I have a part in my code where I am programatically filling out PDF forms using iText Java based on user-entered data, and then I concat a number of such PDFs into one using iText again.
The PDF forms that are getting merged can be (and usually are) different.
The resulting PDF is way too large - looking at it, 98% of the space is taken by fonts.
The way I understand it, what happens is that the individual PDF forms have different font subsets, so when I merge them, I get massive amount of duplicate glyphs, except that the subsets are not identical, so I can't get rid of them without merging the subsets.
The other problem is that the PDF forms themselves might not even contain subsets, but heavily packed fonts that have 2000+ glyphs, so even if I manage to leave only one instance of that font in the PDF, that still can be many megabytes. Hence it seems that I need to be able to 1) create and 2) merge existing font subsets.
The quirk is that I do not control neither the PDF forms (that are being filled out) nor their number, nor the order in which they are concatenated, so it is not possible to solve this by controlling what kind of fonts are embedded in them.
Adobe Acrobat can of course solve such a problem - it can create and also merge font subsets - but I need a programatic, server-side solution. According to google hits, iText cannot do this. Is there another library that I could use (or anything else I can do)?
Related
Is it possible to merge layers of a PDF (OCG) with the base PDF to result in a PDF without layers?
I saw that it's possible to accomplish this using an application as Adobe Acrobat DC using a "Flatten Layers" option but I need this programmed in my Java application using iText7.
EDIT:
#joelgeraci has a useful and good answer that solves the previous question, but I have initially some hidden layers that will be displayed anyway when removing the OCProperties from the catalog.
You don't actually need to "merge" the layers. All of the layer content is already part of the page content. Layers, or more properly Optional Content Groups, are sets of instructions that the viewer can either draw or not, depending on the settings, for viewers that don't support layers, they just all show. To "flatten" the layers, you just need to modify the PDF so that the viewer doesn't think there is any optional content. The easiest way is to delete the OCProperties dictionary from the Catalog. Once you have the catalog object, use "remove" passing the name of the OCPropreties dictionary.
catalog.remove(PdfName.OCPROPERTIES)
I have a program which will be used for building questions database. I'm making it for a site that want user to know that contet was donwloaded from that site. That's why I want the output be PDF - almost everyone can view it, almost nobody can edit it (and remove e.g. footer or watermark, unlike in some simpler file types). That explains why it HAS to be PDF.
This program will be used by numerous users which will create new databases or expand existing ones. That's why having output formed as multple files is extremly sloppy and inefficient way of achieving what I want to achieve (it would complicate things for the user).
And what I want to do is to create PDF files which are still editable with my program once created.
I want to achieve this by implementing my custom file type readable with my program into the output PDF.
I came up with three ways of doing that:
Attach the file to PDF and then corrupting the part of PDF which contains it in a way it just makes the PDF unaware that it contains the file, thus making imposible for user to notice it (easely). Upon reading the document I'd revert the corruption and extract file using one of may PDF libraries.
Hide the file inside an image which would be added to the PDF somwhere on the first or last page, somehow (that is still need to work out) hidden from the public eye. Knowing it's location, it should be relativley easy to retrieve it using PDF library.
I have learned that if you add "%" sign as a first character in line inside a PDF, the whole line will be ignored (similar to "//" in Java) by the PDF reader (atleast Adobe reader), making possible for me to add as many lines as I want to the PDF (if I know where, and I do) whitout the end user being aware of that. I could implement my whole custom file into PDF that way. The problem here is that I actually have to read the PDF using one of the Java's input readers, but I'm not sure which one. I understand that PDF can't be read like a text file since it's a binary file (Right?).
In the end, I decided to go with the method number 3.
Unless someone has any better ideas, and the conditions are:
1. One file only. And that file is PDF.
2. User must not be aware of the addition.
The problem is that I don't know how to read the PDF as a file (I'm not trying to read it as a PDF, which I would do using a PDF library).
So, does anyone have a better idea?
If not, how do I read PDF as a FILE, so the output is array of characters (with newline detection), and then rewrite the whole file with my content addition?
In Java, there is no real difference between text and binary files, you can read them both as an inputstream. The difference is that for binary files, you can't really create a Reader for it, because that assumes there's a way to convert the byte stream to unicode characters, and that won't work for PDF files.
So in your case, you'd need to read the files in byte buffers and possibly loop over them to scan for bytes representing the '%' and end-of-line character in PDF.
