Upload final_app.py
Browse files- final_app.py +271 -0
final_app.py
ADDED
|
@@ -0,0 +1,271 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import pandas as pd
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
class AnonymizationEvaluator:
|
| 6 |
+
def __init__(self):
|
| 7 |
+
# فهرست بانکهای ایران
|
| 8 |
+
self.iranian_banks = [
|
| 9 |
+
'بانک ملی', 'بانک صادرات', 'بانک پاسارگاد', 'بانک کشاورزی',
|
| 10 |
+
'بانک ملت', 'بانک تجارت', 'بانک صنعت و معدن', 'بانک رسالت',
|
| 11 |
+
'بانک دی', 'بانک پارسیان', 'بانک کارآفرین', 'بانک سامان',
|
| 12 |
+
'بانک اقتصاد نوین', 'بانک مهر اقتصاد', 'بانک آینده', 'بانک کشاورزی'
|
| 13 |
+
]
|
| 14 |
+
|
| 15 |
+
# فهرست سازمانهای دولتی
|
| 16 |
+
self.government_orgs = [
|
| 17 |
+
'بانک مرکزی جمهوری اسلامی ایران',
|
| 18 |
+
'دفتر اسناد رسمی',
|
| 19 |
+
'اداره کل مالیات',
|
| 20 |
+
'تامین اجتماعی',
|
| 21 |
+
'وزارت دادگستری'
|
| 22 |
+
]
|
| 23 |
+
|
| 24 |
+
self.patterns = {
|
| 25 |
+
'person_names': {
|
| 26 |
+
'pattern': re.compile(r'(?:آقای|خانم)\s+[\u0600-\u06FF\s]+?(?=\s+با|$)', re.UNICODE),
|
| 27 |
+
'replacement': re.compile(r'person_\d+'),
|
| 28 |
+
'name': 'اسامی اشخاص'
|
| 29 |
+
},
|
| 30 |
+
'national_ids': {
|
| 31 |
+
'pattern': re.compile(r'\b\d{10,11}\b'),
|
| 32 |
+
'replacement': re.compile(r'id_number_\d+'),
|
| 33 |
+
'name': 'کدهای ملی'
|
| 34 |
+
},
|
| 35 |
+
'phone_numbers': {
|
| 36 |
+
'pattern': re.compile(r'(?:09\d{9}|021-\d{8}|0\d{2,3}-\d{8})'),
|
| 37 |
+
'replacement': re.compile(r'phone_\d+'),
|
| 38 |
+
'name': 'شماره تلفنها'
|
| 39 |
+
},
|
| 40 |
+
'account_numbers': {
|
| 41 |
+
'pattern': re.compile(r'\b\d{3}-\d{3}-\d{3}-\d{1}\b'),
|
| 42 |
+
'replacement': re.compile(r'account_\d+'),
|
| 43 |
+
'name': 'شماره حسابها'
|
| 44 |
+
},
|
| 45 |
+
'card_numbers': {
|
| 46 |
+
'pattern': re.compile(r'\b\d{4}-\d{4}-\d{4}-\d{4}\b'),
|
| 47 |
+
'replacement': re.compile(r'card_number_\d+'),
|
| 48 |
+
'name': 'شماره کارتها'
|
| 49 |
+
},
|
| 50 |
+
'amounts': {
|
| 51 |
+
'pattern': re.compile(r'\b\d+\s*تومان'),
|
| 52 |
+
'replacement': re.compile(r'amount_\d+'),
|
| 53 |
+
'name': 'مبالغ مالی'
|
| 54 |
+
},
|
| 55 |
+
'dates': {
|
| 56 |
+
'pattern': re.compile(r'(?:\d{4}\/\d{2}\/\d{2}|۳۰\s*اسفند\s*۱۴۰۳|\b\d{4}\b(?=\s*سال))'),
|
| 57 |
+
'replacement': re.compile(r'date_\d+'),
|
| 58 |
+
'name': 'تاریخها'
|
| 59 |
+
},
|
| 60 |
+
'iranian_banks': {
|
| 61 |
+
'pattern': re.compile(f"({'|'.join(self.iranian_banks)})", re.UNICODE),
|
| 62 |
+
'replacement': re.compile(r'company_\d+'),
|
| 63 |
+
'name': 'بانکهای ایران'
|
| 64 |
+
},
|
| 65 |
+
'government_orgs': {
|
| 66 |
+
'pattern': re.compile(f"({'|'.join(self.government_orgs)})", re.UNICODE),
|
| 67 |
+
'replacement': re.compile(r'company_\d+'),
|
| 68 |
+
'name': 'سازمانهای دولتی'
|
| 69 |
+
},
|
| 70 |
+
'other_companies': {
|
| 71 |
+
'pattern': re.compile(r'شرکت\s+[\u0600-\u06FF\s]+?(?=\s|$|،|\.)', re.UNICODE),
|
| 72 |
+
'replacement': re.compile(r'company_\d+'),
|
| 73 |
+
'name': 'سایر شرکتها'
|
| 74 |
+
},
|
| 75 |
+
'addresses': {
|
| 76 |
+
'pattern': re.compile(r'(?:تهران|کرج|اصفهان|بندر\s*ماهشهر)[،\s]+[^،\.]+', re.UNICODE),
|
| 77 |
+
'replacement': re.compile(r'(?:full_address_\d+|location_\d+)'),
|
| 78 |
+
'name': 'آدرسها'
|
| 79 |
+
},
|
| 80 |
+
'documents': {
|
| 81 |
+
'pattern': re.compile(r'(?:INV-\d{4}-\d{4}|RPT-\d{4}-\d{4}|\b\d{4}(?=\s+تهران)|\b\d{7}\b)'),
|
| 82 |
+
