（评论）

（评论）
(comments)

原始链接: https://news.ycombinator.com/item?id=41377960

建议使用 MUTool 以轻松理解 PDF 文件格式。要从压缩的 PDF 中提取内容，请使用 MUTool 的命令“mutool clean -d in.pdf out.pdf”。 PDF 文件主要由一系列包含数据和指令的互连对象组成。一个对象代表一个页面，并包含诸如“MediaBox”中的大小之类的信息，以及指向存储在 PDF 内另一位置的实际页面内容的指针。了解这些基本结构可以让您导航和分析复杂的 PDF 文件。有关 PDF 标准的详细探索，请参阅此处的 Adobe Document Cloud 开发人员指南：[PDF 标准](https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandard)。值得注意的是，与 iText 相关的某些工具和组件根据 Affero 通用公共许可证 (AGPL) 许可其工作，该许可证对商业使用施加限制，并要求分发所有相关源代码。实施 iText 时要小心，以免侵犯他们对许可的积极态度。或者，考虑使用替代的开源或专有选项。要从 PDF 文件中提取图像，请执行以下步骤： * 使用PDFTK等工具解压PDF文件 * 在提取的数据中查找图像插入代码 * 用相同长度的空格替换图像插入代码的每个实例，以保持生成的 PDF 的结构完整性 * 使用 PDFTK 将更改后的 PDF 压缩回其原始状态此过程将允许您删除 PDF 文件中存在的特定元素，为各种目的提供实际应用。此外，参考 pdfa.org 提供的“PDF 规范档案”可以深入了解给定 PDF 中使用的图像插入技术。

Mutool is the one I suggest to people. The easiest way to understand a PDF is to decompress it and then just read the contents.

    mutool clean -d in.pdf out.pdf

At that point you’ll realise that a PDF is mostly just a list of objects and that those objects can reference each other. After that you’ll journey through the spec understanding what each type of object does and what the fields in it control. The graphics stream itself is just a stack based co-ordinates drawing system that’s easy to follow too.

By way of an example. Here's an object that represents a Page. You can see the dimensions in the MediaBox. The contents themselves are contained at object "9 0 obj" ("9 0 R" is the pointer to it):

    2 0 obj
    <<
      /Type /Page
      /MediaBox [ 0 0 612 792 ]
      /Contents 9 0 R
    >>
    endobj

Meanwhile "9 0 obj" has the drawing instructions. They seem a little weird at first glance but you see the values ".23999999 0 0 -.23999999 0 792" each get pushed on the stack and then "cm" pops them to interpret them as the transformation matrix.

    9 0 obj
    <<
      /Length 18266
    >>
    stream
    .23999999 0 0 -.23999999 0 792 cm
    q
    0 0 2551 3301 re
    ...

The depth and detail of all of the different possible things that can be represented in a PDF is insane. But understanding the structure above is all you need to begin your journey!

EDIT The rest of your journey is contained in this epic document: https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandard...

> mutool clean -d in.pdf out. pdf

My tool can do exactly the same (viewing internal structure, exporting objects, and see the uncompressed raw content for stream) with a graphical interface and without all this kind of flags (which one of the reasons I started to design this project with egui), but thanks for posting yours too.

Thanks for the list, the idea behind my tool was to try to code something that might fit an analyst that would take a fast look at the PDF. I'm also trying to figure out some fast heuristics to mark/highlight some peculiar stuff on the file itself.

Now regarding the tools you mentioned, I haven't checked out all of them, but part of them are interesting (and more mature, speaking of testing and compatibility). However some (at least the ones I was trying) are very basic, and they don't allow the "Save object as.." or uncompress it. I like the feature of displaying the PDF for preview :)

yeah I agree, and while everyone is suggesting tools which are really good but I designed mine to get rid of the flags and CLI interface. Good for tech people that keeps remembering flags, I'm not :(

Do be careful with iText products—they license their stuff under AGPL but their interpretation of AGPL is pretty extreme. If you talk to their team they'll tell you that ~everything your company makes should be AGPL-licensed if you use iText anywhere [0]:

> You may not deploy it on a network without disclosing the full source code of your own applications under the AGPL license. You must distribute all source code, including your own product and web-based applications.

They also have this delightful nagware encoded as a base64 string that spits this out in your logs [1]:

> You are using iText under the AGPL.

> If this is your intention, you have published your own source code as AGPL software too. Please let us know where to find your source code by sending a mail to [email protected] We'd be honored to add it to our list of AGPL projects built on top of iText and we'll explain how to remove this message from your error logs.

> If this wasn't your intention, you are probably using iText in a non-free environment. In this case, please contact us by filling out this form: http://itextpdf.com/sales If you are a customer, we'll explain how to install your license key to avoid this message. If you're not a customer, we'll explain the benefits of becoming a customer.

For using RUPS on a local computer you're probably safe, but I avoid the company because everything about their approach to the AGPL suggests that they chose it as a marketing technique for their paid products (with an extremely strong desire that it never be used commercially without pay), not out of a serious commitment to free software.

[0] https://itextpdf.com/how-buy/AGPLv3-license

[1] https://github.com/itext/itext-dotnet/blob/develop/itext/ite...

