Extract from XMP and insert into XML

Clive Andrews · Post by **Clive Andrews** » Wed Apr 04, 2012 5:30 pm

I know the probable answer here, but nothing gained by not asking....

Is there an easy/quick way to extract some XMP info from a PDF and write it into an XML file?

I have this "cunning plan"....

:)

dkelly · Post by **dkelly** » Wed Apr 04, 2012 10:42 pm

Apago PDFspy can retrieve XMP for document and/or individual images.

Clive Andrews · Post by **Clive Andrews** » Thu Apr 05, 2012 9:26 am

Yes sir, I can get to the XMP as metadata anyway - what I want to do is take a PDF, inject an XML template - but merge some extra data from the PDF itself into the metadata I have attached. It seems the easy way would be to dynamically modify the content of the Template before I inject it....

I'm trying to avoid having lots of similar flows - I have a good preflight, report and archive flow, but we will potentially get files delivered by FTP direct into Switch (iDropper or Switch clients are not an option) I can get them to embed the relevant info the source PDF - probably just the creator file name, so would like to merge that into our HTML email template....

dkelly · Post by **dkelly** » Thu Apr 05, 2012 4:06 pm

In that case, you can extract the XMP using Switch, modify it using an XSLT transform and write it back using XMP update.

Clive Andrews · Post by **Clive Andrews** » Sat Apr 07, 2012 7:59 pm

dkelly wrote: In that case, you can extract the XMP using Switch, modify it using an XSLT transform and write it back using XMP update.

Gotcha - simples!

I'm still trying to get my head round XSLT transformation (ok, stealing a lot of your work) - but I'll get there...

Thank you sir, happy easter!

dkelly · Post by **dkelly** » Mon Apr 09, 2012 3:26 pm

Give me a simple example of what you are trying to do and I'll see if I have an XSLT that I can easily modify for you.

Clive Andrews · Post by **Clive Andrews** » Tue Apr 10, 2012 10:19 am

No worries Dwight.

The xml template I have is:

Customer upload

FTP

Upload

c.andrews@printing-press-press.co.uk

Joe

Bloggs

01388 212716

customer_name@printing-press.co.uk

FTP Upload

Joe Bloggs Inc

I am trying to change the "Reference" - in BOLD and Red using the variable "dc:title" from XMP in the PDF - (I'm renaming it with [Metadata.Text:Path="dc:title/*[@xml:lang='x-default']"]

So it should be easy (if I knew what I was doing)

Obviously, I could use this as a base for a lot more - once I see how to do it.

My plan is to then squirt this folder back into my normal flows for preflight and reporting - and all I need the customer to do is get the ISBN into the Title field on the PDF document info, and I can then use it to marry up to the ISBN that will be raised by an EDI order (JDF)

Thanks Dwight.

dkelly · Post by **dkelly** » Tue Apr 10, 2012 4:56 pm

Try this:

function jobArrived( s : Switch, job : Job )

{

var theXmp = job.getDataset("Xmp");

var theCreator = "Unknown";

if (theXmp != null)

theCreator = theXmp.getString( "dc:creator/*[1]" , null );

var theXML = new Document();

// create the root node

var rootElem = theXML.createElement( "jobInfo", null );

theXML.appendChild( rootElem );

// create the children nodes

var theChildNode = theXML.createElement( "uuid", null );

rootElem.appendChild( theChildNode );

var theValue = theXML.createText( "Customer upload" );

theChildNode.appendChild( theValue );

theChildNode = theXML.createElement( "First_name", null );

rootElem.appendChild( theChildNode );

theValue = theXML.createText( "FTP" );

theChildNode.appendChild( theValue );

theChildNode = theXML.createElement( "Reference", null );

rootElem.appendChild( theChildNode );

theValue = theXML.createText( theCreator );

theChildNode.appendChild( theValue );

// save the XML document

var theDataset = job.createDataset("XML");

var theXMLFilename = theDataset.getPath();

theXML.save(theXMLFilename);

job.setDataset("jobticket", theDataset);

delete theXML;

job.sendToSingle(job.getPath());

}

You'll have to create the other elements but this will get you started.

Cheers,

Dwight Kelly

Apago, Inc.

dkelly@apago.com

tips accepted

Clive Andrews · Post by **Clive Andrews** » Tue Apr 10, 2012 5:16 pm

Yes - I see, that's easier, just use the script to write the XML, with static values for all bar that 1. That appears to mean I don't need to inject a static file anymore?

dkelly · Post by **dkelly** » Tue Apr 10, 2012 5:29 pm

No reason you couldn't read the injected template XML and modify it instead of creating it from scratch.. Document() creator optional takes a filename of existing XML file to read.

Clive Andrews · Post by **Clive Andrews** » Wed Apr 11, 2012 12:15 pm

dkelly wrote: No reason you couldn't read the injected template XML and modify it instead of creating it from scratch.. Document() creator optional takes a filename of existing XML file to read.

Thanks Dwight - I've ended up using your script and just taking it as job ticket info - all I need do is add variables to that as well as the XML (i.e. from iDropper source) - it works well.

It gives me a basis I have for the 100 odd new ideas rattling around inside my brain....

Edit: One small issue occurring now - I cannot get the jobticket info into my template:

A notification has been passed to [Metadata.Text:Path="/XMLSTINGRAYInJobReport/CSR_First",Dataset="Xml",Model=XML][Metadata.Text:Path="/jobInfo/CSR_First",Dataset="Xml",Model=XML][Metadata.Text:Path="/jobInfo/CSR_First",Dataset="jobticket",Model=XML] [Metadata.Text:Path="/jobInfo/CSR_Last",Dataset="Xml",Model=XML][Metadata.Text:Path="/jobInfo/CSR_Last",Dataset="jobticket",Model=XML][Metadata.Text:Path="/XMLSTINGRAYInJobReport/CSR_Last",Dataset="Xml",Model=XML] who will be your contact within Ashford Colour Press.

As you can see - I have 3 different potential sources, only 1 of which will ever be used.

[Metadata.Text:Path="/XMLSTINGRAYInJobReport/CSR_First",Dataset="Xml",Model=XML]

and

[Metadata.Text:Path="/jobInfo/CSR_First",Dataset="Xml",Model=XML]

work fine, however:

[Metadata.Text:Path="/jobInfo/CSR_First",Dataset="jobticket",Model=XML]

Doesn't....

Most perplexing. Any suggestions anyone?

Clive Andrews · Post by **Clive Andrews** » Thu Apr 12, 2012 9:39 am

Is it because the dataset needs to be XML to use in my HTML template...?