data from PDF documents in Lua filters?

[pr-apes]

@pgundlach,

I would like to rewrite for Publisher some basic hacks that were written in ConTeXt.

These scripts play with page imposition and image clipping.

I guess that the filters allow standard Lua code and some modules (it would be great to know which ones are available).

But besides access to variables (option -v), I wonder whether it is possible to have access in the filter to the following PDF info:

1. Dimensions: height, width, number of pages.

2. Metadata: author, title, subject...

Many thanks for your help.

[pgundlach]

No. You can get the number of pages by reading the -aux.xml file and analyse that.

What you can do is write the data you need to an external XML file (SaveDataset) and parse the data after the run.

[pr-apes]

Many thanks for your reply, @pgundlach.

Sorry, but I’m not sure I get the point (probably because I may be aiming at something not achievable by Publisher).

I have a bunch of PDF documents with different page sizes and total numbers of pages (they are generated by other tools).

ConTeXt runs a kind of Lua script (with “PDF superpowers”) that depending on the page dimensions makes a different imposition, such as the following ones:

• If page size is landscape legal, please clip it to landscape A4.
• If page size is portrait legal, please append it to another portrait A4, after clipping it to A4 page size.
• If page size is portrait A5, please clip a predefined image area and impose it centered (both vertically and horizontally) in an A4 page.
• And some other cases...

Lua filters cannot read dimensions or other data from external PDF documents.

To be able to read these dimensions, the following code will not be of use:

<Layout xmlns="urn:speedata.de:2009/publisher/en"
  xmlns:sd="urn:speedata:2009/publisher/functions/en">

  <Record element="data">
    <SetVariable variable="myfile" select="$external_pdfdoc"/>
    <Loop select="sd:number-of-pages($myfile)" variable="page">
      <PlaceObject column="0mm" row="0mm">
        <Image file="{$myfile}" width="210mm" page="{$page}"/>
      </PlaceObject>
      <ClearPage/>
    </Loop>
  </Record>

</Layout>

Because the image is placed on an already given pages, that will be reading page size dimensions of the main PDF documents.

I wonder whether there is something similar in Publisher to what a TEXpage is in ConTeXt (page with no extra dimensions that the ones required by its contents, such as an image).

But I see that I can have that with XPath functions (no metadata, though): sd:aspectratio(), sd:imageheight(), sd:imagewidth() (and sd:aspectratio()).

This could help to do it:

<Layout xmlns="urn:speedata.de:2009/publisher/en"
  xmlns:sd="urn:speedata:2009/publisher/functions/en">

  <Record element="data">
    <SetVariable variable="myfile" select="$external_pdfdoc"/>
	<Output>
	<Text>
	<Paragraph>
    <Value select="sd:number-of-pages($myfile)"/>
	<Value> </Value>
    <Value select="sd:imageheight($myfile)"/>
	<Value> </Value>
    <Value select="sd:imagewidth($myfile)"/>
	</Paragraph>
      </Text>
    </Output>
  </Record>

</Layout>

The only minor issue is that the measurement unit is grid cells.

I think this would set grid cells in millimiters:

<SetGrid width="1mm" height="1mm"/>

Now I see how it might work.

Many thanks for your help.

[pgundlach]

I didn't think of external PDFs. The data from these files are not accessible in the preprocessing step.

Imageheight and imagewitdth have an undocumented feature (sorry - will be documented in the next version):

sd:imageheight($myfile, 'mm') returns the number of millimetres (same with imagewidth()).

[pr-apes]

That would be a great feature.

[pr-apes]

Even greater would be sd:pdfauthor() sd:title() and similar ones...

[pr-apes]

Sorry, I get it know, it was not an unimplemented feature, just an undocumented one.