Get Contour from list

Get Contour from list is the index node of Bmad's OpenCV contour toolchain: a CV_CONTOURS list in, one CV_CONTOUR out, chosen by an integer. If you've ever run a contour-finding node and gotten back a hundred blobs when you wanted just the one that mattered, this is the node that pulls it out.

It slots into the pack's CV pipeline, which mirrors OpenCV's classic workflow: threshold an image, find contours, then do geometry on the results. Contours runs OpenCV's findContours and returns the full list; Get Contour from list picks entry index; and downstream you can feed that single contour into ContourToMask (turn it into a mask), ContourGetBoundingRect (get its box), or Filter Contour (rank candidates by fitness). The practical move is to chain Contours → Filter Contour to get the list down to the shapes you care about, then index into it here.

How it works

The implementation is refreshingly honest:

def get_contour(self, contours, index):
    if index >= len(contours):
        return (None,)
    return (contours[index],)

So unlike the pack's FromListGet1* nodes, this one does not wrap around on a bad index - out of range hands you None instead. That's a deliberate design choice, and the pack leans on it: the contour nodes are built to tolerate None. ContourGetBoundingRect, for instance, prints "Contour = None !" and returns zeros rather than crashing. Which is your clue to keep an index that can legitimately miss - it's a way of saying "no contour here" without blowing up the graph.

Inputs and outputs

• contours - a CV_CONTOURS list, from the Contours node (same pack) or anything that outputs the contour type.
• index - an INT, default 0, minimum 0 - no negatives, no wrap.
• Output: one CV_CONTOUR, or None if the index is past the end.

Install

Ships in bmad4ever/comfyui_bmad_nodes. In ComfyUI Manager, search "comfyui_bmad_nodes" ("Bmad Nodes"), install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/bmad4ever/comfyui_bmad_nodes
cd comfyui_bmad_nodes
pip install -r requirements.txt

Restart ComfyUI after. This node needs OpenCV, which is in the requirements (opencv-python); no model files.

Common issues

The main gotcha is the asymmetry with the FromListGet1* nodes: index past the end here gives you None, not a wrap-around. That's usually what you want, but if you're driving the index from a random node, expect None results to propagate downstream - make sure whatever consumes the contour tolerates it (Bmad's own nodes do). Second, index has a floor of 0: negative indexes are clamped at the input level, so you can't count from the end. If you want the last contour, compute the list length first.