Spaces:
Sleeping
Sleeping
Upload 5 files
Browse files- N_F_skill_output.xlsx +0 -0
- app.py +161 -0
- education_output.xlsx +0 -0
- jobs_output.xlsx +0 -0
- requirements.txt +4 -0
N_F_skill_output.xlsx
ADDED
Binary file (20.4 kB). View file
|
|
app.py
ADDED
@@ -0,0 +1,161 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import pandas as pd
|
2 |
+
import re
|
3 |
+
import json
|
4 |
+
from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline
|
5 |
+
import gradio as gr
|
6 |
+
|
7 |
+
def read_from_excel(file_path):
|
8 |
+
df = pd.read_excel(file_path)
|
9 |
+
items = df['object'].astype(str).tolist() # تبدیل همه مقادیر به رشته
|
10 |
+
return items
|
11 |
+
|
12 |
+
def preprocess_text(text):
|
13 |
+
# حذف کاراکترهای غیرضروری و نرمالسازی متن
|
14 |
+
text = text.replace('\u200c', ' ').strip() # حذف نیمفاصله و فاصلههای اضافی
|
15 |
+
text = re.sub(r'\s+', ' ', text) # حذف فاصلههای تکراری
|
16 |
+
return text
|
17 |
+
|
18 |
+
def extract_items_in_text(text, items):
|
19 |
+
text = preprocess_text(text)
|
20 |
+
found_items = set() # استفاده از مجموعه برای جلوگیری از تکرار
|
21 |
+
for item in items:
|
22 |
+
item_normalized = preprocess_text(item)
|
23 |
+
if item_normalized.lower() in text.lower():
|
24 |
+
found_items.add(item_normalized)
|
25 |
+
return list(found_items)
|
26 |
+
|
27 |
+
def compare_items(items_1, items_2):
|
28 |
+
common_items = set()
|
29 |
+
score = 0 # مقدار پیشفرض برای score
|
30 |
+
for item1 in items_1:
|
31 |
+
for item2 in items_2:
|
32 |
+
words1 = set(item1.lower().split())
|
33 |
+
words2 = set(item2.lower().split())
|
34 |
+
common_words = words1.intersection(words2)
|
35 |
+
num_common = len(common_words)
|
36 |
+
|
37 |
+
if num_common >= 3:
|
38 |
+
common_items.add((item1, item2))
|
39 |
+
score = 100
|
40 |
+
elif num_common == 2:
|
41 |
+
common_items.add((item1, item2))
|
42 |
+
score = 75
|
43 |
+
elif num_common == 1:
|
44 |
+
common_items.add((item1, item2))
|
45 |
+
score = 50
|
46 |
+
|
47 |
+
return score, common_items
|
48 |
+
|
49 |
+
def compare_skills(skill_1, skill_2):
|
50 |
+
common_skill = set(skill_1).intersection(set(skill_2))
|
51 |
+
num_common = len(common_skill)
|
52 |
+
|
53 |
+
if num_common >= 10:
|
54 |
+
score = 100
|
55 |
+
elif num_common == 7:
|
56 |
+
score = 75
|
57 |
+
elif num_common == 5:
|
58 |
+
score = 50
|
59 |
+
else:
|
60 |
+
score = 25
|
61 |
+
|
62 |
+
return score, common_skill
|
63 |
+
|
64 |
+
def extract_ner_info(text, nlp):
|
65 |
+
ner_results = nlp(text)
|
66 |
+
full_name = ''
|
67 |
+
loc = ''
|
68 |
+
age = None
|
69 |
+
|
70 |
+
for i in range(len(ner_results)):
|
71 |
+
if ner_results[i]['entity'] == 'B-PER':
|
72 |
+
full_name = ner_results[i]['word']
|
73 |
+
for j in range(i+1, len(ner_results)):
|
74 |
+
if ner_results[j]['entity'].startswith('I-PER'):
|
75 |
+
full_name += ner_results[j]['word'].replace('##', '')
|
76 |
+
else:
|
77 |
+
break
|
78 |
+
|
79 |
+
if ner_results[i]['entity'] == 'B-LOC' and not loc:
|
80 |
+
loc = ner_results[i]['word']
|
81 |
+
|
82 |
+
age_match = re.search(r'سن\s*:\s*(\d+)', text)
|
83 |
+
if age_match:
|
84 |
+
age = int(age_match.group(1))
|
85 |
+
|
86 |
+
return full_name, loc, age
|
87 |
+
|
88 |
+
def process_text(input_text):
|
89 |
+
# مسیر فایل اکسلها را وارد کنید
|
90 |
+
job_excel_file_path = 'jobs_output.xlsx'
|
91 |
+
education_excel_file_path = 'education_output.xlsx'
