I'm writing a python script to change color of text and borders of inline annotations inserted with okular into a pdf document.
This script instead of changing only text and border seems change the entire annotation red (full filled red rectangle shows up instead).
Is this supported at all? I realize how complex pdf format is, but I hoped that this simple change won't require anything unusual.
sample inline text annotation in pdf as produced by okular:
11533 │ 229·13·obj␍␊
11534 │ <</Length·293·/Subtype·/Form·/BBox·[0·0·232.7894972067·28.173484252·]·/Resources·<</Font·<</popplerfont·227·0·R·>>·>>·>>·stream␍␊
11535 │ q␊
11536 │ []·0·d␊
11537 │ 2.00·w␊
11538 │ 0.00000·0.00000·0.00000·RG␊
11539 │ 1.00·1.00·230.79·26.17·re␊
11540 │ 0.21569·0.18824·1.00000·rg␊
11541 │ B␊
11542 │ 4.00·4.00·224.79·20.17·re·W·n␊
11543 │ 0.00000·0.00000·0.00000·rg␊
11544 │ BT·1·0·0·1·4.00·24.17·Tm␊
11545 │ /popplerfont·10.00·Tf␊
11546 │ 0.00·-10.00·Td␊
11547 │ (\000e\000r\000g\000r\000g\000h\000r\000e\000h\000r\000h\000e\000h)·Tj␊
11548 │ ET·Q␊
11549 │ ␍␊
11550 │ endstream␍␊
11551 │ ␍␊
11552 │ endobj␍␊
11553 │ 230·0·obj␍␊
11554 │ <</Type·/Annot·/Rect·[505.4857653631·525.0882627953·507.5172614261·534.4675688976·]·/Subtype·/FreeText·/DA·(/popplerfont·10.00·Tf\n)·/M·(D:20241117102025Z␀································)·/T·<feff00750073006500720031>·/Contents·<feff>·/NM·(okular-{a598c121-968e-4134-8fa3-b6d17caa86f2})·/CreationDate·(D:20241117102025Z␀································)·/F·20·/C·[0.2156862766·0.1882352978·1·]·/CA·1·/Border·[0·0·2·]·/Q·0·/IT·/FreeText·/P·2·0·R·/AP·<</N·231·0·R·>>·/AS·/N·>>·␍␊
11555 │ endobj␍␊
11556 │ 231·0·obj␍␊
11557 │ <</Length·222·/Subtype·/Form·/BBox·[0·0·2.031496063·9.3793061024·]·/Resources·<</Font·<</popplerfont·227·0·R·>>·>>·>>·stream␍␊
11558 │ q␊
11559 │ []·0·d␊
11560 │ 2.00·w␊
11561 │ 0.00000·0.00000·0.00000·RG␊
11562 │ 1.00·1.00·0.03·7.38·re␊
11563 │ 0.21569·0.18824·1.00000·rg␊
11564 │ B␊
11565 │ 4.00·4.00·-5.97·1.38·re·W·n␊
11566 │ 0.00000·0.00000·0.00000·rg␊
11567 │ BT·1·0·0·1·4.00·5.38·Tm␊
11568 │ /popplerfont·10.00·Tf␊
11569 │ 0.00·-10.00·Td␊
11570 │ ()·Tj␊
11571 │ ET·Q␊
11572 │ ␍␊
11573 │ endstream␍␊
11574 │ ␍␊
11575 │ endobj␍␊
11576 │ xref␍␊
and here converted to qpdf format:
40687 │ 0183013890·00000·n·␊
40688 │ 0185638157·00000·n·␊
40689 │ trailer·<<␊
40690 │ ··/Info·2·0·R␊
40691 │ ··/Root·1·0·R␊
40692 │ ··/Size·372␊
40693 │ ··/ID·[<fa440df8ed780aedfcf5266a1bfed4ad><fa440df8ed780aedfcf5266a1bfed4ad>]␊
40694 │ >>␊
40695 │ startxref␊
40696 │ 185638183␊
40697 │ %%EOF␊
40698 │ 1·0·obj␍␊
40699 │ <</Pages·3·0·R·/Type·/Catalog·/AcroForm·373·0·R·>>·␍␊
40700 │ endobj␍␊
40701 │ 4·0·obj␍␊
