Version 0.0.20 has been released

[JorjMcKie]

This fixes some more bugs.
Here is an overview of other changes:

• Two new parameters ignore_images and ignore_graphics can be used to omit these objects when determining the page layout and will be omitted in the output. Can be helpful with overly crowded pages.
• Parameter table_strategy can now be None to suppress execution of PyMuPDF's table finder. This can significantly increase conversion speed.
• More flexibility in header identification: the new parameter max_levels (an integer between 1 and 6) limits the number of detected header levels. Here is a code snippet:

import pymupdf, pymupdf4llm

doc = pymupdf.open("input.pdf")
headers = pymupdf4llm.IdentifyHeaders(doc, max_levels=3)
md = pymupdf4llm.to_markdown(doc, hdr_info=headers)
...

The output produced above will only have headers up to level 3 ("###"). All text with a smaller font size than that of level 3 will be treated as body text.

The complete changes are here.