diff options
Diffstat (limited to 'src/python/gudhi/point_cloud')
| -rw-r--r-- | src/python/gudhi/point_cloud/knn.py | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/src/python/gudhi/point_cloud/knn.py b/src/python/gudhi/point_cloud/knn.py index dec5f88f..0724ce94 100644 --- a/src/python/gudhi/point_cloud/knn.py +++ b/src/python/gudhi/point_cloud/knn.py @@ -259,9 +259,11 @@ class KNearestNeighbors: neighbors, distances = self.graph.knn_query(X, k, num_threads=self.params["num_threads"]) if numpy.any(numpy.isnan(distances)): import warnings + warnings.simplefilter("always") warnings.warn("NaN value encountered while computing 'distances'", RuntimeWarning) if numpy.any(numpy.isinf(distances)): import warnings + warnings.simplefilter("always") warnings.warn("Overflow value encountered while computing 'distances'", RuntimeWarning) # The k nearest neighbors are always sorted. I couldn't find it in the doc, but the code calls searchKnn, # which returns a priority_queue, and then fills the return array backwards with top/pop on the queue. @@ -298,9 +300,11 @@ class KNearestNeighbors: distances, neighbors = mat.Kmin_argKmin(k, dim=1) if torch.isnan(distances).any(): import warnings + warnings.simplefilter("always") warnings.warn("NaN value encountered while computing 'distances'", RuntimeWarning) if torch.isinf(distances).any(): import warnings + warnings.simplefilter("always") warnings.warn("Overflow encountered while computing 'distances'", RuntimeWarning) if p != numpy.inf: distances = distances ** (1.0 / p) @@ -312,9 +316,11 @@ class KNearestNeighbors: distances = mat.Kmin(k, dim=1) if torch.isnan(distances).any(): import warnings + warnings.simplefilter("always") warnings.warn("NaN value encountered while computing 'distances'", RuntimeWarning) if torch.isinf(distances).any(): import warnings + warnings.simplefilter("always") warnings.warn("Overflow encountered while computing 'distances'", RuntimeWarning) if p != numpy.inf: distances = distances ** (1.0 / p) |