Polish letters in form fields cause text misalignment

[Pshemas]

I'm trying to use pymupdf to fill a pdf form . Initially I thought I got it... with sample data form text fields were filling nicely.

But then I've introduced polish letters into sample data (like ąęśćżź) - and this caused the filled data to be rendered outside text field (below it) - occluded / invisible (wouldn't noticed those if I haven't spotted the very tops of the letters) until I've clicked on it.

Is there a way to make it work?

The form itself has been created in OnlyOffice.

Edit: also I've noticed that if I opened the file filled with my pymupdf script the font used in the fields gets reset to Nimbus or Helvetica. In the fields where are polish letters it is Numbus.

I've added pymupdf-fonts package and tried forcing other fonts - but with no luck. I'm clearly doing something wrong.

Here's the function I'm using to fill the form:

def fill_pdf( data: dict, doc: pymupdf.Document, filename: str, target_folder: Path = Path.cwd(), ): for page in doc: for field in page.widgets(types=(pymupdf.PDF_WIDGET_TYPE_TEXT,)): if field.field_name in data: field.field_value = data[field.field_name] field.text_font = "notos" field.update() output_path = target_folder / filename if target_folder in output_path.parents: doc.save(output_path) else: raise FilepathError("Path outside target path.")

[JorjMcKie]

Please consult the documentation:
We currently only support Base-14 fonts in form fields. This implies that any characters with Unicode values outside range(256) cannot be entered, respectively are not being correctly handled.
All your example characters have Unicode values > 256 ...

[Pshemas]

thx @JorjMcKie . That's unfortunate - from the initial tests of various pdf python libs I've like pymupdf the most. But this makes it sadly unusable outside english speaking (or rather writing ;) ) countries as there's no way to get the characters right.

Hoped there's some sort of workaround where I can use some other tool to fix the font and encoding after filling the form using pymupdf. Ok, so getting back to searching and testing.

[JorjMcKie]

Actually, there does exist a circumvention which I am planning to make part of the official package at some time.
The main part is helper function modify_textwidget. The basic approach is:

1. Define / create some text form field / widget
2. Call modify_textwidget to modify the widget assigning a desired font to it.

After this, the text widget should support all characters defined in the provided font - including for example Polish / extended Latin etc.
Together with the utility function, I have also included a few examples that show the use of special text widgets like multi-line, combs etc.

The chosen example font ("Liberation ...") is of no significance and could be replaced by whatever you find useful.

Please note that there is no guarantee whatsoever and that the work overall remains to be licensed as PyMuPDF itself (GNU AGPL 3.0 or commercial).

modify-textwidget.zip