40702 │ <</Contents·68·0·R·/MediaBox·[0·0·595.276·841.89·]·/Parent·3·0·R·/Resources·<</Font·<</F4·70·0·R·/F5·71·0·R·/F6·72·0·R·>>·/ProcSet·73·0·R·/XObject·<</Im3·74·0·R·/Im7·76·0·R·>>·>>·/Type·/Page·/Annots·378·0·R·>>·␍␊
40703 │ endobj␍␊
40704 │ 372·0·obj␍␊
40705 │ <</Type·/Annot·/Rect·[142.1173581129·531.1166868381·347.6343759828·552.3443721634·]·/Subtype·/FreeText·/DA·(/popplerfont·10.00·Tf\n)·/M·(D:20241117100201Z␀································)·/T·<feff00750073006500720031>·/Contents·<feff00750069006500620072006700730075006b0067006200730073006b00650075>·/NM·(okular-{c67d430b-37dd-4ba6-85c6-45a45db4e512})·/CreationDate·(D:20241117100143Z␀································)·/F·20·/C·[0.2156862766·0.1882352978·1·]·/CA·1·/Border·[0·0·2·]·/Q·0·/IT·/FreeText·/P·4·0·R·/AP·<</N·379·16·R·>>·/AS·/N·>>·␍␊
40706 │ endobj␍␊
40707 │ 373·0·obj␍␊
40708 │ <</Fields·[]·/DR·<</Font·<</popplerfont·377·0·R·>>·>>·>>·␍␊
40709 │ endobj␍␊
40710 │ 374·0·obj␍␊
40711 │ <</Length·210957·/Filter·/FlateDecode·>>·stream␍␊
#!/usr/bin/env python3
import sys
import pikepdf
from pikepdf import Name
import re
def change_annotation_font_and_border_to_red(input_path):
output_path = input_path.replace(".pdf", "-output.pdf")
# Open the PDF file
with pikepdf.open(input_path) as pdf:
for page_number, page in enumerate(pdf.pages, start=1):
# Check if the page has annotations
if "/Annots" in page:
print(f"Page {page_number} contains annotations:")
# Iterate over each annotation in the /Annots array
for annot in page["/Annots"]:
print(f" - Annotation detected: {annot}")
# Set the border color to red (RGB)
annot[Name("/C")] = pikepdf.Array([1, 0, 0]) # RGB for red border
# Update the default appearance string (/DA) to set text color to red
if "/DA" in annot:
da_str = str(annot[Name("/DA")])
print(f" Original /DA: {da_str}")
# Remove any existing color settings in the DA string
da_str = re.sub(r"(\d+\.?\d*\s){1,3}[rg]", "", da_str)
# Append red color setting for text color
da_str = da_str.strip() + " 1 0 0 rg"
annot[Name("/DA")] = da_str
print(f" Updated /DA: {annot[Name('/DA')]}")
# Remove the appearance stream to apply changes if present
if "/AP" in annot:
del annot[Name("/AP")]
print(" Removed /AP to force appearance regeneration")
# Save the modified PDF with '-output.pdf' suffix
pdf.save(output_path)
print(f"Font and border color changed to red. Saved as: {output_path}")
if __name__ == "__main__":
if len(sys.argv) != 2:
print("Usage: python change_annotation_font_and_border_to_red.py <input-pdf-path>")
else:
input_path = sys.argv[1]
change_annotation_font_and_border_to_red(input_path)
I'm writing a python script to change color of text and borders of inline annotations inserted with okular into a pdf document.
This script instead of changing only text and border seems change the entire annotation red (full filled red rectangle shows up instead).
