def recall()

in mapillary_sls/utils/eval.py [0:0]

• 9 lines of code
• 1 McCabe index (conditional complexity)

def recall(ranks, pidx, ks):

	recall_at_k = np.zeros(len(ks))
	for qidx in range(ranks.shape[0]):

		for i, k in enumerate(ks):
			if np.sum(np.in1d(ranks[qidx,:k], pidx[qidx])) > 0:
				recall_at_k[i:] += 1
				break

	recall_at_k /= ranks.shape[0]

	return recall_at_k