Revert "Revert "Added format conversion""

Author: hypnopump

models/distance_pipeline/Tutorials/README.pdf

@@ -0,0 +1,14 @@
+RR_FORMAT = """PFRMAT RR
+MODEL 1
+"""
+def save_rr(seq,sample_pred,file_name_rr):
+  my_file=open(file_name_rr,'w')
+  my_file.write(RR_FORMAT)
+  my_file.write(seq + '\n')
+  for i in range(0,len(seq)):
+    for j in range(i+5,len(seq)):
+      print(max(sample_pred[0][i][j]))
+      my_file.write(str(i+1)+" "+ str(j+1)+" "+"0 8 "+ str(max(sample_pred[0][i][j]))+"\n")
+  my_file.write('END\n')
+  my_file.close()
+save_rr(seq,sample_pred,'file_name_rr')

models/distance_pipeline/Tutorials/RR_format.py

@@ -0,0 +1,23 @@
+with open("file_name.pssm", "r") as f:
+    lines = f.readlines()[2:-6]
+key  = "ACDEFGHIKLMNPQRSTVWY"
+text_keys = lines[0].replace("\n", "").split("   ")[4:]
+
+# PSSM OPTION 1
+text_vals = np.array([" ".join(line.replace("\n", "")[72:-11].split()).split() for line in lines[1:]]).astype(float)
+
+# PSSM OPTION 2
+# text_vals = np.array([" ".join(line.replace("\n", "")[10:72].split()).split() for line in lines[1:]]).astype(float)
+
+# normalize to [0,1]
+# text_vals = text_vals / np.sum(text_vals, axis=1).reshape((len(text_vals), 1))
+for i in range(len(text_vals)):
+    text_vals[i] = (text_vals[i] - np.amin(text_vals[i])) / (np.amax(text_vals[i] - np.amin(text_vals[i])))
+
+# create NxL PSSM
+pssm      = np.zeros_like(text_vals)
+for i,aa in enumerate(text_keys):
+    pssm[key.index(aa), :] = text_vals[i, :] 
+    
+inputs_pssm = wider_pssm(pssm.T, seq)
+inputs = np.concatenate((inputs_aa, inputs_pssm), axis=-1)