Spaces:
Sleeping
Sleeping
Delete personal_information.py
Browse files- personal_information.py +0 -76
personal_information.py
DELETED
@@ -1,76 +0,0 @@
|
|
1 |
-
import json
|
2 |
-
import re
|
3 |
-
from openai_utils import get_ai_response
|
4 |
-
from ocr_extractor import process_file
|
5 |
-
from cv_prompt import get_personal_info_prompt
|
6 |
-
|
7 |
-
# Load the scoring data
|
8 |
-
with open('personal_info_scores.json', 'r') as f:
|
9 |
-
score_data = json.load(f)
|
10 |
-
|
11 |
-
def extract_email(text):
|
12 |
-
email_pattern = r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b'
|
13 |
-
emails = re.findall(email_pattern, text)
|
14 |
-
return emails[0] if emails else None
|
15 |
-
|
16 |
-
def extract_phone(text):
|
17 |
-
phone_pattern = r'\b(?:\+?1[-.\s]?)?(?:\(\d{3}\)|\d{3})[-.\s]?\d{3}[-.\s]?\d{4}\b'
|
18 |
-
phones = re.findall(phone_pattern, text)
|
19 |
-
return phones[0] if phones else None
|
20 |
-
|
21 |
-
def extract_location(text):
|
22 |
-
prompt = get_personal_info_prompt(text)
|
23 |
-
messages = [
|
24 |
-
{"role": "user", "content": prompt}
|
25 |
-
]
|
26 |
-
|
27 |
-
response = get_ai_response(messages)
|
28 |
-
|
29 |
-
if response:
|
30 |
-
try:
|
31 |
-
location_data = json.loads(response)
|
32 |
-
city = list(location_data.get('city', {}).keys())[0] if location_data.get('city') else None
|
33 |
-
country = any(location_data.get('country', {}).values())
|
34 |
-
except json.JSONDecodeError:
|
35 |
-
print("Failed to parse JSON from response")
|
36 |
-
city, country = None, False
|
37 |
-
else:
|
38 |
-
city, country = None, False
|
39 |
-
|
40 |
-
return city, country
|
41 |
-
|
42 |
-
def calculate_score(email_exists, phone_exists, city_exists, country_exists):
|
43 |
-
score = 0
|
44 |
-
if email_exists:
|
45 |
-
score += score_data['email']
|
46 |
-
if phone_exists:
|
47 |
-
score += score_data['phone']
|
48 |
-
if city_exists:
|
49 |
-
score += score_data['city']
|
50 |
-
if country_exists:
|
51 |
-
score += score_data['country']
|
52 |
-
return score
|
53 |
-
|
54 |
-
def analyze_personal_info(file_path):
|
55 |
-
text = process_file(file_path, 'ocr_weights.json')
|
56 |
-
|
57 |
-
email = extract_email(text)
|
58 |
-
phone = extract_phone(text)
|
59 |
-
city, country = extract_location(text)
|
60 |
-
|
61 |
-
email_exists = email is not None
|
62 |
-
phone_exists = phone is not None
|
63 |
-
city_exists = city is not None
|
64 |
-
country_exists = country
|
65 |
-
|
66 |
-
score = calculate_score(email_exists, phone_exists, city_exists, country_exists)
|
67 |
-
|
68 |
-
result = {
|
69 |
-
"email": email_exists,
|
70 |
-
"phone": phone_exists,
|
71 |
-
"city": city if city_exists else None,
|
72 |
-
"country": country_exists,
|
73 |
-
"personal_info_score": score
|
74 |
-
}
|
75 |
-
|
76 |
-
return result
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|