Reading order after flatten method call

[Foohx]

Hello everybody,

I have a template of a tagged PDF file, which has a reading order defined to be properly read by screen readers (Voice Over, NVDA, JAWS ..). The values I want to insert into the model are represented by inputs.

My test template : input.pdf

So I get the fields, add a value to them then I flatten the document before saving it.

import { readFileSync, writeFileSync } from "fs";
import { PDFDocument } from "pdf-lib";

const fileBytes = readFileSync("./input.pdf");
const pdf = await PDFDocument.load(fileBytes);

const form = pdf.getForm();
const fields = form.getFields();
fields.forEach((field) => {
  const type = field.constructor.name;

  if (type === "PDFTextField") {
    field.setText("Test");
  } else if (type === "PDFDropdown") {
    const options = field.getOptions();
    field.select(options[0]);
  }
});

form.flatten({ updateFieldAppearances: true });

const outputBytes = await pdf.save();
writeFileSync("./output.pdf", outputBytes);

The problem is with the output.pdf file.
When I try to read it through a screen reader, my flattened values are present instead of the template inputs, however they are unreadable by screen readers.

It's as if they don't exist..
Do I have to tag/do something or use another solution?

Thank you in advance.

[Hopding]

pdf-lib doesn't provide special support for tagged PDFs. This is something I'd like to improve though, so I've added it to the roadmap for tracking: #998.

[pablodenadai]

Hi @Foohx did you find a solution to this?

[pablodenadai]

@Hopding would you be able to point us in the right direction if we were to create a PR for this issue?

[Foohx]

@pablodenadai not really, but I got around the problem.

I decided to no longer use forms directly within PDFs. I am using a web form and generating the final PDF accordingly.
This way, my reading order is always correct and I can validate all the datas before writing them in the document.

EDIT: I also use another library (more low level) allowing a finer control over the accessibility.

[LucianoGanga]

@Foohx , what's the library you used to get finer control over accessibility? I need to access and manipulate that part of the PDFs but I can't find a way to do it.

Thanks!

[Foohx]

@Foohx , what's the library you used to get finer control over accessibility? I need to access and manipulate that part of the PDFs but I can't find a way to do it.

Thanks!

Hi,

I use the pdfkit lib.