'replacement': re.compile(r'(?:invoice_\d+|report_\d+|contract_\d+|cheque_\d+)'),
|
| 83 |
+
'name': 'شماره اسناد'
|
| 84 |
+
}
|
| 85 |
+
}
|
| 86 |
+
|
| 87 |
+
def analyze_entities(self, original_text, anonymized_text):
|
| 88 |
+
results = {}
|
| 89 |
+
for entity_type, config in self.patterns.items():
|
| 90 |
+
original_matches = config['pattern'].findall(original_text)
|
| 91 |
+
replacement_matches = config['replacement'].findall(anonymized_text)
|
| 92 |
+
|
| 93 |
+
anonymized_count = 0
|
| 94 |
+
for match in original_matches:
|
| 95 |
+
if not anonymized_text.count(match.strip()):
|
| 96 |
+
anonymized_count += 1
|
| 97 |
+
|
| 98 |
+
if len(replacement_matches) > anonymized_count:
|
| 99 |
+
anonymized_count = min(len(replacement_matches), len(original_matches))
|
| 100 |
+
|
| 101 |
+
percentage = (anonymized_count / len(original_matches) * 100) if original_matches else 0
|
| 102 |
+
|
| 103 |
+
results[entity_type] = {
|
| 104 |
+
'name': config['name'],
|
| 105 |
+
'total': len(original_matches),
|
| 106 |
+
'anonymized': anonymized_count,
|
| 107 |
+
'percentage': round(percentage, 1),
|
| 108 |
+
'samples': original_matches[:3] if original_matches else []
|
| 109 |
+
}
|
| 110 |
+
return results
|
| 111 |
+
|
| 112 |
+
def evaluate_csv(self, csv_file):
|
| 113 |
+
try:
|
| 114 |
+
if csv_file is None:
|
| 115 |
+
return "لطفاً یک فایل CSV آپلود کنید."
|
| 116 |
+
|
| 117 |
+
try:
|
| 118 |
+
df = pd.read_csv(csv_file.name, encoding='utf-8')
|
| 119 |
+
except:
|
| 120 |
+
df = pd.read_csv(csv_file.name, encoding='utf-8', sep='\t')
|
| 121 |
+
|
| 122 |
+
if 'original_text' not in df.columns or 'anonymized_text' not in df.columns:
|
| 123 |
+
return "فایل CSV باید شامل ستونهای 'original_text' و 'anonymized_text' باشد."
|
| 124 |
+
|
| 125 |
+
overall_stats = {}
|
| 126 |
+
total_entities = 0
|
| 127 |
+
total_anonymized = 0
|
| 128 |
+
|
| 129 |
+
for _, row in df.iterrows():
|
| 130 |
+
if pd.isna(row['original_text']) or pd.isna(row['anonymized_text']):
|
| 131 |
+
continue
|
| 132 |
+
|
| 133 |
+
row_analysis = self.analyze_entities(str(row['original_text']), str(row['anonymized_text']))
|
| 134 |
+
|
| 135 |
+
for entity_type, data in row_analysis.items():
|
| 136 |
+
if entity_type not in overall_stats:
|
| 137 |
+
overall_stats[entity_type] = {
|
| 138 |
+
'name': data['name'],
|
| 139 |
+
'total': 0,
|
| 140 |
+
'anonymized': 0,
|
| 141 |
+
'samples': []
|
| 142 |
+
}
|
| 143 |
+
|
| 144 |
+
overall_stats[entity_type]['total'] += data['total']
|
| 145 |
+
overall_stats[entity_type]['anonymized'] += data['anonymized']
|
| 146 |
+
overall_stats[entity_type]['samples'].extend(data['samples'])
|
| 147 |
+
|
| 148 |
+
total_entities += data['total']
|
| 149 |
+
total_anonymized += data['anonymized']
|
| 150 |
+
|
| 151 |
+
for entity_type in overall_stats:
|
| 152 |
+
stats = overall_stats[entity_type]
|
| 153 |
+
stats['percentage'] = round((stats['anonymized'] / stats['total'] * 100) if stats['total'] > 0 else 0, 1)
|
| 154 |
+
stats['samples'] = list(set(stats['samples']))[:3]
|
| 155 |
+
|
| 156 |
+
return self.generate_report(overall_stats, total_entities, total_anonymized, len(df))
|
| 157 |
+
|
| 158 |
+
except Exception as e:
|
| 159 |
+
return f"خطا در پردازش فایل: {str(e)}"
|
| 160 |
+
|
| 161 |
+
def generate_report(self, stats, total_entities, total_anonymized, total_rows):
|
| 162 |
+
report = f"""# گزارش ارزیابی ناشناسسازی متن
|
| 163 |
+
|
