Inconsistent information between getText("dict")['blocks'] and getText("html")

[Yichen-fqyd]

Hello<

I really appreciate this great repo!
I am trying to add bounding box information to the extracted html to the corresponding p tag, two problems found during the process.

1. there is more p tags than provided bbox information in "dict",
2. the "dict" will return the (0.0, 0.0, 0.0, 0.0) bounding box, more than p tag
   Since these two using the same function, I would expect the provided information be the same, but it happened, I am wondering why and any other easier way to achieve the purpose of adding bounding box information to the html file.

Thanks

[JorjMcKie]

I am afraid this will not work.
HTML, XHTML and XML extraction options are based on original MuPDF functions and as such must be accepted as they are.
The other options are my own making.
To "my" functions, over time and upon request, I added corrective code where errors were reported and introduced some extended features like reduced glyph heights or reducing the text amount to a given clip rectangle.

So when you see a zero bbox in the *ML files, there is nothing I can do. Such things go back to an inconsistent / erroneous PDF or font information. Any corrective code I may be using in my functions cannot be taken over to the *ML functions.

[Yichen-fqyd]

Thank you for the quick response,

Sorry for the confusion, I am just trying to add the information to the retuned html file. so I am trying to find the correspondence between theses two functions. Currently, for the most cases, line level found by the "dict" and p tag has nice correspondence.

And zero bbox is actually returned by getText("dict")['blocks'] in a file

[JorjMcKie]

And zero bbox is actually returned by getText("dict")['blocks'] in a file

If you want and have no confidentiality concerns, let me have that file and check, whether this issue may be solved by the pending v1.18.10 code.

[Yichen-fqyd]

Can I send it to your email: jorj.x.mckie@outlook.de

[JorjMcKie]

Sure!