add dict utility, functions not pruned

Former-commit-id: cce40bebec78c4bd8ab0ae7fa3adf76edfa54f2b

dict_util.py

@@ -0,0 +1,46 @@
+import csv
+import pickle
+
+### NOTICE: csv only accept two colomn input. but accept multi-time input.
+
+
+# 1_2_3, 1 is action, 2 is supply object, 3 is accept object
+def update_dict_csv(term_dict, f):
+    for rows in csv.reader(f):
+        if rows[0] in term_dict:
+            if rows[1] not in term_dict[rows[0]]:
+                term_dict[rows[0]] = term_dict[rows[0]]+[rows[1]]
+        else:
+            term_dict[rows[0]]=[rows[1]]
+    pass
+
+def export_dict_csv(term_dict, f):
+    for key, val in term_dict.items():
+        csv.writer(f).writerow([key, val])
+    pass
+
+def save_dict_pickle(term_dict, f):
+    pickle.dump(term_dict, f, pickle.HIGHEST_PROTOCOL)
+
+def update_csv_pickle(pickle_f, csv_f):
+    term_dict = pickle.load(pickle_f)
+    for rows in csv.reader(csv_f):
+        if rows[0] in term_dict:
+            if rows[1] not in term_dict[rows[0]]:
+                term_dict[rows[0]] = term_dict[rows[0]]+[rows[1]]
+        else:
+            term_dict[rows[0]]=[rows[1]]
+    #save to pickle file, highest protocal to get better performance
+    pickle.dump(term_dict, pickle_f, pickle.HIGHEST_PROTOCOL)
+
+
+#demo
+term_dict_sc2 = {}
+with open("./finetune_data/dict_enzh.csv", 'r', encoding='utf-8') as f:
+    update_dict_csv(term_dict_sc2,f)
+
+with open("../test.csv", "w", encoding='utf-8') as w:
+    export_dict_csv(term_dict_sc2,w)
+
+## for load pickle, just:
+# pickle.load(f)