normalize before upsampling (nans)

modules/deeplearning/srcnn_l1b_l2.py

@@ -718,6 +718,46 @@ def run_evaluate_static(in_file, out_file, ckpt_dir):
         return out_sr
 
 
+def run_evaluate_static_2(in_file, out_file, ckpt_dir):
+    nda = np.load(in_file)
+    grd_a = nda[:, 0, :, :]
+    grd_a = grd_a[:, 3:131:2, 3:131:2]
+
+    grd_b = nda[:, 2, 3:131, 3:131]
+
+    grd_c = nda[:, 3, :, :]
+    grd_c = grd_c[:, 3:131:2, 3:131:2]
+
+    num, leny, lenx = grd_a.shape
+    x = np.arange(lenx)
+    y = np.arange(leny)
+    x_up = np.arange(0, lenx, 0.5)
+    y_up = np.arange(0, leny, 0.5)
+
+    grd_a = normalize(grd_a, 'temp_11_0um_nom', mean_std_dct)
+    grd_a = resample_2d_linear(x, y, grd_a, x_up, y_up)
+
+    grd_b = normalize(grd_b, 'refl_0_65um_nom', mean_std_dct)
+
+    if label_param == 'cloud_fraction':
+        grd_c = np.where(np.isnan(grd_c), 0, grd_c)
+    else:
+        grd_c = normalize(grd_c, label_param, mean_std_dct)
+    grd_c = resample_2d_linear(x, y, grd_c, x_up, y_up)
+
+    data = np.stack([grd_a, grd_b, grd_c], axis=3)
+    print(data.shape)
+
+    nn = SRCNN()
+    out_sr = nn.run_evaluate(data, ckpt_dir)
+    if label_param != 'cloud_fraction':
+        out_sr = denormalize(out_sr, label_param, mean_std_dct)
+    if out_file is not None:
+        np.save(out_file, out_sr)
+    else:
+        return out_sr
+
+
 def analyze(fpath='/Users/tomrink/clavrx_snpp_viirs.A2019080.0100.001.2019080064252.uwssec_B00038315.level2.h5', param='cloud_fraction'):
     h5f = h5py.File(fpath, 'r')
     grd = get_grid_values_all(h5f, param)