|
92 |
+
skills_excel_file_path = 'N_F_skill_output.xlsx'
|
93 |
+
# خواندن شغلها، تحصیلات و مهارتها از فایلهای اکسل
|
94 |
+
jobs = read_from_excel(job_excel_file_path)
|
95 |
+
education = read_from_excel(education_excel_file_path)
|
96 |
+
skills = read_from_excel(skills_excel_file_path)
|
97 |
+
|
98 |
+
# متن ثابت
|
99 |
+
fixed_text = """استخدام کارآموز هوش مصنوعی (AI-شیراز)"""
|
100 |
+
|
101 |
+
input_text = input_text.replace("آدرس", "")
|
102 |
+
# استخراج شغلها، تحصیلات و مهارتها از متنها
|
103 |
+
jobs_in_fixed_text = extract_items_in_text(fixed_text, jobs)
|
104 |
+
jobs_in_input_text = extract_items_in_text(input_text, jobs)
|
105 |
+
education_in_fixed_text = extract_items_in_text(fixed_text, education)
|
106 |
+
education_in_input_text = extract_items_in_text(input_text, education)
|
107 |
+
skills_in_fixed_text = extract_items_in_text(fixed_text, skills)
|
108 |
+
skills_in_input_text = extract_items_in_text(input_text, skills)
|
109 |
+
|
110 |
+
# مقایسه و نمرهدهی
|
111 |
+
job_score, common_jobs = compare_items(jobs_in_fixed_text, jobs_in_input_text)
|
112 |
+
education_score, common_education = compare_items(education_in_fixed_text, education_in_input_text)
|
113 |
+
skill_score, common_skills = compare_skills(skills_in_fixed_text, skills_in_input_text)
|
114 |
+
|
115 |
+
# تنظیم و آمادهسازی مدل NER
|
116 |
+
model_name_or_path = "HooshvareLab/distilbert-fa-zwnj-base-ner"
|
117 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
|
118 |
+
model = AutoModelForTokenClassification.from_pretrained(model_name_or_path) # Pytorch
|
119 |
+
nlp = pipeline("ner", model=model, tokenizer=tokenizer)
|
120 |
+
|
121 |
+
# استخراج اطلاعات NER
|
122 |
+
full_name, loc, age = extract_ner_info(input_text, nlp)
|
123 |
+
|
124 |
+
# نمرهدهی لوکیشن
|
125 |
+
fixed_loc = "شیراز"
|
126 |
+
loc_score = 100 if loc == fixed_loc else 0
|
127 |
+
|
128 |
+
# نمرهدهی سن
|
129 |
+
age_score = 100 if age and 18 <= age <= 30 else 0
|
130 |
+
|
131 |
+
# محاسبه و نمایش میانگین نمرات
|
132 |
+
average_score = (job_score + education_score + skill_score + loc_score + age_score) / 5
|
133 |
+
|
134 |
+
# ساخت خروجی JSON
|
135 |
+
output = {
|
136 |
+
"average_score": average_score,
|
137 |
+
"full_name": full_name,
|
138 |
+
"age": age,
|
139 |
+
"location": loc,
|
140 |
+
"job_score": job_score,
|
141 |
+
"education_score": education_score,
|
142 |
+
"skill_score": skill_score,
|
143 |
+
"loc_score": loc_score,
|
144 |
+
"age_score": age_score,
|
145 |
+
"common_jobs": list(common_jobs),
|
146 |
+
"common_education": list(common_education),
|
147 |
+
"common_skills": list(common_skills)
|
148 |
+
}
|
149 |
+
|
150 |
+
return json.dumps(output, ensure_ascii=False, indent=4)
|
151 |
+
|
152 |
+
iface = gr.Interface(
|
153 |
+
fn=process_text,
|
154 |
+
inputs=gr.inputs.Textbox(lines=10, placeholder="لطفاً متن خود را وارد کنید..."),
|
155 |
+
outputs="json",
|
156 |
+
title="متن پرداز",
|
157 |
+
description="این ابزار متن شما را پردازش کرده و امتیازات مشابهت را محاسبه میکند."
|
158 |
+
)
|
159 |
+
|
160 |
+
if __name__ == "__main__":
|
161 |
+
iface.launch()
|
education_output.xlsx
ADDED
Binary file (11 kB). View file
|
|
jobs_output.xlsx
ADDED
Binary file (14.3 kB). View file
|
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
gradio
|
2 |
+
transformers
|
3 |
+
pandas
|
4 |
+
openpyxl
|