Highlight pdf

[maivan-hoa]

I highlight the pdf file, but the result is not as smooth as expected. Has anyone encountered this error before?

Here is my code:

    list_text = ["Tiếng Việt được chính thức ghi nhận trong Hiến pháp nước Cộng hòa xã hội chủ nghĩa Việt Nam 2013, tại Chương I Điều 5 
                    Mục 3, là ngôn ngữ quốc gia của Việt Nam .", "Tôi rất mong nhận được sự góp ý, đánh giá của các thầy cô bộ môn để đề tài 
                    của tôi thêm hoàn thiện hơn."]
    doc = fitz.open(input_file)
    
    for page in doc:
        for text in list_text:
            ### SEARCH
            text_instances = page.search_for(text, quads=True)
        
            ### HIGHLIGHT
            for inst in text_instances:
                highlight = page.add_highlight_annot(inst)
          
    doc.save(output_file, garbage=4, deflate=True, clean=True)

[JorjMcKie]

And what is the error? The code does what you told it to do.

Don't understand why you use .update() though.

You did not mention your version.

[JorjMcKie]

Let me convert this to a "Discussions" item. You seem to have questions - as opposed to reporting an issue.

[JorjMcKie]

If you don't share an example PDF and the failing code, this will not lead us to anywhere.

[maivan-hoa]

I used the above code and here is the pdf file: https://drive.google.com/file/d/1Nbvx1JcQg0fcJ09r6fu1WQAGDBQqc8gS/view?usp=sharing

[JorjMcKie]

Thanks.
I have looked at it, and indeed: The behavior of PyMuPDF seems awkward.
Of course one would expect 4 rectangles, because your two strings each cover 2 lines. But the PDF creator, MS Word used 2 different fonts (which is not visible by you and me).
The tricky things is that both fonts have the same name, "TimesNewRomanPSMT", but different properties in terms of line height and other geometry properties.
So the "text_instances" of this page consist of multiple rectangles in total, because PyMuPDF of course does not / cannot decide to join characters in the same rectangle if they do not have the same base line. Here is the list of resulting hit rects

