add chunk merging

utils.py

@@ -139,20 +139,23 @@ def merge_two_chunk(chunk_1, chunk_2, overlap, debug=False):
             idx_list_1.extend([idx] * len(extract_phrase_word(phrase)))
         for idx, phrase in enumerate(list_2):
             idx_list_2.extend([idx] * len(extract_phrase_word(phrase)))
+        # print(idx_list_1, idx_list_2)
         for idx, (idx_1, idx_2) in enumerate(zip(idx_list_1, idx_list_2)):
             if list_1[idx_1].startswith('<delete>') or list_2[idx_2].startswith('<delete>'):
                 continue
-            elif is_equal(list_1[idx_1], list_2[idx_2]) and '1_{}'.format(idx_1) not in mark_term_complete and '2_{}'.format(idx_2) not in mark_term_complete:
-                if idx <= overlap//2:
-                    combine_phrases.append(list_1[idx_1])
-                    mark_term_complete.append('1_{}'.format(idx_1))
-                else:
-                    combine_phrases.append(list_2[idx_2])
-                    mark_term_complete.append('2_{}'.format(idx_2))
+            elif is_equal(list_1[idx_1], list_2[idx_2]):
+                # print(list_1[idx_1])
+                if '1_{}'.format(idx_1) not in mark_term_complete and '2_{}'.format(idx_2) not in mark_term_complete:
+                    if idx <= overlap//2:
+                        combine_phrases.append(list_1[idx_1])
+                        mark_term_complete.append('1_{}'.format(idx_1))
+                    else:
+                        combine_phrases.append(list_2[idx_2])
+                        mark_term_complete.append('2_{}'.format(idx_2))
             else:
                 combine_phrases.append(list_raw[idx])
                 mark_term_complete.extend(['1_{}'.format(idx_1), '2_{}'.format(idx_2)])
-
+        # print(mark_term_complete)
         return combine_phrases
 
     remain_1, compete_1 = extract_compete_region(chunk_1, is_head=True)