cannot find builtin CJK font

[HazyGuo]

I got this error while running the pdfviewer(wx).py file. What is the CJK font?

[JorjMcKie]

CJK = China, Japa, Korea
Can you send me an example PDF? Maybe there is a bigger issue underneath ...

[JorjMcKie]

Using PDFdisplay.py works ok?

[HazyGuo]

sorry, the pdf is confidential data getting from the financial customer.

[HazyGuo]

there is no error appear, but the text content can't display.

[JorjMcKie]

uhhh that's bad!!!

I am afraid I have to change the way how I generate PyMuPDF and MuPDF then.
Currently I am excluding fonts from the MuPDF generation whenever I thought I can do that. Obviously I have to re-consider this ...

How urgent is your issue? I need a few days to change and upload a new version ...

[HazyGuo]

thanks for your reply, and maybe one week remains for my issue.

[JorjMcKie]

I just checked what will happen if I include minimal CJK support in my MuPDF and subsequent PyMuPDF generation.

1. I do not know if this already solves your problem
2. It increases the size of the PyMuPDF binary (_fitz<...>.so or _fitz.pyd) from about 3 or 4 MB to 8 or 9 MB.

Are you using Windows? In that case I could send you a wheel to check if your issue is already solved ...

[HazyGuo]

yes, i write test code in windows.
it's pleasure that you can send it to my email(guoqinghe99@gmail.com)

[JorjMcKie]

I Need your Python and bitness combination, like Python 3.6, 64bit or similar Gesendet von Mail<https://go.microsoft.com/fwlink/?LinkId=550986> für Windows 10

…

________________________________ Von: HazyGuo <notifications@github.com> Gesendet: Monday, July 23, 2018 7:16:39 AM An: rk700/PyMuPDF Cc: Jorj X. McKie; Comment Betreff: Re: [rk700/PyMuPDF] cannot find builtin CJK font (#189) yes, i write test code in windows. it's pleasure that you can send it to my email(guoqinghe99@gmail.com<mailto:guoqinghe99@gmail.com>) — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#189 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AH6BohWiVl99CUNRHxbUsMCDQXAE2rDjks5uJbCXgaJpZM4Vaoed>.

[HazyGuo]

my python version is 3.5, and the operating system is windows 64

[JorjMcKie]

You should have received PyMuPDF-1.13.15-cp35-cp35m-win_amd64.whl by now. Please install it via

python -m pip install --upgrade PyMuPDF-1.13.15-cp35-cp35m-win_amd64.whl

[HazyGuo]

thank you very much, it can extract text content from my pdf.
but the structure of the html page rendered by the extracted content seems not like the original in the pdf.

in addition, can i distinguish which content is come from the table and which is not?

[JorjMcKie]

Did you mean the HTML page created from page.getText("html")?
This is the same as it is created by mutool draw ....
This is a different issue - has nothing to do with the font support. The same problem arises with HTML output for ASCII or LATIN fonts. The basic reason seems to be that the generated HTML contains font specifications that are not complete.

To your final question:
No, there is no way to find out whether some text comes from a table or not. The only way would be using text position information. This type of information is contained in page.getText("dict"), page.getTextBlocks(), page.getTextWords() and page.getText("xml").
All these methods provide various levels of info detail. By using them you can determine, where a line ("dict" method), a paragraph ("blocks" method), a word ("words" method), or even a character ("xml") method is located on the page.
You may want to have a look at the most current Wiki about extracting text from rectangles.

[HazyGuo]

thanks for your careful answer, and i get it now.