extract table headings along with table contents

[poojitharamachandra]

hi,
how to extract the table headings in addition to table contents ?

[jsvine]

Hi @poojitharamachandra; thanks for your interest in this library. There's no built-in automated functionality for what you describe, because the logic ends up being quite custom to the particular layout and structure of any given PDF. But the general idea would be to use page.find_tables() to find the tables on a page; each table's .bbox property will give you its coordinates, which you can use along with page.crop((x0, top, x1, bottom)).extract_text() to select an area above the table (perhaps optionally with page.filter(...), depending on what else is in that area), with those coordinates determined by you based on the spacing between the table and its heading.

[poojitharamachandra]

thanks, that helped

[poojitharamachandra]

is there a way to extract the section headings under which a particular table is present in the pdf?

[dhdaines]

There isn't any general way to do this, because section headings could be represented visually in a multitude of ways.

If it's a tagged PDF then the section headings themselves may be tagged with a header-level tag, e.g. H1, H2, H3, etc, which you can access in the tag entry of the dictionaries returned by .chars or .extract_words(). But even in tagged PDFs there isn't any notion of a "section", so you would just have to assume that the most recent section heading is the one which contains a given table.