Is this supported at all? I realize how complex pdf format is, but I hoped that this simple change won't require anything unusual.
sample inline text annotation in pdf as produced by okular:
11533 │ 229·13·obj␍␊
11534 │ <</Length·293·/Subtype·/Form·/BBox·[0·0·232.7894972067·28.173484252·]·/Resources·<</Font·<</popplerfont·227·0·R·>>·>>·>>·stream␍␊
11535 │ q␊
11536 │ []·0·d␊
11537 │ 2.00·w␊
11538 │ 0.00000·0.00000·0.00000·RG␊
11539 │ 1.00·1.00·230.79·26.17·re␊
11540 │ 0.21569·0.18824·1.00000·rg␊
11541 │ B␊
11542 │ 4.00·4.00·224.79·20.17·re·W·n␊
11543 │ 0.00000·0.00000·0.00000·rg␊
11544 │ BT·1·0·0·1·4.00·24.17·Tm␊
11545 │ /popplerfont·10.00·Tf␊
11546 │ 0.00·-10.00·Td␊
11547 │ (\000e\000r\000g\000r\000g\000h\000r\000e\000h\000r\000h\000e\000h)·Tj␊
11548 │ ET·Q␊
11549 │ ␍␊
11550 │ endstream␍␊
11551 │ ␍␊
11552 │ endobj␍␊
11553 │ 230·0·obj␍␊
11554 │ <</Type·/Annot·/Rect·[505.4857653631·525.0882627953·507.5172614261·534.4675688976·]·/Subtype·/FreeText·/DA·(/popplerfont·10.00·Tf\n)·/M·(D:20241117102025Z␀································)·/T·<feff00750073006500720031>·/Contents·<feff>·/NM·(okular-{a598c121-968e-4134-8fa3-b6d17caa86f2})·/CreationDate·(D:20241117102025Z␀································)·/F·20·/C·[0.2156862766·0.1882352978·1·]·/CA·1·/Border·[0·0·2·]·/Q·0·/IT·/FreeText·/P·2·0·R·/AP·<</N·231·0·R·>>·/AS·/N·>>·␍␊
11555 │ endobj␍␊
11556 │ 231·0·obj␍␊
11557 │ <</Length·222·/Subtype·/Form·/BBox·[0·0·2.031496063·9.3793061024·]·/Resources·<</Font·<</popplerfont·227·0·R·>>·>>·>>·stream␍␊
11558 │ q␊
11559 │ []·0·d␊
11560 │ 2.00·w␊
11561 │ 0.00000·0.00000·0.00000·RG␊
11562 │ 1.00·1.00·0.03·7.38·re␊
11563 │ 0.21569·0.18824·1.00000·rg␊
11564 │ B␊
11565 │ 4.00·4.00·-5.97·1.38·re·W·n␊
11566 │ 0.00000·0.00000·0.00000·rg␊
11567 │ BT·1·0·0·1·4.00·5.38·Tm␊
11568 │ /popplerfont·10.00·Tf␊
11569 │ 0.00·-10.00·Td␊
11570 │ ()·Tj␊
11571 │ ET·Q␊
11572 │ ␍␊
11573 │ endstream␍␊
11574 │ ␍␊
11575 │ endobj␍␊
11576 │ xref␍␊
and here converted to qpdf format:
40687 │ 0183013890·00000·n·␊
40688 │ 0185638157·00000·n·␊
40689 │ trailer·<<␊
40690 │ ··/Info·2·0·R␊
40691 │ ··/Root·1·0·R␊
40692 │ ··/Size·372␊
40693 │ ··/ID·[<fa440df8ed780aedfcf5266a1bfed4ad><fa440df8ed780aedfcf5266a1bfed4ad>]␊
40694 │ >>␊
40695 │ startxref␊
40696 │ 185638183␊
40697 │ %%EOF␊
40698 │ 1·0·obj␍␊
40699 │ <</Pages·3·0·R·/Type·/Catalog·/AcroForm·373·0·R·>>·␍␊
40700 │ endobj␍␊
40701 │ 4·0·obj␍␊
40702 │ <</Contents·68·0·R·/MediaBox·[0·0·595.276·841.89·]·/Parent·3·0·R·/Resources·<</Font·<</F4·70·0·R·/F5·71·0·R·/F6·72·0·R·>>·/ProcSet·73·0·R·/XObject·<</Im3·74·0·R·/Im7·76·0·R·>>·>>·/Type·/Page·/Annots·378·0·R·>>·␍␊
40703 │ endobj␍␊
40704 │ 372·0·obj␍␊
