Chapter 11
Document Information and Metadata

11.1 Reading Document Information

The -info operation prints entries from the document information dictionary, and from any XMP metadata to standard output.

$cpdf -info pdf_reference.pdf
Encryption: Not encrypted
Permissions:
Linearized: true
Object streams: true
ID: <0b1f990718e2a92c0c112fbf08b233fb> <b2f1dbee369e11d9b951000393c97fd8>
Version: 1.5
Pages: 1236
Title: PDF Reference, version 1.6
Author: Adobe Systems Incorporated
Subject: Adobe Portable Document Format (PDF)
Keywords:
Creator: FrameMaker 7.0
Producer: Acrobat Distiller 6.0.1 for Macintosh
Created: D:20041114084116Z
Modified: D:20041114163850-08'00'
Trapped: False
PageMode: UseOutlines
PageLayout:
HideToolbar:
HideMenubar:
HideWindowUI:
FitWindow:
CenterWindow:
DisplayDocTitle: true
NonFullScreenPageMode:
AcroForm: False
MediaBox: 0.000000 0.000000 612.000000 792.000000
CropBox: 41.000000 63.000000 572.000000 729.000000
BleedBox:
TrimBox: various
ArtBox: various
Subformats:
XMP dc:title: PDF Reference, version 1.6
XMP dc:creator: Adobe Systems Incorporated
XMP dc:description: Adobe Portable Document Format (PDF)

By default, cpdf strips to ASCII, discarding character codes in excess of 127. In order to preserve the original unicode, add the -utf8 option. To disable all post-processing of the string, add -raw. See Section 1.17 for more information.

The -info-json operation prints the information in JSON format instead. For example:

{
  "Encryption": "Not encrypted",
  "Permissions": [],
  "Linearized": true,
  "Object streams": true,
  "ID": [
    "0b1f990718e2a92c0c112fbf08b233fb", "b2f1dbee369e11d9b951000393c97fd8"
  ],
  "Version": [ 1, 5 ],
  "Pages": 1236,
  "Title": "PDF Reference, version 1.6",
  "Author": "Adobe Systems Incorporated",
  "Subject": "Adobe Portable Document Format (PDF)",
  "Keywords": null,
  "Creator": "FrameMaker 7.0",
  "Producer": "Acrobat Distiller 6.0.1 for Macintosh",
  "Created": "D:20041114084116Z",
  "Modified": "D:20041114163850-08'00'",
  "Trapped": false,
  "PageMode": "UseOutlines",
  "PageLayout": null,
  "HideToolbar": null,
  "HideMenubar": null,
  "HideWindowUI": null,
  "FitWindow": null,
  "CenterWindow": null,
  "DisplayDocTitle": true,
  "NonFullPageScreenMode": null,
  "AcroForm": false,
  "MediaBox": [ 0.0, 0.0, 612.0, 792.0 ],
  "CropBox": [ 41.0, 63.0, 572.0, 729.0 ],
  "BleedBox": null,
  "TrimBox": "various",
  "ArtBox": "various",
  "Subformats": [],
  "XMP dc:title": "PDF Reference, version 1.6",
  "XMP dc:creator": "Adobe Systems Incorporated",
  "XMP dc:description": "Adobe Portable Document Format (PDF)"
}

The -page-info operation prints the page label, media box and other boxes, and number of annotations page-by-page to standard output, for all pages in the current range.

$cpdf -page-info 14psfonts.pdf
Page 1:
Label: i
MediaBox: 0.000000 0.000000 600.000000 450.000000
CropBox: 200.000000 200.000000 500.000000 500.000000
BleedBox:
TrimBox:
ArtBox:
Rotation: 0
Annotations: 0

Note that the format for boxes is minimum x, minimum y, maximum x, maximum y. Using -page-info-json we can get the information in JSON format. For example:

[
  {
    "Page": 1,
    "Label": "i",
    "MediaBox": [ 0.0, 0.0, 600.0, 450.0 ],
    "CropBox": [ 200.0, 200.0, 500.0, 500.0 ],
    "BleedBox": null,
    "TrimBox": null,
    "ArtBox": null,
    "Rotation": 0,
    "Annotations": 0
  }
]

The -pages operation prints the number of pages in the file.

cpdf -pages Archos.pdf
8

11.2 Setting Document Information

The document information dictionary in a PDF file specifies various pieces of information about a PDF. These can be consulted in a PDF viewer (for instance, Acrobat).

