Remove pkg_resources dependency from sambar.py

netZooPy/condor/condor.py

@@ -305,13 +305,14 @@ def matrices(self, c,resolution):
             # Computes weighted biadjacency matrix.
             A = np.matrix(np.zeros((p, q)))
             for edge in self.net.iterrows():
-                A[gn[edge[1][1]], rg[edge[1][0]]] = edge[1][2]
+                row = edge[1]
+                A[gn[row.iloc[1]], rg[row.iloc[0]]] = row.iloc[2]
 
             # Computes node degrees for the nodesets.
             ki = A.sum(1)
             dj = A.sum(0)
             # Computes sum of edges and bimodularity matrix.
-            m = float(sum(ki))
+            m = float(ki.sum())
             B = A - resolution*((ki @ dj) / m)
 
             # d = self.index_dict

netZooPy/sambar/sambar.py

@@ -5,7 +5,6 @@
 import networkx as nx
 from scipy.spatial.distance import pdist,cosine,squareform
 from scipy.cluster.hierarchy import linkage,cut_tree
-import pkg_resources
 """
 Description:
     Python implementation of the Subtyping Agglomerated Mutations By Annotation Relations (SAMBAR) method as implemented in R https://github.com/mararie/SAMBAR.
@@ -40,10 +39,11 @@
             This package includes the functions ```sambar```,```desparsify```,```corgenelength```,```convertgmt```,```clustering``` as well as an implementation of the binomial distance (Millar dissimilarity from the package vegdist from R. To see the full description of each of this functions use ```help(pysambar.function)```.
 """
 ## Default toydata files
-esize = pkg_resources.resource_filename(__name__, 'esizef.csv')
-genes = pkg_resources.resource_filename(__name__, 'genes.txt')
-sign  = pkg_resources.resource_filename(__name__, 'h.all.v6.1.symbols.gmt')
-mut   = pkg_resources.resource_filename(__name__, 'mut.ucec.csv')
+_DATA_DIR = os.path.dirname(__file__)
+esize = os.path.join(_DATA_DIR, 'esizef.csv')
+genes = os.path.join(_DATA_DIR, 'genes.txt')
+sign  = os.path.join(_DATA_DIR, 'h.all.v6.1.symbols.gmt')
+mut   = os.path.join(_DATA_DIR, 'mut.ucec.csv')
 
 def corgenelength(mut,cangenes,esize,normbysample=True,subcangenes=True):
     """

tests/test_cobra.py

@@ -37,7 +37,8 @@ def test_cobra():
     pd.testing.assert_frame_equal(G, G_gt, rtol=1e-10, check_exact=False)
 
     q = psi.shape[0]
+    X_mean = np.mean(X.to_numpy(), axis=0)
     for i in range(q):
-        C = Q.to_numpy().dot(np.mean(X, axis=0)[i] * np.diag(psi.to_numpy()[i, :])).dot(Q.to_numpy().T)
-        C_gt = Q_gt.to_numpy().dot(np.mean(X, axis=0)[i] * np.diag(psi_gt.to_numpy()[i, :])).dot(Q_gt.to_numpy().T)
+        C = Q.to_numpy().dot(X_mean[i] * np.diag(psi.to_numpy()[i, :])).dot(Q.to_numpy().T)
+        C_gt = Q_gt.to_numpy().dot(X_mean[i] * np.diag(psi_gt.to_numpy()[i, :])).dot(Q_gt.to_numpy().T)
         pd.testing.assert_frame_equal(pd.DataFrame(C), pd.DataFrame(C_gt), rtol=1e-10, check_exact=False)