varir / copac Goto Github PK

The inner loop of the correlation distance computation should be vectorized for increased performance.

Unit tests would be great.

Code needs to be clean up for

• unused imports
• unused variables
• unused parameters

Also, check PEP8 compliance.

Description

Error in: from sklearn.cluster import DBSCAN when running: from copac import COPAC

Files

copac.py

To Reproduce

from copac import COPAC

Tasks

• line 19: ".dbscan_" should be removed from: sklearn.cluster.dbscan_ import dbscan.

Generalized DBSCAN is very similar to standard DBSCAN. It seems, sklearn's DBSCAN routines can be used. If so, the (incomplete) GDBSCAN functions in the copac module can be removed.
Otherwise, we need to implement them fully.

Description

AttributeError: module 'numpy' has no attribute 'float' 'int'

Screenshots

Files

copac.py

Tasks

• lines 143, 144: use 'float' and 'int' without 'np.'

Description

Name and Type errors related to dbscan when running copac.fit_predict

Screenshots

Files

copac.py

Tasks

The following modifications to lines 168-170 solved the errors.