Near as I can tell, the problem relates to the fact that the page includes both a table that includes a bar graph. At least there's no problem with extracting tables on any other page in that file, and that's the only thing that stood out to me as being different about that page.
If that's actually the case might it be that the bars are being interpreted as really thick table lines?
Regardless, I was able to fix it locally by changing line 374 of utils.py from
if bbox_intersect(ba, bb):
to
if bbox_intersect(ba, bb) and bbox_area(ba):
which I'm hoping will be a clue to what's causing the problem.
camelot.read_pdf(fn, pages='all') fails occasionally with a ZeroDivisionError.
I chased the problem down to page 20 on https://www.accessdata.fda.gov/cdrh_docs/pdf6/P060040B.pdf
Near as I can tell, the problem relates to the fact that the page includes both a table that includes a bar graph. At least there's no problem with extracting tables on any other page in that file, and that's the only thing that stood out to me as being different about that page.
If that's actually the case might it be that the bars are being interpreted as really thick table lines?
Regardless, I was able to fix it locally by changing line 374 of utils.py from
if bbox_intersect(ba, bb):toif bbox_intersect(ba, bb) and bbox_area(ba):which I'm hoping will be a clue to what's causing the problem.