Merge pull request #128 from carlgogo/master

Update to ``svd_wrapper``, pytorch support. Removed ascii logo when importing vip

docs/source/readme.rst

@@ -21,6 +21,22 @@ VIP - Vortex Image Processing package
     :target: http://vip.readthedocs.io/en/latest/?badge=latest
 
 
+
+::
+
+    ---------------------------------------------------
+            oooooo     oooo ooooo ooooooooo.
+             `888.     .8'  `888' `888   `Y88.
+              `888.   .8'    888   888   .d88'
+               `888. .8'     888   888ooo88P'
+                `888.8'      888   888
+                 `888'       888   888
+                  `8'       o888o o888o
+    ---------------------------------------------------
+             Vortex Image Processing package
+    ---------------------------------------------------
+
+
 Introduction
 ------------
 

docs/source/vip_hci.leastsq.rst

@@ -0,0 +1,20 @@
+vip\_hci\.leastsq package
+=========================
+
+.. automodule:: vip_hci.leastsq
+    :members:
+    :undoc-members:
+    :show-inheritance:
+
+Submodules
+----------
+
+vip\_hci\.leastsq\.leastsq module
+---------------------------------
+
+.. automodule:: vip_hci.leastsq.leastsq
+    :members:
+    :undoc-members:
+    :show-inheritance:
+
+

docs/source/vip_hci.madi.rst

@@ -17,4 +17,12 @@ vip\_hci\.madi\.adi\_source module
     :undoc-members:
     :show-inheritance:
 
+vip\_hci\.madi\.adi\_utils module
+---------------------------------
+
+.. automodule:: vip_hci.madi.adi_utils
+    :members:
+    :undoc-members:
+    :show-inheritance:
+
 

docs/source/vip_hci.rst

@@ -15,6 +15,7 @@ Subpackages
     vip_hci.exlib
     vip_hci.fits
     vip_hci.frdiff
+    vip_hci.leastsq
     vip_hci.llsg
     vip_hci.madi
     vip_hci.negfc

readme.rst

@@ -21,6 +21,22 @@ VIP - Vortex Image Processing package
     :target: http://vip.readthedocs.io/en/latest/?badge=latest
 
 
+
+::
+
+    ---------------------------------------------------
+            oooooo     oooo ooooo ooooooooo.
+             `888.     .8'  `888' `888   `Y88.
+              `888.   .8'    888   888   .d88'
+               `888. .8'     888   888ooo88P'
+                `888.8'      888   888
+                 `888'       888   888
+                  `8'       o888o o888o
+    ---------------------------------------------------
+             Vortex Image Processing package
+    ---------------------------------------------------
+
+
 Introduction
 ------------
 

vip_hci/__init__.py

@@ -17,15 +17,5 @@
 __version__ = "0.9.2"
 
 
-print("---------------------------------------------------")
-print("        oooooo     oooo ooooo ooooooooo.           ")
-print("         `888.     .8'  `888' `888   `Y88.         ")
-print("          `888.   .8'    888   888   .d88'         ")
-print("           `888. .8'     888   888ooo88P'          ")
-print("            `888.8'      888   888                 ")
-print("             `888'       888   888                 ")
-print("              `8'       o888o o888o                ")
-print("---------------------------------------------------")
-print("     Vortex Image Processing package v"+__version__)
-print("---------------------------------------------------")
+
 

vip_hci/conf/utils_conf.py

@@ -8,7 +8,7 @@
 from __future__ import print_function
 
 __author__ = 'Carlos Alberto Gomez Gonzalez'
-__all__ = ['eval_func_tuple', 'redirect_output', 'sep']
+__all__ = []
 
 import sys
 

vip_hci/frdiff/framediff.py

@@ -20,7 +20,7 @@
 from ..conf import time_ini, timing
 from ..pca.utils_pca import pca_annulus
 from ..madi.adi_utils import _compute_pa_thresh, _find_indices, _define_annuli
-from ..conf import eval_func_tuple as EFT
+from ..conf.utils_conf import eval_func_tuple as EFT
 