40705 │ <</Type·/Annot·/Rect·[142.1173581129·531.1166868381·347.6343759828·552.3443721634·]·/Subtype·/FreeText·/DA·(/popplerfont·10.00·Tf\n)·/M·(D:20241117100201Z␀································)·/T·<feff00750073006500720031>·/Contents·<feff00750069006500620072006700730075006b0067006200730073006b00650075>·/NM·(okular-{c67d430b-37dd-4ba6-85c6-45a45db4e512})·/CreationDate·(D:20241117100143Z␀································)·/F·20·/C·[0.2156862766·0.1882352978·1·]·/CA·1·/Border·[0·0·2·]·/Q·0·/IT·/FreeText·/P·4·0·R·/AP·<</N·379·16·R·>>·/AS·/N·>>·␍␊
40706 │ endobj␍␊
40707 │ 373·0·obj␍␊
40708 │ <</Fields·[]·/DR·<</Font·<</popplerfont·377·0·R·>>·>>·>>·␍␊
40709 │ endobj␍␊
40710 │ 374·0·obj␍␊
40711 │ <</Length·210957·/Filter·/FlateDecode·>>·stream␍␊
#!/usr/bin/env python3
import sys
import pikepdf
from pikepdf import Name
import re
def change_annotation_font_and_border_to_red(input_path):
output_path = input_path.replace(".pdf", "-output.pdf")
# Open the PDF file
with pikepdf.open(input_path) as pdf:
for page_number, page in enumerate(pdf.pages, start=1):
# Check if the page has annotations
if "/Annots" in page:
print(f"Page {page_number} contains annotations:")
# Iterate over each annotation in the /Annots array
for annot in page["/Annots"]:
print(f" - Annotation detected: {annot}")
# Set the border color to red (RGB)
annot[Name("/C")] = pikepdf.Array([1, 0, 0]) # RGB for red border
# Update the default appearance string (/DA) to set text color to red
if "/DA" in annot:
da_str = str(annot[Name("/DA")])
print(f" Original /DA: {da_str}")
# Remove any existing color settings in the DA string
da_str = re.sub(r"(\d+\.?\d*\s){1,3}[rg]", "", da_str)
# Append red color setting for text color
da_str = da_str.strip() + " 1 0 0 rg"
annot[Name("/DA")] = da_str
print(f" Updated /DA: {annot[Name('/DA')]}")
# Remove the appearance stream to apply changes if present
if "/AP" in annot:
del annot[Name("/AP")]
print(" Removed /AP to force appearance regeneration")
# Save the modified PDF with '-output.pdf' suffix
pdf.save(output_path)
print(f"Font and border color changed to red. Saved as: {output_path}")
if __name__ == "__main__":
if len(sys.argv) != 2:
print("Usage: python change_annotation_font_and_border_to_red.py <input-pdf-path>")
else:
input_path = sys.argv[1]
change_annotation_font_and_border_to_red(input_path)
- thanks for looking into this, I've updated the question with examples of inline text annotations as produced by okular – vqqkomb0 Commented Nov 17, 2024 at 10:26
1 Answer
Reset to default 0You can round-trip the annotations with Cpdf. First:
cpdf -list-annotations-json in.pdf > out.json
This creates a JSON file with the annotations (and any associated objects). Now, edit the JSON as you wish. Then, we can re-apply them:
cpdf -remove-annotations in.pdf AND -set-annotations-json out.json -o out.pdf
More details in Chapter 10 of the manual.