Review the doc and fix imports.

wrapper/matlab/+matfaust/+fact/eigtj.m

@@ -24,7 +24,8 @@
 %> @param 'tol', number [optional if nGivens is set] the tolerance error at which the algorithm stops. The default value is zero so that stopping is based on reaching the targeted nGivens.
 %> @param 'order', char [optional, default is ‘ascend’] order of eigenvalues, possible choices are ‘ascend, 'descend' or 'undef' (to avoid a sorting operation and save some time).
 %> @param 'nGivens_per_fac', integer [optional, default is <code>floor(size(M, 1)/2)</code>] targeted number of Givens rotations per factor of V. Must be an integer between 1 to <code>floor(size(M, 1)/2)</code>.
-%> @param 'relerr', bool [optional, default is True] the type of error used as stopping criterion. (true) for the relative error norm(V*D*V'-M, 'fro')/norm(M, 'fro'), (false) for the absolute error norm(V*D*V'-M, 'fro').
+%> @param 'relerr', bool [optional, default is true] the type of error used as stopping criterion. (true) for the relative error norm(V*D*V'-M, 'fro')/norm(M, 'fro'), (false) for the absolute error norm(V*D*V'-M, 'fro').
+%> @param 'enable_large_Faust', bool [optional, default is false] if true, it allows to compute a transform that doesn't worth it regarding its complexity relatively to the matrix M. Otherwise, by default, an exception is raised before the algorithm starts.
 %> @param 'verbosity', integer [optional] verbosity level. The greater the value the more info is displayed. It can be helpful to understand for example why the algorithm stopped before reaching the tol error or the number of Givens (nGivens).
 %>
 %>
@@ -45,7 +46,7 @@
 %> % get a Laplacian to diagonalize
 %> load('Laplacian_256_community.mat')
 %> % do it
-%> [Uhat, Dhat] = eigtj(Lap, 'nGivens', size(Lap,1)*100, 'nGivens_per_fac', size(Lap, 1)/2, 'verbosity', 2)
+%> [Uhat, Dhat] = eigtj(Lap, 'nGivens', size(Lap,1)*100, 'nGivens_per_fac', size(Lap, 1)/2, 'verbosity', 2, 'enable_large_Faust', true)
 %> % Uhat is the Fourier matrix/eigenvectors approximation as a Faust (200 factors)
 %> % Dhat the eigenvalues diagonal matrix approx.
 %> % Computing the decomposition of the same matrix but targeting a precise relative error

wrapper/matlab/+matfaust/+fact/hierarchical.m

@@ -50,7 +50,7 @@
 %> import matfaust.*
 %> import matfaust.fact.hierarchical
 %> % generate a Hadamard Faust of size 32x32
-%> FH = wh(32);
+%> FH = wht(32);
 %> H = full(FH); % the full matrix version
 %> % factorize it
 %> FH2 = hierarchical(H, 'squaremat');

wrapper/matlab/+matfaust/+fact/svdtj.m

@@ -19,6 +19,7 @@
 %> @param 'tol', number see fact.eigtj (NB: as described below, the error tolerance is not exactly for the approximate SVD but for the subsequent eigtj calls).
 %> @param 'relerr', bool see fact.eigtj
 %> @param 'nGivens_per_fac',integer see fact.eigtj
+%> @param 'enable_large_Faust',bool see fact.eigtj
 %>
 %> @retval [U,S,V]: such that U*S*V' is the approximate of M with:
 %>      - S: (sparse real diagonal matrix) the singular values in descendant order.

wrapper/python/pyfaust/fact.py

@@ -102,6 +102,7 @@ def svdtj(M, nGivens=None, tol=0, order='ascend', relerr=True,
             the svd but for the subsequent eigtj calls).
             relerr: see fact.eigtj
             nGivens_per_fac: see fact.eigtj
+            enable_large_Faust: see fact.eigtj
 
 
         Returns:
@@ -208,7 +209,9 @@ def eigtj(M, nGivens=None, tol=0, order='ascend', relerr=True,
         info is displayed. It can be helpful to understand for example why the
         algorithm stopped before reaching the tol error or the number of Givens
         (nGivens).
-
+        enable_large_Faust: (bool)  if true, it allows to compute a transform
+        that doesn't worth it regarding its complexity compared to the matrix
+        M. Otherwise by default, an exception is raised before the algorithm starts.
 
 
     Returns:
@@ -248,7 +251,7 @@ def eigtj(M, nGivens=None, tol=0, order='ascend', relerr=True,
         demo_path = sep.join((get_data_dirpath(),'Laplacian_256_community.mat'))
         data_dict = loadmat(demo_path)
         Lap = data_dict['Lap'].astype(np.float)
-        Dhat, Uhat = eigtj(Lap, nGivens=Lap.shape[0]*100)
+        Dhat, Uhat = eigtj(Lap, nGivens=Lap.shape[0]*100, enable_large_Faust=True)
         # Uhat is the Fourier matrix/eigenvectors approximation as a Faust
         # (200 factors)
         # Dhat the eigenvalues diagonal matrix approx.
@@ -260,8 +263,7 @@ def eigtj(M, nGivens=None, tol=0, order='ascend', relerr=True,
         # and then asking for an absolute error
         Dhat3, Uhat3 = eigtj(Lap, tol=0.1, relerr=False)
         assert(norm(Lap-Uhat3*np.diag(Dhat3)*Uhat3.H) < .11)
-        # now recompute Uhat2, Dhat2 but asking a descending order of
-        eigenvalues
+        # now recompute Uhat2, Dhat2 but asking a descending order of eigenvalues
         Dhat4, Uhat4 = eigtj(Lap, tol=0.01)
         assert((Dhat4[::-1] == Dhat2[::]).all())
         # and now with no sort
@@ -572,6 +574,7 @@ def hierarchical(M, p, ret_lambda=False, ret_params=False):
        <b> 3. Simplified Parameters for a Rectangular Matrix Factorization
        (the BSL demo MEG matrix)</b>
        >>> from pyfaust import *
+       >>> from pyfaust.fact import hierarchical
        >>> from scipy.io import loadmat
        >>> from pyfaust.demo import get_data_dirpath
        >>> d = loadmat(get_data_dirpath()+'/matrix_MEG.mat')

wrapper/python/pyfaust/factparams.py

@@ -540,6 +540,7 @@ class ParamsFact(ABC):
         self.is_update_way_R2L = is_update_way_R2L
         self.init_lambda = init_lambda
         self.step_size = step_size
+        import pyfaust.proj
         if((isinstance(constraints, list) or isinstance(constraints, tuple))
            and np.array([isinstance(constraints[i],pyfaust.proj.proj_gen) for i in
                     range(0,len(constraints))]).all()):
@@ -578,6 +579,7 @@ class ParamsHierarchical(ParamsFact):
                  is_fact_side_left=False,
                  is_verbose=False,
                  grad_calc_opt_mode=ParamsFact.EXTERNAL_OPT):
+        import pyfaust.proj
         if((isinstance(fact_constraints, list) or isinstance(fact_constraints, tuple))
            and np.array([isinstance(fact_constraints[i],pyfaust.proj.proj_gen) for i in
                     range(0,len(fact_constraints))]).all()):