Update Widget text

[Mominadar]

Description of the bug

Hi, I'm trying to process PDFs to translate them. Now, working with widgets, I am facing the issue of not being able to update the text correctly. I get that I can use the button_caption field to update text, but I need the original text as well to translate it. However, when I see the original button caption, it's None.

This is a sample dict representation of the widget:

 {'border_color': None, 'border_style': 'Solid', 'border_width': 1, 'border_dashes': None, 'choice_values': None, 'rb_parent': None, 'field_name': 'TOC.Page 2', 'field_label': None, 'field_value': '', 'field_flags': 65536, 'field_display': 0, 'field_type': 1, 'field_type_string': 'Button', 'fill_color': None, 'button_caption': None, 'is_signed': None, 'text_color': (0, 0, 0), 'text_font': 'Helv', 'text_fontsize': 0, 'text_maxlen': 0, 'text_format': 0, '_text_da': '', 'script': None, 'script_stroke': None, 'script_format': None, 'script_change': None, 'script_calc': None, 'script_blur': None, 'script_focus': None, 'rect': Rect(778.302001953125, 6.64398193359375, 793.7509765625, 29.593994140625), 'xref': 654, 'parent': <weakproxy at 0x741d3acba2a0 to Page at 0x741d3ac47580>, '_annot': 'Widget' annotation on page 1 of <memory, doc# 1>}
heree None {'border_color': None, 'border_style': 'Solid', 'border_width': 1, 'border_dashes': None, 'choice_values': None, 'rb_parent': None, 'field_name': 'References.Page 2', 'field_label': None, 'field_value': '', 'field_flags': 65536, 'field_display': 0, 'field_type': 1, 'field_type_string': 'Button', 'fill_color': None, 'button_caption': None, 'is_signed': None, 'text_color': (0, 0, 0), 'text_font': 'Helv', 'text_fontsize': 0, 'text_maxlen': 0, 'text_format': 0, '_text_da': '', 'script': None, 'script_stroke': None, 'script_format': None, 'script_change': None, 'script_calc': None, 'script_blur': None, 'script_focus': None, 'rect': Rect(808.3410034179688, 6.957000732421875, 821.7559814453125, 29.593994140625), 'xref': 653, 'parent': <weakproxy at 0x741d3acba2a0 to Page at 0x741d3ac47580>, '_annot': 'Widget' annotation on page 1 of <memory, doc# 1>}

The field_name is what I see on the PDF, i.e., TOC in this case, this is true for all the widgets on the PDF. My question is, is that the norm? Can I generalize this for all PDFs to get the button text, and get the field name minus the page number? Or is there more to it? Appreciate any guidance that can help cover all bases.

How to reproduce the bug

You can try this pdf: superstar-messaging.pdf

Code:

 def extract_text_blocks_1(self, doc):
        pages_blocks = []
        for page_num in range(len(doc)):
            page = doc.load_page(page_num)
            blocks = page.get_text("dict", sort=True, clip=pymupdf.INFINITE_RECT())
            blocks = blocks["blocks"]
            page_blocks = []
            for block in blocks:
                if "lines" in block:
                    for line in block["lines"]:
                        for span in line["spans"]:
                            # logger.info(f"properties {span}, flag, {span}")
                            font_properties = {"size": span["size"], "color": "#%06x" % span["color"], "font-family": '%s' % span["font"], "font-weight": "bold" if self.flags_decomposer(span['flags']) else "normal", "font-style": "italic" if self.flags_decomposer(span['flags']) else "normal"}
                            if span["text"].strip():
                                bbox = span["bbox"]
                                if isinstance(bbox, (list, tuple)) and len(bbox) == 4:
                                    bbox = [float(coord) for coord in bbox]
                                    if span["text"]:
                                        page_blocks.append({
                                            "text": span["text"],
                                            "bbox": bbox,
                                            "font_properties": font_properties,
                                            "angle": self.get_angle(bbox, line["dir"])
                                        })
            pages_blocks.append(page_blocks)
        return pages_blocks   
    

doc = pymupdf.open(stream=file, filetype="pdf")
pages_blocks = self.extract_text_blocks_1(doc)
        
count_rc = 0
logger.info(f"{len(pages_blocks)} Pages found")
for page_num, page_blocks in enumerate(pages_blocks[:2]):
    page = doc.load_page(page_num)
    logger.info(f"Page no: {page_num}")
    for widget in page.widgets():
         if widget.field_type == pymupdf.PDF_WIDGET_TYPE_BUTTON:
               print("heree", widget.button_caption, widget.__dict__)
              text = ""  #how to get text?
               if any(c.isalpha() for c in text):
                   translated_text = translate_file_text_function(text) #how to get text?
                   widget.button_caption = translated_text
                    widget.update()

PyMuPDF version

1.25.4

Operating system

Linux

Python version

3.10

[JorjMcKie]

This is not an issue. Transferring to "Discussions" tab.

[Mominadar]

Alright, any help would be appreciated here, guys. Thanks in advance!

[JorjMcKie]

Field names are mandatory. They also must be unique in the PDF. Uniqueness understood as / applied to the concatenation of field hierarchy levels (e.g. like "customer-number.adress.street").
Many other field attributes are optional, e.g. button_caption and field_label.
Especially field_label is often used for displaying help text when the cursor hovers over a field. This seems to be what you are interested in.
Other text - outside fields - on the page is often present to the left or right of buttons, too. Probably of interest as well, but there is no technical connection to the widget they are referring to. This means, that you must find an appropriate algorithm to find this connection / relationship. Best candidate information obviously being the geometrical position of such text ...