Added recent SPARCO example

demos/example_image_reconstruction_sparco_grey_v1295.jl

@@ -3,34 +3,35 @@
 
 using OITOOLS
 oifitsfile = "./data/2019_v1295Aql.WL_SMOOTH.A.oifits"
-pixsize = 0.25 #mas/pixel
-nx = 32 # pixels
+pixsize = 0.125 #mas/pixel
+nx = 64 # pixels
 data = readoifits(oifitsfile, filter_bad_data=true)[1,1];
 ft = setup_nfft(data, nx, pixsize);
 params_start = [0.444, 0, 0, 4, 1.6e-6] #V2:12.36 T3A: 4.14 T3P: 1.61
 x_start = rand(nx,nx)
 x_start /= sum(x_start)
-chi2_sparco_f(x_start,  params_start, ft, data, weights=[1.0,0.0,1.0])
+weights=[1.0, 0.0, 1.0] # For this dataset, the T3amp are a bit fishy
+chi2_sparco_f(x_start,  params_start, ft, data, weights=weights)
 
 # Initial image, no regularization except positivity
 regularizers = []
 params_start = [0.5, 0, 0, 4, 1.6e-6] 
 params, x = reconstruct_sparco_gray(x_start, params_start, data, ft, regularizers=regularizers, weights=[1.0, 0.0, 1.0], verb=true, maxiter=200); #grey environment
-minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = [1.0,1.0,1.0] )
+minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = weights )
 params, x = reconstruct_sparco_gray(x, params, data, ft, regularizers=regularizers, weights=[1.0, 0.0, 1.0], verb=true, maxiter=200); #grey environment
 imdisp(x, pixsize = pixsize, figtitle="v1295 Aql - A")
+# The idea is to slowly decrease the regularization level
 regularizers = [["tvsq", 1e7]]
 params, x = reconstruct_sparco_gray(x, params, data, ft, regularizers=regularizers, weights=[1.0, 0.0, 1.0], verb=true, maxiter=200); #grey environment
-minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = [1.0,1.0,1.0] )
+minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = weights )
 regularizers = [["tvsq", 1e5]]
 params, x = reconstruct_sparco_gray(x, params, data, ft, regularizers=regularizers, weights=[1.0, 0.0, 1.0], verb=true, maxiter=200); #grey environment
-minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = [1.0,1.0,1.0] )
+minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = weights )
 regularizers = [["tvsq", 1e3]]
 params, x = reconstruct_sparco_gray(x, params, data, ft, regularizers=regularizers, weights=[1.0, 0.0, 1.0], verb=true, maxiter=200); #grey environment
-minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = [1.0,1.0,1.0] )
+minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = weights )
 regularizers = [["tvsq", 1e1]]
 params, x = reconstruct_sparco_gray(x, params, data, ft, regularizers=regularizers, weights=[1.0, 0.0, 1.0], verb=true, maxiter=200); #grey environment
-minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = [1.0,1.0,1.0] )
+minchi2, params,ret = optimize_sparco_parameters(params_start, x, ft, data; weights = weights )
 imdisp(x, pixsize = pixsize, figtitle="v1295 Aql - A")
-g=similar([params_start;x_start])
-chi2_f = chi2_sparco_fg([params_start;x_start],  g, ft, data, 5, weights=[1.0, 0.0, 1.0])
+chi2_sparco_f(x,  params, ft, data, weights=weights)