--- a
+++ b/deepdta-toy/testdatahelper.py
@@ -0,0 +1,85 @@
+import csv, pickle, json, os, math, sys
+from collections import OrderedDict
+from bioservices import UniProt
+import numpy as np
+import pickle
+
+
+PROT_FILE = "proteins.fasta"
+CHEM_FILE = "ligands.tab"
+AFF_FILE = "Y.tab"
+
+def prepare_new_data(fpath, test=True):
+
+    prots = read_proteins(fpath)
+    chems = read_chemicals(fpath)
+    Y = np.zeros((len(chems), len(prots)))
+
+    if  os.path.exists(fpath + AFF_FILE):
+        Y = np.loadtxt(fpath + AFF_FILE)
+
+    pickle.dump(Y, open(fpath + "Y","wb"), protocol=pickle.HIGHEST_PROTOCOL)
+    label_row_inds, label_col_inds = np.where(np.isnan(Y)==False)
+
+    #json.dump(linepos, open(FLAGS.test_path  + "csv_pos_match.txt","w"))
+    indic = set(range(len(label_row_inds)))
+    indic = sorted(indic, key=os.urandom)
+
+
+    if not os.path.exists(fpath + "folds/"):
+        os.makedirs(fpath + "folds/")
+    if test:
+        json.dump(indic, open(fpath + "folds/test_fold.txt","w"))
+    else:
+        json.dump(indic, open(fpath + "folds/train_fold.txt","w"))
+
+
+def read_chemicals(datafolder):
+    counter =0
+    filepath = datafolder + CHEM_FILE
+    chemicals = {}
+    with open(filepath) as file:
+         next(file)
+         for row in file:
+            chem_id = row.split('\t')[0]
+            smiles = (row.split('\t')[1]).strip()
+            chemicals[chem_id] = smiles
+            counter +=1
+
+    print("%d number(s) of chemical(s)" % counter)
+    json.dump(chemicals, open(datafolder + 'ligands.txt', 'w'))
+
+    return chemicals
+
+
+def read_proteins(datafolder):
+    proteins = {}
+    counter =0
+    fa=""
+    filename = datafolder + PROT_FILE
+    with open(filename) as f:
+        fa = f.readlines()
+
+    idindex=[]
+    for i, line in enumerate(fa):
+        if ">" in line:
+            idindex.append(i)
+    idindex.append(i)
+
+    for i, idx in enumerate(idindex):
+
+        if i < len(idindex)-1:
+            idx1 = idindex[i+1]
+            info = fa[idx].split()
+
+            pid = info[0][4:10]
+            seq = "".join(fa[idx+1:idx1])
+            seq = seq.replace("\n","")
+            proteins[pid] = seq
+            counter +=1
+
+    print("%d number(s) of protein(s)" % counter)
+    json.dump(proteins, open(datafolder + 'proteins.txt', 'w'))
+
+    return proteins
+