A better way is to use another existing way of encoding data in a PDF: XMP tags. This is allows any sort of complex Key-Value pairs to be encoded in XML and embedded in PDF's, JPEGs etc. See http://partners.adobe.com/public/developer/en/xmp/sdk/XMPspecification.pdf.
There's an open source library in Java that allows you to manipulate that: http://pdfbox.apache.org/userguide/metadata.html. See also a related question from another guy who succeeded in it: custom schema to XMP metadata or http://plindenbaum.blogspot.co.uk/2010/07/pdfbox-insertextract-metadata-frominto.html
It's all just 1's and 0's - just use RandomAccessFile and start reading. The PDF specification defines what a valid newline character(s) is/are (there are several). Grab a hex editor and open a PDF and you can at least start getting a feel for things. Be careful of where you insert your lines though - you'll need to add them towards the end of the file where they won't screw up the xref table offsets to the obj entries.
Here's a related question that may be of interest: PDF parsing file trailer
I would suggest putting your comment immediately before the startxref line. If you put it anywhere else, you could wind up shifting things around and breaking the xref table pointers.
So a simple algorithm for inserting your special comment will be:
Go to the end of the file
Search backwards for startxref
Insert your special comment immediately before startxref - be sure to insert a newline character at the end of your special comment
Save the PDF
You can (and should) do this manually in a hex editor.
Really important: are your users going to be saving changes to these files? i.e. if they fill in the form field, are they going to hit save? If they are, your comment lines may be removed during the save (and different versions of different PDF viewers could behave differently in this regard).
XMP tags are the correct way to do what you are trying to do - you can embed entire XML segments, and I think you'd be hard pressed to come up with a data structure that couldn't be expressed as XML.
I personally recommend using iText for this, but I'm biased (I'm one of the devs). The iText In Action book has an excellent chapter on embedding XMP data into PDFs. Here's some sample code from the book (which I definitely recommend): http://itextpdf.com/examples/iia.php?id=217
I have created a program that should one day become a PDF editor
It's purpose will be saving GUI's textual content to the PDF, and loading it from it. GUI resembles text editor, but it only has certain fields(JTextAreas, actually).
It can look like this (this is only one page, it can have many more, also upper and lower margins are cut out of the picture) It should actually resemble A4 in pixel size.
I have looked around for a bit for PDF libraries and found out that iText could suit my PDF creating needs, however, if I understood it correct, it retirevs text from a whole page as a string which won't work for me, because I will need to detect diferent fields/paragaphs/orsomething to be able to load them back into the program.
Now, I'm a bit lazy, but I don't want to spend hours going trough numerus PDF libraries just to find out that they won't work for me.
Instead, I'm asking someone with a bit more Java PDF handling experience to recommend me one according to my needs.
Or maybe recommend me how to add invisible parts to PDF which will help my program to determine where is it exactly situated insied a PDF file...
Just to be clear (I formed my question wrong before), only thing I need to put in my PDF is text, and that's all I need to later be able to get out. My program should be able to read PDF's which he created himself...
Also, because of the designated use of files created with this program, they need to be in the PDF format.
Short Answer: Use an intermediate format like JSON or XML.
Long Answer: You're using PDF's in a manner that they wasn't designed for. PDF's were not designed to store data; they were designed to present and format data in an portable form. Furthermore, a PDF is a very "heavy" way to store data. I suggest storing your data in another manner, perhaps in a format like JSON or XML.
The advantage now is that you are not tied to a specific output-format like PDF. This can come in handy later on if you decide that you want to export your data into another format (like a Word document, or an image) because you now have a common representation.
I found this link and another link that provides examples that show you how to store and read back metadata in your PDF. This might be what you're looking for, but again, I don't recommend it.
If you really insist on using PDF to store data, I suggest that you store the actual data in either XML or RDF and then attach that to the PDF file when you generate it. Then you can read the XML back for the data.
Assuming that your application will only consume PDF files generated by the same application, there is one part of the PDF specification called Marked Content, that was introduced precisely for this purpose. Using Marked Content you can specify the structure of the text in your document (chapter, paragraph, etc).
Read Chapter 14 - Document Interchange of the PDF Reference Document for more details.
In a current project i need to display PDFs in a webpage. Right now we are embedding them with the Adobe PDF Reader but i would rather have something more elegant (the reader does not integrate well, it can not be overlaid with transparent regions, ...).
I envision something close google documents, where they display PDFs as image but also allow text to be selected and copied out of the PDF (an requirement we have).
Does anybody know how they do this? Or of any library we could use to obtain a comparable result?
