any idea for get drawings

[kalle07]

base is a pdf:
text-blocks all fine
tables all fine
images all fine
but drawings (like a diagram or a skech) are lines and rectangles and points.

for me would be okay to get an image from this regions but it is not recognized as image.

[JorjMcKie]

Vector graphics usually / often consist of many "atomic" draw commands like lines, curves, rectangles. They appear as a holistic graphic because of the geometrical vicinity of the atoms.
To be clear: These are no images and therefore cannot be extracted as such.
You can extract the drawings on a page via page.get_drawings(), which gives yo all this details. But these are Python dictionaries - no images.
You can also use page.cluster_drawings() which detects and clusters drawing atoms that are sufficiently close to each other.
It returns a list of rectangles which represent those clusters. Make a pixmap using a cluster rectangle as clip value - then you can save it as an image somewhere.

[kalle07]

thy , i will try ...

[kalle07]

works well if i restrict all clusters below 10 "atoms" (mostly some lines or a rectangle arround an image)