[Rect(72.02400207519531, 175.25314331054688, 83.54688262939453, 192.54177856445312),
 Rect(83.54399871826172, 177.3511962890625, 89.2962646484375, 191.70338439941406),
 Rect(89.30400085449219, 175.25314331054688, 119.41008758544922, 192.54177856445312),
 Rect(119.41999816894531, 177.3511962890625, 153.51336669921875, 191.70338439941406),
 Rect(153.5, 175.25314331054688, 206.7396697998047, 192.54177856445312),
 Rect(206.80999755859375, 177.3511962890625, 213.83421325683594, 191.70338439941406),
 Rect(213.88999938964844, 175.25314331054688, 257.66888427734375, 192.54177856445312),
 Rect(257.69000244140625, 177.3511962890625, 263.4422607421875, 191.70338439941406),
 Rect(263.45001220703125, 175.25314331054688, 318.8410339355469, 192.54177856445312),
 Rect(318.8900146484375, 177.3511962890625, 385.2833557128906, 191.70338439941406),
 Rect(385.2699890136719, 175.25314331054688, 403.9842224121094, 192.54177856445312),
 Rect(403.989990234375, 177.3511962890625, 410.4700012207031, 191.70338439941406),
 Rect(410.4700012207031, 175.25314331054688, 473.559326171875, 192.54177856445312),
 Rect(473.739990234375, 177.3511962890625, 480.2200012207031, 191.70338439941406),
 Rect(480.2200012207031, 175.25314331054688, 500.2431945800781, 192.54177856445312),
 Rect(500.260009765625, 177.3511962890625, 506.7400207519531, 191.70338439941406),
 Rect(506.8599853515625, 175.25314331054688, 510.0999755859375, 192.54177856445312),
 Rect(511.05999755859375, 177.3511962890625, 543.330322265625, 191.70338439941406),
 Rect(72.02400207519531, 191.45309448242188, 84.98688507080078, 208.74172973632812),
 Rect(84.98400115966797, 193.5511474609375, 90.73626708984375, 207.90333557128906),
 Rect(90.74400329589844, 191.45309448242188, 162.12767028808594, 208.74172973632812),
 Rect(162.13999938964844, 193.5511474609375, 246.5222625732422, 207.90333557128906),
 Rect(246.52999877929688, 191.45309448242188, 277.5951232910156, 208.74172973632812),
 Rect(277.6099853515625, 193.5511474609375, 284.0899963378906, 207.90333557128906),
 Rect(284.0899963378906, 191.45309448242188, 360.981689453125, 208.74172973632812),
 Rect(361.0299987792969, 193.5511474609375, 368.0542297363281, 207.90333557128906),
 Rect(368.1099853515625, 191.45309448242188, 384.3100280761719, 208.74172973632812),
 Rect(384.42999267578125, 193.5511474609375, 390.9100036621094, 207.90333557128906),
 Rect(390.9100036621094, 191.45309448242188, 424.7701721191406, 208.74172973632812),
 Rect(424.75, 193.5511474609375, 431.2300109863281, 207.90333557128906),
 Rect(431.3500061035156, 191.45309448242188, 453.321533203125, 208.74172973632812),
 Rect(453.3399963378906, 193.5511474609375, 459.0922546386719, 207.90333557128906),
 Rect(459.1000061035156, 191.45309448242188, 497.72076416015625, 208.74172973632812)]
--------------------------------------------------
[Rect(292.8500061035156, 296.2431335449219, 318.7440490722656, 313.5317687988281),
 Rect(318.7699890136719, 298.3411865234375, 324.5222473144531, 312.693359375),
 Rect(324.54998779296875, 296.2431335449219, 377.4267883300781, 313.5317687988281),
 Rect(377.4700012207031, 298.3411865234375, 413.4833679199219, 312.693359375),
 Rect(413.4700012207031, 296.2431335449219, 427.622314453125, 313.5317687988281),
 Rect(427.6300048828125, 298.3411865234375, 434.65423583984375, 312.693359375),
 Rect(434.8299865722656, 296.2431335449219, 438.0699768066406, 313.5317687988281),
 Rect(438.19000244140625, 298.3411865234375, 534.219970703125, 312.693359375),
 Rect(534.3400268554688, 296.2431335449219, 543.3400268554688, 313.5317687988281),
 Rect(72.02400207519531, 312.3231201171875, 102.63841247558594, 329.61175537109375),
 Rect(102.62000274658203, 314.4211730957031, 108.37226867675781, 328.7733459472656),
 Rect(108.37999725341797, 312.3231201171875, 140.0542449951172, 329.61175537109375),
 Rect(140.17999267578125, 314.4211730957031, 146.6599884033203, 328.7733459472656),
 Rect(146.66000366210938, 312.3231201171875, 149.90000915527344, 329.61175537109375),
 Rect(149.89999389648438, 314.4211730957031, 188.53225708007812, 328.7733459472656),
 Rect(188.5399932861328, 312.3231201171875, 191.77999877929688, 329.61175537109375),
 Rect(191.92999267578125, 314.4211730957031, 204.16226196289062, 328.7733459472656),
 Rect(204.1699981689453, 312.3231201171875, 229.3527374267578, 329.61175537109375),
 Rect(229.3699951171875, 314.4211730957031, 235.84999084472656, 328.7733459472656),
 Rect(235.85000610351562, 312.3231201171875, 333.432861328125, 329.61175537109375),
 Rect(333.42999267578125, 314.4211730957031, 372.05657958984375, 328.7733459472656)]
--------------------------------------------------

This is the list of fonts used:

pprint(page.get_fonts())
[(6, 'n/a', 'TrueType', 'TimesNewRomanPSMT', 'F1', 'WinAnsiEncoding'),
 (9, 'ttf', 'Type0', 'TimesNewRomanPSMT', 'F2', 'Identity-H')]

There is no easy way to cope with this.
If you are searching for long phrases, I recommend joining the results by line yourself before adding the highlight annot.

[JorjMcKie]

This script is better:

import fitz

input_file = "input_test.pdf"
list_text = [
    """Tiếng Việt được chính thức ghi nhận trong Hiến pháp nước Cộng hòa xã hội chủ nghĩa Việt Nam 2013, tại Chương I Điều 5 Mục 3, là ngôn ngữ quốc gia của Việt Nam .""",
    """Tôi rất mong nhận được sự góp ý, đánh giá của các thầy cô bộ môn để đề tài của tôi thêm hoàn thiện hơn.""",
]
doc = fitz.open(input_file)

for page in doc:
    page.clean_contents()
    for text in list_text:
        ### SEARCH
        rlist = page.search_for(text)
        rect = rlist[0]
        i = 1
        for i, r in enumerate(rlist[i:], start=1):
            if r.tl.x > rect.tl.x:  # still in same line
                rect |= r
            else:  # new line started
                page.add_highlight_annot(rect)  # add highlight previous line
                rect = r
        page.add_highlight_annot(rect)

doc.save(__file__.replace(".py", "-2.pdf"), garbage=4, deflate=True, clean=True)

[maivan-hoa]

Thank you very much🎉