I use it literally everyday, not only to see the structure but also modify pdf's on the fly when I need to tests edge cases.

Stuff I do with it: Modify content streams, extract images/content, just investigate general structure of the pdf documents, remove pages, repair documents,... it's literally a swiss army knife when working with pdf's

Nice! My tool should be runnable in the browser thanks to wasm compatibility with Rust + egui :) Btw I've just tried it, and it's a little bit buggy in Safari with a 504kb PDF (lots of objects though). Apart from that, is there a way to export the raw stream? Is there any reason of do you print all the raw streams as a text?

I don’t remember much about the work - it was just a quick and dirty app to help me debug PDF for my ReportMill work (10 years ago). I remember thinking there probably weren’t more than 100 people on the planet who would even care about it.

That demo page told me everything I needed to know about not using that framework. There's no right-click to copy the text, and there's similarly no context menu to open links in a new tab, or copy their URL

Yeah, I knew I was in for some onoz when I saw "compiled to WebAssembly and rendered with WebGL". In their defense, it's stunning that any text operations work at all

Also, "There is no DOM, HTML, JS or CSS" is some uh-huh given the considerable amount of silliness involved in view-source:https://www.egui.rs/

Is there actually a open source good alternative of Adobe Acrobat? I mean one where you can easily 1) merge and extract pages 2) edit content, e.g. change a single line in a simple vector graphics, add text ornother elements 3) place a signature and create a "protected" pdf. AFAIK, there isn't a single good open source program to do all of these things. Why?

:D Well, I'm sure that half of reverse engineering community needs to thank you, and Zynamics for the important contribution for tools of static analysis. I just take the occasion to thank you for being an inspiration with such awesome tools like in BinNavi, BinDiff, and ultimately PDF dissector. When I was reading that it got discontinued, I just had that idea and started to reason about something focused on analysis, and applying some approaches we've already seen for the binary analysis tools.

Does anyone have any recommendations for a good tool that allows both programmatic inspection and modification of PDF primitives. For example, let's say someone wants to iterate through every embedded image in a PDF and apply some form of signal processing to the images in-place, then re-save the PDF?

My tool (PDFSyntax[1], mentioned in this thread) is a Python library that is able to both inspect and transform PDF files.

Depending on your transformation use case, you may write an incremental update with only a few bytes at the end of the original file instead of rewriting it entirely. To my knowledge this feature of the PDF specification is often overlooked and not a lot of libraries implements it.

It is a work in progress and I have not developed functions for images yet, though.

[1] https://github.com/desgeeko/pdfsyntax

I've been using pypdf for working with PDFs in Python. My uses are pretty humble. I create Jupyter notebooks for managing sheet music that I receive in PDF format, allowing me to do things like break up a book of tunes into individual files, and so forth. This in turns makes it easier to pull up individual tunes on my tablet during a performance. But it looks like you can treat the PDF as a tree structure. I've used that feature for writing some recursive functions.

yeah, I've been using pypdf mainly, camelot-py for some table stuff, and a bit of pdfminer

I've been needing something to see the x/y bounds of tables to fix some edge cases with camelot, seem to be some good links in the comments here

I was curious to try this out as it might actually solve a minor problem of mine right now, but it crashed as soon as I tried to open a PDF.

Installed from git using cargo 1.80.1 on Ubuntu 22.04 on an AMD Framework laptop if that's of any help.

argh, that's too bad, feel free to open an issue, what's happening in the console? It's panicking, isn't it? Feel free to contact me via email if you prefer

I recently wanted to edit out a huge background image repeating on almost every page of a PDF and found out there's no obvious way to do it.

Would appreciate any tool suggestions!

This is probably a simple find-and-replace task, so I wouldn't bother with proper PDF parsing or libraries. I would:

1. Use pdftk to uncompress it: pdftk input.pdf output uncompressed.pdf uncompress

2. Look at the PDF code (it's text based) to find the image insertion code.

3. Replace all instances of the image insertion code with strings of spaces the same length (there's a table of object byte offsets at the end that you don't want to mess up).

4. Use pdftk to compress it again: pdftk edited.pdf output output.pdf compress

I have a script that does this to remove pen strokes of particular colours so I can e.g. strip out marking rubric on test solutions written on a tablet.

Get the PDF 1.7 spec from https://pdfa.org/resource/pdf-specification-archive/. You're looking for the "Do" operator invoking a named image object defined elsewhere with "/Subtype /Image". See section 4.8, particularly the example on p343. Or, if it's badly done, it might instead be an inline image using the "BI" operator (a bit later in the same section).

what's a good tool to check if a pdf is not tampered with eg. as a tool to check before loading a pdf from a public bucket to your backend application?

They trust their AWS EC2 instance doing the processing but not their AWS S3 bucket doing the storing? I don't really understand the threat model here.

And what's "public bucket"?

So the model here is, first it gets uploaded to a staging bucket, a lambda/callback checks the validity of the file and then puts it into a safe bucket of which content I trust to put in my server(backend)

（评论） (comments)

（评论）
(comments)