Here is a summary of the commands for setting entries in the document information dictionary:

(The details of the format for creation and modification dates can be found in Appendix 19.11. Using the date "now" uses the time and date at which the command is executed. Note also that -producer and -creator may be used to set the producer and/or the creator when writing any file, separate from the operations described in this chapter.)

For example, to set the title, the full command line would be

To set also any field in the XMP metadata, add -also-set-xmp. The field must exist already. To set only the field (not the document information dictionary), add -just-set-xmp instead.

To delete existing non-XMP metadata in line with PDF 2.0, use -remove-dict-entry "/Info" as described in chapter 19.

11.3 XMP Metadata

PDF files can contain a piece of arbitrary metadata, often in XMP format. This is typically stored in an uncompressed stream, so that other applications can read it without having to decode the whole PDF. To set the metadata:

11.4 Upon Opening a Document

When considering the following commands, also NB: If the file has a valid /OpenAction setting, which tells the PDF reader to open at a certain page or position on a page, this can override the page layout or display options. To prevent this, preprocess the file with the -remove-dict-entry functionality from Section 19.9:

11.4.1 Page Layout

The -set-page-layout operation specifies the page layout to be used when a document is opened in, for instance, Acrobat. The possible (case-sensitive) values are:

For instance:

11.4.2 Page Mode

The page mode in a PDF file defines how a viewer should display the document when first opened. The possible (case-sensitive) values are:

For instance:

If full screen mode is selected for document opening, we can also set a mode to be used when the user exits from full-screen mode:

As would be expected, FullScreen is not allowed here.

11.4.3 Display Options

The appearance of the PDF viewer upon opening a document may be set with these options. Each is boolean - supply true or false:

For instance:

The page a PDF file opens at can be set using -open-at-page:

To have that page scaled to fit the window in the viewer, use -open-at-page-fit instead:

(Here, we used end to open at the last page. Any page specification describing a single page is ok here.)

Alternatively, we may specify a full destination, of the kind described on page 39:

11.5 Page Labels

It is possible to add page labels to a document. These are not the printed on the page, but may be displayed alongside thumbnails or in print dialogue boxes by PDF readers. We use -add-page-labels to do this, by default with decimal arabic numbers (1,2,3…). We can add -label-style to choose what type of labels to add from these kinds:

We can use -label-prefix to add a textual prefix to each label. Consider a file with twenty pages and no current page labels (a PDF reader will assume 1,2,3…if there are none). We will add the following page labels:

i, ii, iii, iv, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, A-0, A-1, A-2, A-3, A-4, A-5

Here are the commands, in order:

By default the labels begin at page number 1 for each range. To override this, we can use -label-startval (we used 0 in the final command), where we want the numbers to begin at zero rather than one. The option -labels-progress can be added to make sure the start value progresses between sub-ranges when the page range specified is disjoint, e.g 1-9, 30-40 or odd.

Page labels may be removed altogether by using -remove-page-labels command. To print the page labels from an existing file, use -print-page-labels. For example:

$ cpdf -print-page-labels in.pdf
labelstyle: LowercaseRoman
labelprefix: None
startpage: 1
startvalue: 1
labelstyle: DecimalArabic
labelprefix: A
startpage: 9
startvalue: 1

Or, in JSON format with -print-page-labels-json:

[
  {
    "labelstyle": "LowercaseRoman",
    "labelprefix": null,
    "startpage": 1,
    "startvalue": 1
  },
  {
    "labelstyle": "DecimalArabic",
    "labelprefix": "A",
    "startpage": 9,
    "startvalue": 1
  }
]

11.6 Composition of a PDF

The -composition and -composition-json operations show how much space in a PDF is used by each kind of data. Here is the output of -composition for this manual:

$ cpdf -composition cpdfmanual.pdf
Images: 0 bytes (0.00%)
Fonts: 144731 bytes (46.72%)
Content streams: 132767 bytes (42.85%)
Structure Info: 0 bytes (0.00%)
Attached Files: 0 bytes (0.00%)
XRef Table: 21082 bytes (6.80%)
Unclassified: 11229 bytes (3.62%)

And here it is in JSON format:

$ cpdf -composition-json cpdfmanual.pdf
[
  ("Images", 0, 0.0),
  ("Fonts", 144731, 46.71620256351494),
  ("Content streams", 132767, 42.854468398271194),
  ("Structure Info", 0, 0.0),
  ("Attached Files", 0, 0.0),
  ("XRef Table", 21082, 6.8048378194306816),
  ("Unclassified", 11229, 3.6244912187831857)
]

Note that, due to small inaccuracies in the method, it is possible for the Unclassified number to be negative.