first commit
Browse files- app.py +10 -0
- requirements.txt +4 -0
- utils.py +84 -0
app.py
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
from annotated_text import annotated_text
|
3 |
+
from utils import ner_extraction
|
4 |
+
|
5 |
+
|
6 |
+
input_text = 'Bill Gates lives in USA'
|
7 |
+
|
8 |
+
ner_extraction = ner_extraction(input_text)
|
9 |
+
|
10 |
+
print(ner_extraction.entity_position())
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
streamlit
|
2 |
+
st-annotated-text
|
3 |
+
requests
|
4 |
+
python-dotenv
|
utils.py
ADDED
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import requests
|
2 |
+
from dotenv import load_dotenv
|
3 |
+
import os
|
4 |
+
|
5 |
+
# load the .env file
|
6 |
+
load_dotenv()
|
7 |
+
|
8 |
+
API_KEY = os.getenv("API")
|
9 |
+
|
10 |
+
API_URL = "https://api-inference.huggingface.co/models/Sadashiv/BERT-ner"
|
11 |
+
headers = {"Authorization": f"Bearer {API_KEY}"}
|
12 |
+
|
13 |
+
tag_color_combination = {'O': '#FF5733',
|
14 |
+
'PER': '#35B7FF',
|
15 |
+
'ORG': '#00FF00',
|
16 |
+
'LOC': '#FFA500',
|
17 |
+
'MISC': '#BA55D3'}
|
18 |
+
|
19 |
+
|
20 |
+
class ner_extraction:
|
21 |
+
def __init__(self, input_text):
|
22 |
+
self.input_text = input_text
|
23 |
+
|
24 |
+
def query(self):
|
25 |
+
response = requests.post(API_URL, headers=headers, json=self.input_text)
|
26 |
+
return response.json()
|
27 |
+
|
28 |
+
def entity_position_locator(self):
|
29 |
+
output = self.query()
|
30 |
+
entity_position = {}
|
31 |
+
|
32 |
+
for i in range(len(output)):
|
33 |
+
entity_position[i]={}
|
34 |
+
entity_position[i]["start"]=output[i]['start']
|
35 |
+
entity_position[i]["end"]=output[i]['end']
|
36 |
+
|
37 |
+
return entity_position
|
38 |
+
|
39 |
+
def entity_update(self):
|
40 |
+
entity_list = []
|
41 |
+
output = self.query()
|
42 |
+
|
43 |
+
for i in range(len(output)):
|
44 |
+
entity_list.append(
|
45 |
+
(
|
46 |
+
output[i]['word'],
|
47 |
+
output[i]['entity_group'],
|
48 |
+
tag_color_combination.get(output[i]['entity_group'])
|
49 |
+
)
|
50 |
+
)
|
51 |
+
|
52 |
+
return entity_list
|
53 |
+
|
54 |
+
def text_list(self):
|
55 |
+
|
56 |
+
input_text = self.input_text
|
57 |
+
entity_position = self.entity_position_locator()
|
58 |
+
|
59 |
+
split_text = []
|
60 |
+
|
61 |
+
for i in entity_position:
|
62 |
+
split_text.append(input_text[entity_position[i]['start']:entity_position[i]['end']])
|
63 |
+
|
64 |
+
if entity_position[i]['end']!=len(input_text):
|
65 |
+
|
66 |
+
if i+1<len(entity_position):
|
67 |
+
split_text.append(input_text[entity_position[i]['end']:entity_position[i+1]['start']])
|
68 |
+
|
69 |
+
else:
|
70 |
+
split_text.append(input_text[entity_position[i]['end']:])
|
71 |
+
|
72 |
+
return split_text
|
73 |
+
|
74 |
+
|
75 |
+
def entity_position(self):
|
76 |
+
split_text = self.text_list()
|
77 |
+
entity_list = self.entity_update()
|
78 |
+
for i in range(len(split_text)):
|
79 |
+
for j in range(len(entity_list)):
|
80 |
+
if type(split_text[i])!= tuple:
|
81 |
+
if split_text[i].lower()==entity_list[j][0]:
|
82 |
+
split_text[i]=entity_list[j]
|
83 |
+
|
84 |
+
return tuple(split_text)
|