I know we could split the PDFs into images on server side, but this would not allow for the selection of text ...
Thanks in advance for any help
PS: Java based project, using wicket.
I have some suggestions, but it'll be definitely hard to implement this stuff. Good luck!
First approach:
First, use a library like pdf-renderer (https://pdf-renderer.dev.java.net/) to convert the PDF into an image. Store these images on your server or use a caching-technique. Converting PDF into an image is not hard.
Then, use the Type Select JavaScript library (http://www.typeselect.org/) to overlay textual data over your text. This text is selectable, while the real text is still in the original image. To get the original text, see the next approach, or do it yourself, see the conclusion.
The original text then must be overlaid on the image, which is a pain.
Second approach:
The PDF specifications allow textual information to be linked to a Font. Most documents use a subset of Type-3 or Type-1 fonts which (often) use a standard character set (I thought it was Unicode, but not sure). If your PDF document does not contain a standard character set, (i.e. it has defined it's own) it's impossible to know what characters are which glyphs (symbols) and thus are you unable to convert to a textual representation.
Read the PDF document, read the graphics-objects, parse the instructions (use the PDF specification for more insight in this process) for rendering text, converting them to HTML. The HTML conversion can select appropriate tags (like <H1> and <p>, but also <b> and <i>) based on the parameters of the fonts (their names and attributes) used and the instructions (letter spacing, line spacing, size, face) in the graphics-objects.
You can use the pdf-renderer library for reading and parsing the PDF files and then code a HTML translator yourself. This is not easy, and it does not cover all cases of PDF documents.
In this approach you will lose the original look of the document. There are some PDF generation libraries which do not use the Adobe Font techniques. This also is a problem with the first approach, even you can see it you can not select it (but equal behavior with the official Adobe Reader, thus not a big deal you'd might say).
Conclusion:
You can choose the first approach, the second approach or both.
I wouldn't go in the direction of Optical Character Recognition (OCR) since it's really overkill in such a problem, since it also has several drawbacks. This approach is Google using. If there are characters which are unrecognized, a human being does the processing.
If you are into the human-processing thing; you can only use the Type Select library and PDF to Image conversion and do the OCR yourself, which is probably the easiest (human as a machine = intelligently cheap, lol) way to solve the problem.
I have built a web application that can be seen as an overcomplicated application form. There are bunch of text areas with a given character limit. After the form submission various things happen and one of them is PDF generation.
The text is queried from the DB and inserted in the PDF template created in iReports. This works fine but the major pain is overflowing text.
The maximum number of characters is set based on 'average' text. But sometimes people prefer to write with CAPS or add plenty of linefeeds to format their text. These then cause user's text to overflow the space given in PDF. Unfortunately the PDF document must look like a real application form so I cannot allow unlimited space.
What kind of approaches you have used to tackle this?
Clean/restrict user input?
Calculate the space requirement of the text based on font metrics?
Provide preview of the PDF? (too bad users are not allowed to change their input after submission...)
Ideally, calculate the requirement based on metrics. I don't know how iReports handles text, but with iText, it lays everything out itself, you just present the data as a streaming document, so we don't worry about overflowing text.
However, iReport may not support that, or you may need to have the PDF layout fit within certain bounds. I'd try to clean the input (ie: if it's all caps, lowercase/sentence case/proper case it), strip extra whitespace. If cleaning the input can't be reliably done, or people are still getting past that, I'd also restrict it.
As a last resort, I'd present the PDF for the user to authorize. Really, users shouldn't be given more work to do, and they're not going to do it anyways.
Your own suggested solutions to your problem are all good. Probably the most important question to have answered is what should your PDF look like when the data to be displayed in a field won't fit? Do you ever need the "full answer" for anything else? When you know the answer to these, you'll have your options reduced.
For example if a field must be limited to 1/2 a page, and users sometimes enter more than 1/2 a page of text you can either
1) limit the user input - on submission calculate the size (using font-metrics as you said) and reject the submission until corrected. This assumes you can legitimately force the user to reduce their data entry.
2) accept the user input and truncate in the display of this report. Some systems use "..." to indicate data has been truncated, and can provide a hyperlink (even within the PDF) to get more information.
Providing a preview would work really well, but only if the users are good at checking and correcting and your system can handle the extra load this will generate.
Do you have control of the font that is used when generating the PDF? If so, I would look for a font in the Monospace family. This will give you consistent length for a given number of chars, regardless of puncuation, capitalization, etc.