cleanup before code push

generate_dataset3.py → generate_RML2016.04c.py

@@ -12,9 +12,7 @@
 '''
 
 apply_channel = True
-
 output = {}
-#snr_vals = [-10,-8,-6,-4,-2,0,2,4,6,8,10,12,14,16,18,20]
 min_length = 9e9
 snr_vals = range(-20,20,2)
 for snr in snr_vals:
@@ -30,7 +28,6 @@
                 tx_len = int(20e3)
             src = source_alphabet(alphabet_type, tx_len, True)
             mod = mod_type()
-            #chan = channels.selective_fading_model(8, 20.0/1e6, False, 4.0, 0, (0.0,0.1,1.3), (1,0.99,0.97), 8)
             fD = 1
             delays = [0.0, 0.9, 1.7]
             mags = [1, 0.8, 0.3]
@@ -39,8 +36,6 @@
             print noise_amp
             #noise_amp = 0.1
             chan = channels.dynamic_channel_model( 200e3, 0.01, 1e2, 0.01, 1e3, 8, fD, True, 4, delays, mags, ntaps, noise_amp, 0x1337 )
-            #chan = channels.dynamic_channel_model( 200e3, 0, 1e3, 0, 1e3, 8, fD, True, 4, delays, mags, ntaps, noise_amp, 0x1337 )
-            #chan = channels.dynamic_channel_model( 200e3, 0.1, 1e3, 0.1, 1e3, 8, fD, True, 4, delays, mags, ntaps, noise_amp, 0x1337 )
 
             snk = blocks.vector_sink_c()
 
@@ -59,31 +54,13 @@
                 min_length_mod = mod_type
             output[(mod_type.modname, snr)] = modulated_vector
 
-            #plt.figure()
-            #plt.subplot(2,1,1)
-            x = snk.data()
-            #plt.plot(10*np.log10(numpy.fft.fftshift(numpy.fft.fft(x[0:100000]))))
-            #plt.title("Power Spectrum of Modulated %s"%(mod_type.modname))
-            #plt.subplot(2,1,2)
-            #plt.plot(x[0:100000])
-            #plt.title("Time Plot of Modulated %s"%(mod_type.modname))
-            #plt.savefig('dataset1/%s.png'%(mod_type.modname))
-
 print "min length mod is %s with %i samples" % (min_length_mod, min_length)
 # trim the beginning and ends, and make all mods have equal number of samples
 start_indx = 100
 fin_indx = min_length-100
 for mod, snr in output:
  output[(mod,snr)] = output[(mod,snr)][start_indx:fin_indx]
-#X = timeseries_slicer.slice_timeseries_dict(output, 64, 32, 1000)
 X = timeseries_slicer.slice_timeseries_dict(output, 128, 64, 1000)
-cPickle.dump( X, file("X_3_dict.dat", "wb" ) )
-#cPickle.dump( X, gzip.open("X_1_dict.pkl.gz", "wb" ) )
-#print len(X), X[X.keys()[0]].shape
+cPickle.dump( X, file("RML2014.04c_dict.dat", "wb" ) )
 X = np.vstack(X.values())
-print X.shape
-cPickle.dump( X, file("X_3.dat", "wb" ) )
-#cPickle.dump( X, gzip.open("X_1.pkl.gz", "wb" ) )
-
-#plt.pause(5)
-#plt.show()
+cPickle.dump( X, file("RML2016.04c.dat", "wb" ) )

generate_dataset4.py → generate_RML2016.10a.py

@@ -38,7 +38,6 @@
                   tx_len = int(30e3)
               src = source_alphabet(alphabet_type, tx_len, True)
               mod = mod_type()
-              #chan = channels.selective_fading_model(8, 20.0/1e6, False, 4.0, 0, (0.0,0.1,1.3), (1,0.99,0.97), 8)
               fD = 1
               delays = [0.0, 0.9, 1.7]
               mags = [1, 0.8, 0.3]
@@ -83,15 +82,5 @@
                   insufficient_modsnr_vectors = False
                   # TODO: do a np permuter to shuffle here ?
 
-
-              #plt.figure()
-              #plt.subplot(2,1,1)
-              #x = snk.data()
-              #plt.plot(10*np.log10(numpy.fft.fftshift(numpy.fft.fft(x[0:100000]))))
-              #plt.title("Power Spectrum of Modulated %s"%(mod_type.modname))
-              #plt.subplot(2,1,2)
-              #plt.plot(x[0:100000])
-              #plt.title("Time Plot of Modulated %s"%(mod_type.modname))
-              #plt.savefig('dataset1/%s.png'%(mod_type.modname))
 print "all done. writing to disk"
-cPickle.dump( dataset, file("X_4_dict.dat", "wb" ) )
+cPickle.dump( dataset, file("RML2016.10a_dict.dat", "wb" ) )