lmcinnes
commented
1 year ago It looks as if numba is balking at some type issues, but it is not clear to me what those could be from the error. This has worked acceptably on a few different machines and configurations I tried, so it isn't the default state of things.
What version of numba are you using?
On Wed, Mar 15, 2023 at 7:21 PM zilch42 @.***> wrote:
Hi there, I'm getting an error trying to import HDBSCAN from fast_hdbscan
from fast_hdbscan import HDBSCAN
TypingError Traceback (most recent call last) Cell In[1], line 10 8 from sklearn.feature_extraction.text import CountVectorizer 9 from umap import UMAP ---> 10 from fast_hdbscan import HDBSCAN 11 import pickle 12 import sys
File c:\Users\abb064\AppData\Local\miniconda3\envs\csiro-horizon-scanning39\lib\site-packages\fast_hdbscan__init__.py:7 5 random_state = np.random.RandomState(42) 6 random_data = random_state.random(size=(100, 3)) ----> 7 HDBSCAN(allow_single_cluster=True).fit(random_data) 8 HDBSCAN(cluster_selection_method="leaf").fit(random_data) 10 all = ["HDBSCAN", "fast_hdbscan"]
File c:\Users\abb064\AppData\Local\miniconda3\envs\csiro-horizon-scanning39\lib\site-packages\fast_hdbscan\hdbscan.py:217, in HDBSCAN.fit(self, X, y, fit_params) 207 clean_data = X 209 kwargs = self.getparams() 211 ( 212 self.labels, 213 self.probabilities_, 214 self._single_linkage_tree, 215 self._condensed_tree, 216 self._min_spanning_tree, --> 217 ) = fast_hdbscan(clean_data, return_trees=True, kwargs) 219 self._condensed_tree = to_numpy_rec_array(self._condensed_tree) 221 if not self._all_finite: 222 # remap indices to align with original data in the case of non-finite entries.
File c:\Users\abb064\AppData\Local\miniconda3\envs\csiro-horizon-scanning39\lib\site-packages\fast_hdbscan\hdbscan.py:149, in fast_hdbscan(data, min_samples, min_cluster_size, cluster_selection_method, allow_single_cluster, return_trees) 147 sklearn_tree = KDTree(data) 148 numba_tree = kdtree_to_numba(sklearn_tree) --> 149 edges = parallel_boruvka( 150 numba_tree, min_samples=min_cluster_size if min_samples is None else min_samples 151 ) 152 sorted_mst = edges[np.argsort(edges.T[2])] 153 linkage_tree = mst_to_linkage_tree(sorted_mst)
File c:\Users\abb064\AppData\Local\miniconda3\envs\csiro-horizon-scanning39\lib\site-packages\fast_hdbscan\boruvka.py:270, in parallel_boruvka(tree, min_samples) 267 while n_components > 1: 268 candidate_distances, candidate_indices = boruvka_tree_query(tree, node_components, point_components, 269 core_distances) --> 270 new_edges = merge_components(components_disjoint_set, candidate_indices, candidate_distances, point_components) 271 update_component_vectors(tree, components_disjoint_set, node_components, point_components) 273 edges = np.vstack((edges, new_edges))
File c:\Users\abb064\AppData\Local\miniconda3\envs\csiro-horizon-scanning39\lib\site-packages\numba\core\dispatcher.py:468, in _DispatcherBase._compile_for_args(self, *args, **kws) 464 msg = (f"{str(e).rstrip()} \n\nThis error may have been caused " 465 f"by the following argument(s):\n{args_str}\n") 466 e.patch_message(msg) --> 468 error_rewrite(e, 'typing') 469 except errors.UnsupportedError as e: 470 # Something unsupported is present in the user code, add help info 471 error_rewrite(e, 'unsupported_error') ... File "..........\AppData\Local\miniconda3\envs\csiro-horizon-scanning39\lib\site-packages\fast_hdbscan\boruvka.py", line 9: def merge_components(disjoint_set, candidate_neighbors, candidate_neighbor_distances, point_components): component_edges = {0: (0, np.int32(1), np.float32(0.0)) for i in range(0)}
Python 3.9.16
— Reply to this email directly, view it on GitHub https://github.com/TutteInstitute/fast_hdbscan/issues/1, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC3IUBIUEQ6USTT5QXR53NTW4JFIDANCNFSM6AAAAAAV4PR6HY . You are receiving this because you are subscribed to this thread.Message ID: @.***>
zilch42
VolkerH
FotiosBistas
Hi there, I'm getting an error trying to import HDBSCAN from fast_hdbscan
from fast_hdbscan import HDBSCANPython 3.9.16