57 | 		write-verbose $regex
58 | 		#regex
59 | 		$matches = ([regex]$regex).Matches($text)
60 | 		If ($matches.Count -gt 0) 
61 | 		{
62 | 			write-verbose $matches
63 | 			#select all parsable dates between set range (clean out as many false positives and mistakes as possible)
64 | 			$matches = $matches | Where-Object { $_.Value.Length -gt 5 } | % { Try { [DateTime]::Parse($_) } Catch {} } | Where-Object { $_.Year -ge $script:min_year -and $_.Year -le $script:max_year }
65 | 			Write-Verbose ("Matches==Null: " + ($matches -eq $null) + ", Total Matches: " + $matches.Count)
66 | 			If ($matches.Count -gt 0) { $dates += $matches[0] }
67 | 		}
68 | 
69 | 		#clear matches collection for next iteration
70 | 		$matches = $null
71 | 	}
72 | 
73 | 	#final logic
74 | 	If ($dates -ne $null -and $dates.Count -gt 0) 
75 | 	{
76 | 		#optional: custom date selection logic eg:
77 | 		#If ($InputText -match "statement|bank|account") { $dates = $dates | Sort -Descending }
78 | 		
79 | 		return $dates[0]
80 | 	} 
81 | 	Else 
82 | 	{ 
83 | 		If ($InputText.Contains(' ') -eq $false)  { throw "No date could be found." } 
84 | 		Else 
85 | 		{
86 | 			#recurse with text minus spaces
87 | 			return Find-Date -InputText $InputText.Replace(' ', '') 
88 | 		}
89 | 	}
90 | }


--------------------------------------------------------------------------------
/itextsharplib.psm1:
--------------------------------------------------------------------------------
 1 | #
 2 | # Title:     itextsharplib.psm1
 3 | # Author:    Jourdan Templeton
 4 | # Email:     hello@jourdant.me
 5 | # Modified:  10/01/2015 21:49PM NZDT
 6 | #
 7 | 
 8 | Add-Type -Path "$PSScriptRoot\Lib\itextsharp.dll"
 9 | 
10 | <#
11 | .SYNOPSIS
12 | 
13 | This cmdlet loads a PDF file and returns the text content.
14 | .DESCRIPTION
15 | 
16 | This cmdlet loads a PDF file and returns the text content. NOTE: this only applies to documents that have text fields embedded. This does not apply to text contained in images of the PDF.
17 | .PARAMETER Path
18 | 
19 | The path to the image to be processed.
20 | .EXAMPLE
21 | 
22 | Get-ItsTextFromImage -Path "C:\temp\test.pdf"
23 | .EXAMPLE
24 | 
25 | $text = Get-ChildItem "C:\Temp" -Filter *.pdf | Get-ItsTextFromImage
26 | #>
27 | Function Get-ItsTextFromPdf()
28 | {
29 | 	Param(
30 | 		[Parameter(Mandatory=$true, ValueFromPipeline=$true)][Alias("FullName")][String]$Path
31 | 	)
32 | 	Process {
33 | 		#construct reader object and prepare for reading
34 | 		$reader = New-Object iTextSharp.text.pdf.PdfReader($Path)
35 | 		
36 | 		#read pdf
37 | 		$ret = [iTextSharp.text.pdf.parser.PdfTextExtractor]::GetTextFromPage($reader, 1)
38 | 		
39 | 		#clean up references
40 | 		$reader.Dispose()
41 | 		return $ret
42 | 	}
43 | }


--------------------------------------------------------------------------------
/tesseractlib.psm1:
--------------------------------------------------------------------------------
 1 | #
 2 | # Title:     tesseractlib.psm1
 3 | # Author:    Jourdan Templeton
 4 | # Email:     hello@jourdant.me
 5 | # Modified:  04/01/2015 08:30PM NZDT
 6 | #
 7 | 
 8 | Add-Type -AssemblyName "System.Drawing"
 9 | Add-Type -Path "$PSScriptRoot\Lib\Tesseract.dll"
10 | $tesseract = New-Object Tesseract.TesseractEngine((Get-Item ".\Lib\tessdata").FullName, "eng", [Tesseract.EngineMode]::Default, $null)
11 | 
12 | <#
13 | .SYNOPSIS
14 | 
15 | This cmdlet loads either a file path or image and returns the text contained with the confidence.
16 | .DESCRIPTION
17 | 
18 | This cmdlet loads either a file path or image and returns the text contained with the confidence.
19 | You can pipe in either System.Drawing.Image file or a child-item object.
20 | .PARAMETER Image
21 | 
22 | The image file already loaded into memory.
23 | .PARAMETER FullName
24 | 
25 | The path to the image to be processed.
26 | .EXAMPLE
27 | 
28 | $image = New-Object System.Drawing.Bitmap("c:\test.jpg")
29 | Get-TessTextFromImage -Image $image
30 | .EXAMPLE
31 | 
32 | New-Object System.Drawing.Bitmap("C:\test.jpg") | Get-TessTextFromImage
33 | .EXAMPLE
34 | 
35 | $image = New-Object System.Drawing.Bitmap("c:\test.jpg")
36 | Get-TessTextFromImage -Image $image
37 | #>
38 | Function Get-TessTextFromImage()
39 | {
40 | 	Param(
41 | 		[Parameter(Mandatory=$true, ValueFromPipeline=$true, ParameterSetName="ImageObject")][System.Drawing.Image]$Image,
42 | 		[Parameter(Mandatory=$true, ValueFromPipeline=$true, ParameterSetName="FilePath")][Alias("FullName")][String]$Path
43 | 	)
44 | 	Process {
45 | 		#load image if path is a param
46 | 		If ($PsCmdlet.ParameterSetName -eq "FilePath") { $Image = New-Object System.Drawing.Bitmap((Get-Item $path).Fullname) } 
47 | 
48 | 		#perform OCR on image
49 | 		$pix = [Tesseract.PixConverter]::ToPix($image)
50 | 		$page = $tesseract.Process($pix)
51 | 		
52 | 		#build return object
53 | 		$ret = New-Object PSObject -Property @{"Text"= $page.GetText();
54 | 											   "Confidence"= $page.GetMeanConfidence()}
55 | 
56 | 		#clean up references
57 | 		$page.Dispose()
58 | 		If ($PsCmdlet.ParameterSetName -eq "FilePath") { $image.Dispose() } 
59 | 		return $ret
60 | 	}
61 | }


--------------------------------------------------------------------------------