Changes to work with updated cvxpy Variable

glrm/glrm.py

@@ -11,13 +11,13 @@
 class GLRM(object):
 
     def __init__(self, A, loss, regX, regY, k, missing_list = None, converge = None, scale=True):
-        
+
         self.scale = scale
         # Turn everything in to lists / convert to correct dimensions
         if not isinstance(A, list): A = [A]
         if not isinstance(loss, list): loss = [loss]
         if not isinstance(regY, list): regY = [regY]
-        if len(regY) == 1 and len(regY) < len(loss): 
+        if len(regY) == 1 and len(regY) < len(loss):
             regY = [copy(regY[0]) for _ in range(len(loss))]
         if missing_list and not isinstance(missing_list[0], list): missing_list = [missing_list]
 
@@ -30,8 +30,8 @@ def __init__(self, A, loss, regX, regY, k, missing_list = None, converge = None,
 
         # initialize cvxpy problems
         self._initialize_probs(A, k, missing_list, regX, regY)
-       
-        
+
+
     def factors(self):
         # return X, Y as matrices (not lists of sub matrices)
         return self.X, hstack(self.Y)
@@ -45,7 +45,7 @@ def predict(self):
         return hstack([L.decode(self.X.dot(yj)) for Aj, yj, L in zip(self.A, self.Y, self.L)])
 
     def fit(self, max_iters=100, eps=1e-2, use_indirect=False, warm_start=False):
-        
+
         Xv, Yp, pX = self.probX
         Xp, Yv, pY = self.probY
         self.converge.reset()
@@ -57,7 +57,7 @@ def fit(self, max_iters=100, eps=1e-2, use_indirect=False, warm_start=False):
             Xp.value[:,:-1] = copy(Xv.value)
 
             # can parallelize this
-            for ypj, yvj, pyj in zip(Yp, Yv, pY): 
+            for ypj, yvj, pyj in zip(Yp, Yv, pY):
                 objY = pyj.solve(solver=cp.SCS, eps=eps, max_iters=max_iters,
                         use_indirect=use_indirect, warm_start=warm_start)
                 ypj.value = copy(yvj.value)
@@ -67,7 +67,7 @@ def fit(self, max_iters=100, eps=1e-2, use_indirect=False, warm_start=False):
         return self.X, self.Y
 
     def _initialize_probs(self, A, k, missing_list, regX, regY):
-        
+
         # useful parameters
         m = A[0].shape[0]
         ns = [a.shape[1] for a in A]
@@ -107,7 +107,7 @@ def _initialize_A(self, A, missing_list):
 
         # compute stdev for entries that are not missing
         for ni, sv, mu, ai, missing, L in zip(ns, stdev, mean, A, missing_list, self.L):
-            
+
             # collect non-missing terms
             for j in range(ni):
                 elems = array([ai[i,j] for i in range(m) if (i,j) not in missing])
@@ -123,7 +123,7 @@ def _initialize_A(self, A, missing_list):
 
             # zero-out missing entries (for XY initialization)
             for (i,j) in missing: bi[i,j], mask[i,j] = 0, 0
-             
+
             B.append(bi) # save
             masks.append(mask)
             offsets.append(offset)
@@ -156,14 +156,14 @@ def _initialize_XY(self, B, k, missing_list):
         ns = cumsum([bj.shape[1] for bj in B])
         if len(ns) == 1: Y0 = [Y0]
         else: Y0 = split(Y0, ns, 1)
-        
+
         return X0, Y0
 
     def _finalize_XY(self, Xv, Yv):
         """ Multiply by std, offset by mean """
-        m, k = Xv.shape.size
+        m, k = Xv.size
         self.X = asarray(hstack((Xv.value, ones((m,1)))))
         self.Y = [asarray(yj.value)*tile(mask[0,:],(k+1,1)) \
                 for yj, mask in zip(Yv, self.masks)]
         for offset, Y in zip(self.offsets, self.Y): Y[-1,:] += offset[0,:]
-            
+

setup.py

@@ -10,5 +10,6 @@
         url="http://github.com/cehorn/GLRM/",
         license="MIT",
         install_requires=[  "numpy >= 1.8",
-                            "scipy >= 0.13"]
+                            "scipy >= 0.13",
+                            "cvxpy >= 0.4.10"]
 )