rich table(or chart?) content is not fully extracted

[shaojun]

Is your feature request related to a problem? Please describe.
I'm trying to develop a pdf to ppt tool, and the first step is to parse everything from pdf.

Now I have a pdf file has a page with 2 tables(maybe 2 charts?) content in it like:

and the library do can detect table in this page, but just 1, by use:

tabs = page.find_tables()  # locate and extract any tables on page
if tabs.tables:  # at least one table found?
    for tabno, tab in enumerate(tabs, start=1):
        table_extract = tab.extract()  # get the table extracted from the page
        # print(table_extract)  # print content of first table
        pdfPageDescriptor.tables.append(table_extract)

the table_extract is actually just one 2 dimention array of:

can see both the table or chart structure and data are in mess.

and also, I tried page.get_images() on this page, none of them related to the tables or charts area.
Describe the solution you'd like
parse out the tables or charts.

Describe alternatives you've considered
crop the tables or charts area as a image is also acceptable.

Additional context
Add any other context or screenshots about the feature request here.

[JorjMcKie]

There is no possible enhancement here I'm afraid.
This is a typical "Discussions" item, so let me transfer this post first.

[JorjMcKie]

We are not looking at a table at all on the shown page. There are vector graphics with some explanatory text - that's all.

You have to analyze this situation yourself, for example by checking if some vector graphic or image is contained in the detected table's bbox.
Vector graphics can be extracted via paths = page.get_drawings(). If the table finder thinks it has detected a table, you can check:

paths = page.get_drawings()
tab_bbox = fitz.Rect(tab.bbox)
if [p for p in paths if p["rect"] in tab_bbox] != []:  # some graphics are in table area!
        print("not a table")

[shaojun]

thanks @JorjMcKie.

your code sample:

if [p for p in paths if p["rect"] in tab_bbox] != []:  # some graphics are in table area!
        print("not a table")

indicates a suspicious chart area is found at position tab_bbox , right?
then the next step is to extract this area, I believe the most easy way to keep it is to crop this area to a image, could you show me how to crop an area and save it? I just find only thing about crop is targeting a whole page: https://pymupdf.readthedocs.io/en/latest/the-basics.html#cropping-a-pdf

[JorjMcKie]

I believe you were taken in the wrong direction by using the verb "crop" which is not applicable here.
What you seem to mean is taking a picture of only the drawing. This works by using the drawing's bbox as a clip area like this: pix = page.get_pixmap(clip=bbox). Then do pix.save("some.jpg") (or .png).
If you want, you can improve image resolution by using the dpi parameter when creating the pixmap.