 
 array = None

vip_hci/leastsq/leastsq.py

@@ -20,7 +20,7 @@
 from ..preproc import cube_derotate, cube_collapse, check_pa_vector
 from ..conf import time_ini, timing
 from ..madi.adi_utils import _find_indices, _define_annuli
-from ..conf import eval_func_tuple as EFT
+from ..conf.utils_conf import eval_func_tuple as EFT
 
 
 array = None

vip_hci/llsg/llsg.py

@@ -20,7 +20,7 @@
 from ..var import get_annulus_segments, cube_filter_highpass
 from ..pca.svd import svd_wrapper, get_eigenvectors
 from .thresholding import thresholding
-from ..conf import eval_func_tuple as EFT
+from ..conf.utils_conf import eval_func_tuple as EFT
 
 
 cube_init = None

vip_hci/madi/adi_source.py

@@ -16,7 +16,7 @@
 from ..conf import time_ini, timing
 from ..var import get_annulus, mask_circle
 from ..preproc import cube_derotate, cube_collapse, check_pa_vector
-from ..conf import eval_func_tuple as EFT
+from ..conf.utils_conf import eval_func_tuple as EFT
 from .adi_utils import _find_indices, _define_annuli
 
 

vip_hci/negfc/mcmc_sampling.py

@@ -25,7 +25,8 @@
 from scipy.stats import norm
 from ..fits import open_adicube, open_fits
 from ..phot import cube_inject_companions
-from ..conf import time_ini, timing, sep
+from ..conf import time_ini, timing
+from ..conf.utils_conf import sep
 from .simplex_fmerit import get_values_optimize
 import warnings
 warnings.filterwarnings("ignore", category=DeprecationWarning)

vip_hci/negfc/simplex_optim.py

@@ -11,7 +11,9 @@
 from scipy.optimize import minimize
 from .simplex_fmerit import chisquare
 from ..var import frame_center
-from ..conf import time_ini, timing, sep
+from ..conf import time_ini, timing
+from ..conf.utils_conf import sep
+
 
 __all__ = ['firstguess']
 

vip_hci/pca/pca_local.py

@@ -16,7 +16,7 @@
 from multiprocessing import Pool, cpu_count
 from ..preproc import cube_derotate, cube_collapse, check_pa_vector
 from ..conf import time_ini, timing
-from ..conf import eval_func_tuple as EFT 
+from ..conf.utils_conf import eval_func_tuple as EFT
 from ..var import get_annulus_segments, matrix_scaling, get_annulus
 from ..madi.adi_utils import _find_indices, _define_annuli
 from ..stats import descriptive_stats

vip_hci/pca/svd.py

@@ -42,7 +42,7 @@
 
 
 def svd_wrapper(matrix, mode, ncomp, debug, verbose, usv=False,
-                random_state=None):
+                random_state=None, to_numpy=True):
     """ Wrapper for different SVD libraries (CPU and GPU). 
       
     Parameters
@@ -76,9 +76,17 @@ def svd_wrapper(matrix, mode, ncomp, debug, verbose, usv=False,
         If True the explained variance ratio is computed and displayed.
     verbose: bool
         If True intermediate information is printed out.
-    usv : {False, True}, bool optional
+    usv : bool optional
         If True the 3 terms of the SVD factorization are returned.
-    
+    random_state : int, RandomState instance or None, optional
+        If int, random_state is the seed used by the random number generator.
+        If RandomState instance, random_state is the random number generator.
+        If None, the random number generator is the RandomState instance used
+        by np.random. Used for ``randsvd`` mode.
+    to_numpy : bool, optional
+        If True (by default) the arrays computed in GPU are transferred from
+        VRAM and converted to numpy ndarrays.
+
     Returns
     -------
     The right singular vectors of the input matrix. If ``usv`` is True it 
@@ -250,22 +258,26 @@ def reconstruction(ncomp, U, S, V, var=1):
         u_gpu, s_gpu, vh_gpu = cupy.linalg.svd(a_gpu, full_matrices=True,
                                                compute_uv=True)
         V = vh_gpu[:ncomp]
-        V = cupy.asnumpy(V)
+        if to_numpy:
+            V = cupy.asnumpy(V)
         if usv:
             S = s_gpu[:ncomp]
-            S = cupy.asnumpy(S)
+            if to_numpy:
+                S = cupy.asnumpy(S)
             U = u_gpu[:, :ncomp]
-            U = cupy.asnumpy(U)
+            if to_numpy:
+                U = cupy.asnumpy(U)
         if verbose:
             print('Done SVD/PCA with cupy (GPU)')
 
