Do you like Quick, Draw? Well what if you could train/predict doodles drawn inside Streamlit? Also draws lines, circles and boxes over background images for annotation.
I want to use the canvas rect (initiated via initial drawing) object as data info for cropping application, within transform mode.
I am however having a problem with UX when user mistakenly delete the box by double click during dragging action.
Options?
Is there anyway I can:
disable double clicking within the canvas
OR have a way to keep at least one rect inside the canvas_result.json_data
OR auto re-add one rect once empty
in the same topic, is there a way to manually inject the shape into the object besides initial drawing?
I know there seems to be multiple questions, but there are all for this single use case. Either one will do.
Research done so far
I have searched and found streamlit_cropper library, but there is another use case for canvas and thus if possible I wanted to use this library. Also encountered a weird bug with this, might need to raise a bug there as well later.
Tried searching for double click disable via JS, but it seems to be a XSS issue due to the iframe
Tried searching for delete in issues list & streamlit forum but couldn't find relevant portion
Might need to find a way to fork and use local version rather than PIP, but I really hope it won't come to that. Do give me a feedback if this is indeed the way to go.
One intermediate solution is just to include big notice for user to click undo if no more rect present on the page.
Hi, first of all, thank you for the great library
Use case
I want to use the canvas rect (initiated via initial drawing) object as data info for cropping application, within transform mode. I am however having a problem with UX when user mistakenly delete the box by double click during dragging action.
Options?
Is there anyway I can:
in the same topic, is there a way to manually inject the shape into the object besides initial drawing?
I know there seems to be multiple questions, but there are all for this single use case. Either one will do.
Research done so far