Merge pull request #73 from MattScicluna/use_numba

Add option for numba acceleration. Improve tests.

Author: bjoaofelipe

README.rst

@@ -44,6 +44,10 @@ Or, to install the latest version from github::
 
     pip install --user git+git://github.com/KrishnaswamyLab/graphtools.git
 
+For improved performance with large datasets, install with numba acceleration::
+
+    pip install --user graphtools[fast]
+
 Usage example
 -------------
 

graphtools/api.py

@@ -31,6 +31,7 @@ def Graph(
     adaptive_k=None,
     n_landmark=None,
     n_svd=100,
+    random_landmarking=False,
     n_jobs=-1,
     verbose=False,
     random_state=None,
@@ -141,6 +142,11 @@ def Graph(
     n_svd : `int`, optional (default: 100)
         number of SVD components to use for spectral clustering
 
+    random_landmarking : `bool`, optional (default: False)
+        If True, use random landmark selection instead of spectral clustering.
+        Randomly selects n_landmark points and assigns samples to nearest landmark.
+        Only used when n_landmark is not None.
+
     random_state : `int` or `None`, optional (default: `None`)
         Random state for random PCA