     elif mode == 'randcupy':
         if no_cupy:
             raise RuntimeError('Cupy is not installed')
         U, S, V = randomized_svd_gpu(matrix, ncomp, n_iter=2, lib='cupy')
-        V = cupy.asnumpy(V)
-        S = cupy.asnumpy(S)
-        U = cupy.asnumpy(U)
+        if to_numpy:
+            V = cupy.asnumpy(V)
+            S = cupy.asnumpy(S)
+            U = cupy.asnumpy(U)
         if debug:
             reconstruction(ncomp, U, S, V)
         if verbose:
@@ -286,43 +298,52 @@ def reconstruction(ncomp, U, S, V, var=1):
         for i in range(V.shape[1]):
             V[:, i] /= S                    # scaling by the square root of eigenvalues
         V = V[:ncomp]
-        V = cupy.asnumpy(V)
-        S = cupy.asnumpy(S)
+        if to_numpy:
+            V = cupy.asnumpy(V)
         if verbose:
             print('Done PCA with cupy eigh function (GPU)')
 
     elif mode == 'pytorch':
         if no_torch:
             raise RuntimeError('Pytorch is not installed')
-        a_gpu = torch.Tensor.cuda(torch.from_numpy(matrix.T))
+        a_gpu = torch.Tensor.cuda(torch.from_numpy(matrix.astype('float32').T))
         u_gpu, s_gpu, vh_gpu = torch.svd(a_gpu)
-        V = np.array(vh_gpu)[:ncomp]
-        S = np.array(s_gpu)[:ncomp]
-        U = np.array(u_gpu)[:, :ncomp]
+        V = vh_gpu[:ncomp]
+        S = s_gpu[:ncomp]
+        U = torch.transpose(u_gpu, 0, 1)[:ncomp]
+        if to_numpy:
+            V = np.array(V)
+            S = np.array(S)
+            U = np.array(U)
         if verbose:
             print('Done SVD/PCA with pytorch (GPU)')
 
     elif mode == 'eigenpytorch':
         if no_torch:
             raise RuntimeError('Pytorch is not installed')
-        a_gpu = torch.Tensor.cuda(torch.from_numpy(matrix))
+        a_gpu = torch.Tensor.cuda(torch.from_numpy(matrix.astype('float32')))
         C = torch.mm(a_gpu, torch.transpose(a_gpu, 0, 1))
         e, EV = torch.eig(C, eigenvectors=True)
-        pc = torch.mm(torch.transpose(EV, 0, 1), a_gpu)
-        V = np.array(pc)
-        S = np.array(torch.sqrt(e[:, 0]))
+        V = torch.mm(torch.transpose(EV, 0, 1), a_gpu)
+        S = torch.sqrt(e[:, 0])
         if debug:
             reconstruction(ncomp, None, S, None)
         for i in range(V.shape[1]):
             V[:, i] /= S
-        V = np.array(V)[:ncomp]
+        V = V[:ncomp]
+        if to_numpy:
+            V = np.array(V)
         if verbose:
             print('Done PCA with pytorch eig function')
 
     elif mode == 'randpytorch':
         if no_torch:
             raise RuntimeError('Pytorch is not installed')
         U, S, V = randomized_svd_gpu(matrix, ncomp, n_iter=2, lib='pytorch')
+        if to_numpy:
+            V = np.array(V)
+            S = np.array(S)
+            U = np.array(U)
         if debug:
             reconstruction(ncomp, U, S, V)
         if verbose:
@@ -332,13 +353,20 @@ def reconstruction(ncomp, U, S, V, var=1):
         raise ValueError('The SVD mode is not available')
 
