Inserting MathML into a PDF

[XHJ-TS9527]

Hi,
I'm trying to insert some latex equations into a pdf file. Since the program will be deployed in an environment without any latex compiler, I would like to try another method: Page.insert_htmlbox(). I read the document and noticed that this method can handle HTML texts. Therefore, I think that I can achieve my goal by converting latex texts into MathML and embed it into the pdf file utilizing the insert_htmlbox function. However, this method doesn't seem to work. The PDF generated does display some symbols in the original equations (like \theta and \rightarrow), but it doesn't display some import symbols like the square root symbol and fraction line. I copied the same MathML text into another html file and opened it with Chrome, and these symbols can be displayed correctly in the browser. I would like to know if I have missed some parameters when calling the insert_htmlbox function. If this function cannot achieve this, how can I embed equations and make them display correctly in the processed PDF using PyMuPDF?

[JorjMcKie]

Interesting question! As usual, our base library MuPDF does the lion's share of achieving the results of .insert_htmlbox. It supports a syntax subset of HTML version 4 and CSS version 2.
This probably explains the observed differences.

Maybe you can make a SVG version of your equation. This can be mixed in other HTML text as the attached example shows. It it a script that outputs a HTML table where some cells display images - of which one is a SVG.
example.zip

[XHJ-TS9527]

Thank you for your response. This seems to be a good solution. However, I encountered another problem when I tried to follow this method. I found a library named ziamath, which can directly transform a latex equation into an SVG. This library needs specific fonts (like DejaVuMathTeXGyre) to display the equations correctly. For demonstration, I use page.insert_htmlbox(page.rect, '<img src="the equation svg file path">', archive=fitz.Archive('.')) to insert this svg file into the page, but the PDF generated only shows some lines (I observed my equation and found that these lines are fraction lines or parts of square root symbol), and no symbols (like the equation symbol or the right arrow symbol) are displayed. I guess this may be caused by missing this font in the page, so I add page.insert_font('DejaVuMathTeXGryre', the path to this font) before the calling of the insert_htmlbox function, but this doesn't work as the generated PDF file displays no symbol either. I would like to know if there is anything wrong with my operations. Thank you for your help!

[JorjMcKie]

Your SVG is not self-contained obviously: it needs a specific font to be displayable. So if you take it to a machine without that font, it won't be shown correctly by the browser either.

Try to persuade that ziamath package to make no reference to a font, but instead store the characters using their graphical representations WRT to the font.
There is a similar difference in PyMuPDF: If you make a SVG image of the page, you have the option to either refer to fonts or store characters as little vector graphics: Page.get_svg_image(text_as_path=True).

[XHJ-TS9527]

Thanks for your reminding. I have permission to install new fonts on a new machine, so it won't be a problem to get access to new fonts. I set the 'svg2' flag as False in the ziamath library, and the equation SVG can now be fully displayed in the generated PDF file. That's a big progress. However, it seems that the inserted SVG is a bitmap. I discovered image aliasing when I zoomed in a PDF page. I noticed that the PDF file in the example zip file treats the included SVG as a bitmap as well, since the same thing happens when I zoom in. This phenomenon will not happen if I directly save the same HTML script as an HTML file and open it in the browser. I wonder whether all images, either bitmaps or vector graphics, are treated as bitmaps in the insert_htmlbox function?

[JorjMcKie]

Yes, right this indeed seems to be the case.

[XHJ-TS9527]

Alright, that shouldn't be a problem. I will adjust the font size to ensure the equation displays appropriately. Thanks for your patience and timely assistance!