How to erase text in a specific area

[xiaolibuzai-ovo]

I want to extract text, translate it, and then paste it back, but I don't want to cover it with any color. So I would like to ask if there is an API to erase text in a specific area, or to get the background color for filling.

thanks for you help!

[JorjMcKie]

Best use redaction annotations:

page.add_redact_annot(rect)
page.apply_redactions()
page.insert_htmlbox(rect,...)

[xiaolibuzai-ovo]

This will lose a lot of background elements, such as borders, etc. I only want to delete the area where the text is located.

[xiaolibuzai-ovo]

The left side is the generated demo.

code：

docs = pymupdf.open("Nuxtjs-Cheat-Sheet.pdf")  
for page in docs:
    blocks=page.get_text("dict",flags=0,sort=False)["blocks"]
    spans=[s for b in blocks for l in b['lines'] for s in l['spans']]
    spanss.append(spans)
    for span in spans:
        bbox = span['bbox']
        text = span['text']
        color = pymupdf.sRGB_to_rgb(span["color"])
        WHITE = pymupdf.pdfcolor["white"]
        page.add_redact_annot(bbox)
        page.apply_redactions()
        text = text.replace('<',"《")
        text = text.replace('>',"》")
        color_str = f'rgb({color[0]}, {color[1]}, {color[2]})'

        target = ''
        for t in text:
            target+='a'
        html = f'<span style="color: {color_str};">{target}</span>'
        page.insert_htmlbox(
            bbox, html
        )
docs.ez_save("Nuxtjs-Cheat-Sheet.pdf-new.pdf")

[JorjMcKie]

Have you ever looked at the documentation of page.apply_rdactions()? There are options to remove only text and leave images and vector graphics untouched.
The default removes everything under the specified rectangle: images, vector graphics and text.

[xiaolibuzai-ovo]

Yes, I added page.apply_redactions(image=0,graphics=0), but it still covers everything.

[xiaolibuzai-ovo]

the file is:
Nuxtjs-Cheat-Sheet.pdf

[JorjMcKie]

Maybe you can use some of this as idea for improvements:

import pymupdf

doc = pymupdf.open("test.pdf")
page = doc[0]
text = "replaced"
words = page.get_text("words")
for w in words:
    page.add_redact_annot(w[:4])
page.apply_redactions(images=0, graphics=0, text=0)

# walk through rectangles again and insert replacement text
for w in words:
    page.insert_htmlbox(w[:4], text)
doc.subset_fonts()  # subset the used font
# save using maximum compression
doc.ez_save("x.pdf", garbage=4)

Output PDF:
x.pdf

[xiaolibuzai-ovo]

Sorry, I'm a beginner with PyMuPDF. What does this mean, and how is it different from my code? I noticed you used get_text("word").

[JorjMcKie]

Ok, look at this then:

import pymupdf

doc = pymupdf.open("test.pdf")
page = doc[0]
blocks = page.get_text("dict", flags=pymupdf.TEXTFLAGS_TEXT)["blocks"]
spans = [s for b in blocks for l in b["lines"] for s in l["spans"]]
for s in spans:
    page.add_redact_annot(s["bbox"])
page.apply_redactions(images=0, graphics=0, text=0)

for s in spans:
    text = "".join([("x" if c != " " else " ") for c in s["text"]])
    color = "%02x%02x%02x" % pymupdf.sRGB_to_rgb(s["color"])
    html = f'<span style="color: {color};font-size:{s["size"]}px;">{text}</span>'
    page.insert_htmlbox(s["bbox"], html)
page.clean_contents()
doc.subset_fonts()  # subset the used font
# save using maximum compression
doc.ez_save("x.pdf", garbage=4)

x.pdf

[xiaolibuzai-ovo]

oh！thanks for you help!

[JorjMcKie]

I forgot the '#' in front of color: "#{color}"

[xiaolibuzai-ovo]

sorry！ I allowed the same code, but got different results.

import pymupdf

doc = pymupdf.open("Nuxtjs-Cheat-Sheet.pdf")
page = doc[0]
blocks = page.get_text("dict", flags=pymupdf.TEXTFLAGS_TEXT)["blocks"]
spans = [s for b in blocks for l in b["lines"] for s in l["spans"]]
for s in spans:
    page.add_redact_annot(s["bbox"])
page.apply_redactions(images=0, graphics=0, text=0)

for s in spans:
    text = "".join([("x" if c != " " else " ") for c in s["text"]])
    color = "%02x%02x%02x" % pymupdf.sRGB_to_rgb(s["color"])
    html = f'<span style="color: #{color};font-size:{s["size"]}px;">{text}</span>'
    page.insert_htmlbox(s["bbox"], html)
page.clean_contents()
doc.subset_fonts()  # subset the used font
# save using maximum compression
doc.ez_save("Nuxtjs-Cheat-Sheet.pdf-new.pdf", garbage=4)

Nuxtjs-Cheat-Sheet.pdf-new.pdf
ew.pdf