Extract content from a PDF file, how?

The API documentation being down is frustrating...  Trust me.

Doing what you want via VBScript will be painful.  Have you exhausted using some of the off-the-shelf IML components for this?

The IML solutions can probably do this - not sure that they have fully changed from making suggestions to actually setting metadata automatically but it should happen around this time. However, the customer's current license would have to be upgraded meaning triple payment compared to their current license. In that situation it is well worth my time spending an afternoon or a day creating a script. The use case is for standardized incoming purchase orders where it is quite simple to extract the relevant data with regex, and where file length is quite limited. So once I get the text content into a variable in the script is is fairly straight forward. Obviously, it would fail if the incoming file format changes...!

Was able to find some old postings mentioning the GetTextContentForFile. From the context it seems like it might be the way forward. Hope to get access to the documentation sometime soon to confirm this.

let me second that request ;)

Beware that computers do not always read content of a PDF in the same order as humans!
To me the document has "Dato: " and then a date in the format DD-MM-YYYY.
However the computer gets the date first and then the text string in this particular instance.
I discovered this by copying the content of the PDF into Notepad and then examine it there. This procedure would be helpful anytime you need to create a regex!

'Extract metadata from file content using regex.
'Suited for documents with a fixed content format and limited file size.
'2022.06.27 Karl Lausten
Option Explicit

Dim objID
set objID = ObjVer.ObjID
Dim objVersion
set objVersion = Vault.ObjectOperations.GetLatestObjectVersionAndProperties(objID, true)
Dim myFiles : Set myFiles = objVersion.VersionData.Files
Dim myFile
Dim fileVer
For Each myFile in myFiles
	set fileVer = myFile.FileVer
Next
'This script was made to run on single file objects. The script is not prepared to handle multiple files!
'Running it on documents with multiple files will create problems such as propertyvalues being overwritten in the proces below.


Dim szFullText
szFullText = Vault.ObjectFileOperations.GetTextContentForFile(ObjVer, fileVer)
Dim szMatch
Dim oPropertyValue
Dim oPattern : Set oPattern = New RegExp

'get Document date, this section can be repeated with different properties and patterns as needed
Dim iPDDocDate : iPDDocDate = Vault.PropertyDefOperations.GetPropertyDefIDbyAlias("PD.DocumentDate")
With oPattern
	.Pattern = "(\d{2}\-\d{2}\-\d{4})(?=\s*Dato:)"
	.IgnoreCase = True
	.Global = True
End With
set szMatch = oPattern.Execute(szFullText)
Set oPropertyValue = PropertyValues.SearchForProperty(iPDDocDate)
oPropertyValue.TypedValue.SetValue MFDatatypeDate, szMatch.Item(0).SubMatches(0)
Vault.ObjectPropertyOperations.SetProperty ObjVer,oPropertyValue

MANY thanks !

What do you guys think about creating a github repo for collecting such code snippets?

It's something I (at M-Files) have toyed with.  The issue is maintaining all the random ones.  If this were a community-oriented one, very explicitly saying that it's nothing to do with our official ones, then that may work.

I do very much worry about the maintenance though.

I guess me and a few others would surely help to maintain it.. Maybe I'll create a repo if I have time and start uploading a few examples.. If you have any suggestions or ideas let me know 

Edit:

Maybe a gist would be more suitable than a traditional repo..

There are lots of people around (including myself) who are not familiar with github, repo or gist. We wouldn't know the first thing about searching for stuff in those places. It needs to be on a platform that looks familiar to ordinary people without professional skills as developers.

Most obvious and familiar solution would be using an M-Files Vault itself but I cannot provide the ressources for that.. But I think a gist would not be complicated at all.. But to outline the terminology, GitHub is the Platform and there are 2 repository variants where ghist is intended for code snippets while i would create a classic repo for a whole project..

I created a PDF from Word with only "Dato: 22-01-2001" and copied your script to a workflow state (of course also created the "PD.DocumentDate" but it fails at line 35.. szFullText contains the text but szMatch.Count is 0 so I guess he hasn't found anything..

Update:

for me this expression works: "(Dato:\s*)(\d{2}\-\d{2}\-\d{4})"

Yes, as I mentioned in my comment to the code - you always have to adapt the regex to the way the computer sees the text. In my case computer read the "Dato:" part as being next line after the actual date even though it appear to be in front of the date.

Yes, as I mentioned in my comment to the code - you always have to adapt the regex to the way the computer sees the text. In my case computer read the "Dato:" part as being next line after the actual date even though it appear to be in front of the date.