     if usv:
-        if mode == 'lapack' or mode=='pytorch':
+        if mode == 'lapack':
             return V.T, S, U.T
+        elif mode == 'pytorch':
+            if to_numpy:
+                return V.T, S, U.T
+            else:
+                return torch.transpose(V, 0, 1), S, torch.transpose(U, 0, 1)
         else:
             return U, S, V
     else:
-        if mode == 'lapack' or mode=='pytorch':
+        if mode == 'lapack':
             return U.T
+        elif mode == 'pytorch':
+            return U
         else:
             return V
 
@@ -512,22 +540,22 @@ def randomized_svd_gpu(M, n_components, n_oversamples=10, n_iter='auto',
             return U[:, :n_components], s[:n_components], V[:n_components, :]
 
     elif lib == 'pytorch':
-        M = torch.Tensor.cuda(torch.from_numpy(M.astype('float32')))
+        M_gpu = torch.Tensor.cuda(torch.from_numpy(M.astype('float32')))
 
         # Generating normal random vectors with shape: (M.shape[1], n_random)
-        Q = torch.cuda.FloatTensor(M.shape[1], n_random).normal_()
+        Q = torch.cuda.FloatTensor(M_gpu.shape[1], n_random).normal_()
 
         # Perform power iterations with Q to further 'imprint' the top
         # singular vectors of M in Q
         for i in range(n_iter):
-            Q = torch.mm(M, Q)
-            Q = torch.mm(torch.transpose(M, 0, 1), Q)
+            Q = torch.mm(M_gpu, Q)
+            Q = torch.mm(torch.transpose(M_gpu, 0, 1), Q)
 
         # Sample the range of M using by linear projection of Q. Extract an orthonormal basis
-        Q, _ = torch.qr(torch.mm(M, Q))
+        Q, _ = torch.qr(torch.mm(M_gpu, Q))
 
         # project M to the (k + p) dimensional space using the basis vectors
-        B = torch.mm(torch.transpose(Q, 0, 1), M)
+        B = torch.mm(torch.transpose(Q, 0, 1), M_gpu)
 
         # compute the SVD on the thin matrix: (k + p) wide
         Uhat, s, V = torch.svd(B)

vip_hci/phot/contrcurve.py

@@ -22,7 +22,8 @@
 from skimage.draw import circle
 from matplotlib import pyplot as plt
 from .fakecomp import cube_inject_companions, frame_inject_companion, psf_norm
-from ..conf import time_ini, timing, sep
+from ..conf import time_ini, timing
+from ..conf.utils_conf import sep
 from ..var import frame_center, dist
 
 

vip_hci/phot/detection.py

@@ -24,7 +24,7 @@
 from skimage.feature import peak_local_max
 from ..var import (mask_circle, pp_subplots, get_square, frame_center,
                    frame_filter_gaussian2d, fit_2dgaussian)
-from ..conf import sep
+from ..conf.utils_conf import sep
 from .snr import snr_ss
 from .frame_analysis import frame_quick_report
 

vip_hci/phot/frame_analysis.py

@@ -14,7 +14,7 @@
 import photutils
 from skimage import draw
 from .snr import snr_ss
-from ..conf import sep
+from ..conf.utils_conf import sep
 
 
 def frame_quick_report(array, fwhm, source_xy=None, verbose=True):

vip_hci/phot/snr.py

@@ -21,7 +21,8 @@
 from astropy.convolution import convolve, Tophat2DKernel
 from astropy.stats import median_absolute_deviation as mad
 from multiprocessing import Pool, cpu_count
-from ..conf import eval_func_tuple, time_ini, timing
+from ..conf.utils_conf import eval_func_tuple
+from ..conf import time_ini, timing
 from ..var import get_annulus, frame_center, dist, pp_subplots