Spaces:
Sleeping
Sleeping
Commit
·
06001d1
1
Parent(s):
788312d
update
Browse files
app.py
ADDED
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from transformers import pipeline
|
2 |
+
|
3 |
+
get_completion = pipeline("ner", model="dslim/bert-base-NER")
|
4 |
+
|
5 |
+
def ner(input):
|
6 |
+
output = get_completion(input)
|
7 |
+
return {"text": input, "entities": output}
|
8 |
+
|
9 |
+
def merge_tokens(tokens):
|
10 |
+
'''
|
11 |
+
WHAT: Faz um loop entre os tokens para concatenar os tokens
|
12 |
+
com entidades I-* (intermediate token) aos B-* (begining token).
|
13 |
+
|
14 |
+
'''
|
15 |
+
merged_tokens = []
|
16 |
+
for token in tokens:
|
17 |
+
if merged_tokens and token['entity'].startswith('I-') and merged_tokens[-1]['entity'].endswith(token['entity'][2:]):
|
18 |
+
# Se a lista merged_tokens não estiver vazia.
|
19 |
+
# o token atual for um token intermediário (começa com 'I-').
|
20 |
+
# a entidade do último token processado terminar com a mesma entidade do token atual, excluindo o prefixo 'I-'.
|
21 |
+
|
22 |
+
last_token = merged_tokens[-1]
|
23 |
+
last_token['word'] += token['word'].replace('##', '')
|
24 |
+
last_token['end'] = token['end']
|
25 |
+
last_token['score'] = (last_token['score'] + token['score']) / 2
|
26 |
+
else:
|
27 |
+
merged_tokens.append(token)
|
28 |
+
|
29 |
+
return merged_tokens
|
30 |
+
|
31 |
+
def ner(input):
|
32 |
+
"""
|
33 |
+
WHAT: Aplica a task de NER com o hugginface pipeline e concatena os tokens com a mesma entidade.
|
34 |
+
RETURN: retorna um dicionário com o token e suas entidades.
|
35 |
+
"""
|
36 |
+
output = get_completion(input)
|
37 |
+
merged_tokens = merge_tokens(output)
|
38 |
+
return {"text": input, "entities": merged_tokens}
|
39 |
+
|
40 |
+
gr.close_all()
|
41 |
+
demo = gr.Interface(fn=ner,
|
42 |
+
inputs=[gr.Textbox(label="Text to find entities", lines=2)],
|
43 |
+
outputs=[gr.HighlightedText(label="Text with entities")],
|
44 |
+
title="NER with dslim/bert-base-NER",
|
45 |
+
description="Find entities using the `dslim/bert-base-NER` model under the hood!",
|
46 |
+
allow_flagging="never",
|
47 |
+
examples=["My name is Andrew, I'm building DeeplearningAI and I live in California", "My name is Poli, I live in Vienna and work at HuggingFace"])
|
48 |
+
|
49 |
+
demo.launch()
|