Access to plain text of a page? about pymupdf

[rk700]

@JorjMcKie, yes there should be. I'll add the API for extracting image/text from a page.

[JorjMcKie]

Hi @rk700,
it is good to see you could find some time in the middle of your big change of life! I hope everything is turning out well for you.
In the meantime I have been trying to continue some of the work.

• some stuff indocumentation
• a solution for a table of contents
• added a close function for PDFs, so that the running process can delete / rename the input PDF. I have been guessing here: is it sufficient to just do a pdf_close_document? What about other file types?

I am afraid, I might have destroyed your change concerning text extraction? I will try it out ...

[rk700]

@JorjMcKie, thank you for your work on the project! Especially when I could not devote too much time on it.

For closing doc, I think we can open a new issue. Currently ,fz_drop_document, in which doc is closed, happens in the destructor of the Document object. That is, the document will not be closed before there are still python references to it. But usually, the destructor is not called as soon as there're no references. So we should add an explicit closing function and maybe then remove the Document destructor. BTW, fz_drop_document can deal with all the supported doc types.

And don't worry about the text extraction code, it can be recovered easily from the commit history:)

[JorjMcKie]

@rk700, I have done a "repair" to the extractText, so it again works.
For the close() issue: I did add a call to pdf_close_document in fitz.i. And it worked (see this large example for editing table of contents)! Of course I do not know all consequences.

[rk700]

@JorjMcKie, the global context would live as long as the module, so it is freed when the program exits. And we can use the general function fz_drop_document for other types of files, since it will call different closing function for different types.

[JorjMcKie]

Text extraction tested - closing the issue.