You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Jan 20, 2021. It is now read-only.
This worked quite well for all the columns and rows, but for some reason the comments column wasn't extracted (it's all text of course). It looked like this:
@CMCDragonkai, there's a known issue (related to #99) where the detected selection area is just a little bit too small. If you expand the selection on the right just a little bit -- so it includes the border, not just the text -- then the comments column should be extracted too.
The bounding box detection -- for PDFs that support it -- uses the line elements in the PDF. So it totally should work, I just think we're at a bit of an impasse about the solution.
Activity
jeremybmerrill commentedon Jan 3, 2016
@CMCDragonkai, there's a known issue (related to #99) where the detected selection area is just a little bit too small. If you expand the selection on the right just a little bit -- so it includes the border, not just the text -- then the comments column should be extracted too.
CMCDragonkai commentedon Jan 4, 2016
I see, does is the bounding box detection based on image processing or something else? Surely there must be a way to match what is contained.
jeremybmerrill commentedon Jan 4, 2016
The bounding box detection -- for PDFs that support it -- uses the line elements in the PDF. So it totally should work, I just think we're at a bit of an impasse about the solution.