Update generic_ner.py

generic_ner.py

@@ -9,6 +9,9 @@ import torch.nn.functional as F
 import re
 
 
+import re, string
+
+
 def tokenize(text):
     # print(text)
     for punctuation in string.punctuation:
@@ -109,14 +112,18 @@ def get_entities(tokens, tags, confidences, text):
                         "score": np.average(confidences[idx : idx + len(subtree)]),
                         "index": (idx, idx + len(subtree)),
                         "word": original_string,
-                        "start": entity_start_position,
+                        "start": entity_start_position-1,
                         "end": entity_end_position,
-                        "text": text,
                     }
                 )
                 # assert (
                 #     text[entity_start_position:entity_end_position] == original_string
                 # )
+                print(
+                    text[entity_start_position:entity_end_position],
+                    "------",
+                    original_string,
+                )
             idx += len(subtree)
 
             # Update the current character position
@@ -148,8 +155,6 @@ def realign(
     return words_list, preds_list, confidence_list
 
 
-import re, string
-
 # List of additional "strange" punctuation marks
 additional_punctuation = "‘’“”„«»•–—―‣◦…§¶†‡‰′″〈〉"