| 164 |
+
## خلاصه کلی
|
| 165 |
+
- **تعداد ردیفهای پردازش شده**: {total_rows:,} ردیف
|
| 166 |
+
- **تعداد موجودیتهای حساس شناسایی شده**: {total_entities:,} مورد
|
| 167 |
+
- **تعداد موجودیتهای ناشناس شده**: {total_anonymized:,} مورد
|
| 168 |
+
- **درصد پوشش کلی**: {(total_anonymized/total_entities*100) if total_entities > 0 else 0:.1f}%
|
| 169 |
+
|
| 170 |
+
## تحلیل تفصیلی دستهبندی موجودیتها
|
| 171 |
+
|
| 172 |
+
"""
|
| 173 |
+
|
| 174 |
+
excellent = []
|
| 175 |
+
good = []
|
| 176 |
+
poor = []
|
| 177 |
+
not_found = []
|
| 178 |
+
|
| 179 |
+
for entity_type, data in stats.items():
|
| 180 |
+
if data['total'] == 0:
|
| 181 |
+
not_found.append((entity_type, data))
|
| 182 |
+
elif data['percentage'] == 100:
|
| 183 |
+
excellent.append((entity_type, data))
|
| 184 |
+
elif data['percentage'] >= 80:
|
| 185 |
+
good.append((entity_type, data))
|
| 186 |
+
else:
|
| 187 |
+
poor.append((entity_type, data))
|
| 188 |
+
|
| 189 |
+
if excellent:
|
| 190 |
+
report += "### ✅ عملکرد عالی (100% موفقیت)\n"
|
| 191 |
+
for entity_type, data in excellent:
|
| 192 |
+
report += f"- **{data['name']}**: {data['anonymized']}/{data['total']} (100%)\n"
|
| 193 |
+
report += "\n"
|
| 194 |
+
|
| 195 |
+
if good:
|
| 196 |
+
report += "### 🟡 عملکرد خوب (80-99% موفقیت)\n"
|
| 197 |
+
for entity_type, data in good:
|
| 198 |
+
report += f"- **{data['name']}**: {data['anonymized']}/{data['total']} ({data['percentage']}%)\n"
|
| 199 |
+
report += "\n"
|
| 200 |
+
|
| 201 |
+
if poor:
|
| 202 |
+
report += "### 🔴 عملکرد ضعیف (<80% موفقیت)\n"
|
| 203 |
+
for entity_type, data in poor:
|
| 204 |
+
missed = data['total'] - data['anonymized']
|
| 205 |
+
report += f"- **{data['name']}**: {data['anonymized']}/{data['total']} ({data['percentage']}%) - {missed} مورد جا مانده\n"
|
| 206 |
+
report += "\n"
|
| 207 |
+
|
| 208 |
+
if not_found:
|
| 209 |
+
report += "### ⚪ موجودیتهای یافت نشده\n"
|
| 210 |
+
for entity_type, data in not_found:
|
| 211 |
+
report += f"- **{data['name']}**: هیچ موجودیتی یافت نشد\n"
|
| 212 |
+
report += "\n"
|
| 213 |
+
|
| 214 |
+
report += "## جدول خلاصه متریکها\n\n"
|
| 215 |
+
report += "| دسته موجودیت | یافته شده | ناشناس شده | درصد موفقیت | موارد جا مانده |\n"
|
| 216 |
+
report += "|---------------|-----------|-------------|-------------|----------------|\n"
|
| 217 |
+
|
| 218 |
+
for entity_type, data in stats.items():
|
| 219 |
+
if data['total'] > 0:
|
| 220 |
+
missed = data['total'] - data['anonymized']
|
| 221 |
+
report += f"| {data['name']} | {data['total']} | {data['anonymized']} | {data['percentage']}% | {missed} |\n"
|
| 222 |
+
|
| 223 |
+
major_issues = [(k, v) for k, v in stats.items() if v['total'] > 0 and v['percentage'] < 80]
|
| 224 |
+
major_issues.sort(key=lambda x: x[1]['total'] - x[1]['anonymized'], reverse=True)
|
| 225 |
+
|
| 226 |
+
if major_issues:
|
| 227 |
+
report += "\n## 🚨 مشکلات اصلی شناسایی شده\n\n"
|
| 228 |
+
for i, (entity_type, data) in enumerate(major_issues, 1):
|
| 229 |
+
missed = data['total'] - data['anonymized']
|
| 230 |
+
impact = round(missed / total_entities * 100, 1) if total_entities > 0 else 0
|
| 231 |
+
report += f"### {i}. {data['name']}\n"
|
| 232 |
+
report += f"- **وضعیت**: {data['percentage']}% موفقیت\n"
|
| 233 |
+
report += f"- **موارد جا مانده**: {missed} مورد از {data['total']} مورد\n"
|
| 234 |
+
report += f"- **تاثیر بر کل**: {impact}% از کل موجودیتها\n\n"
|
| 235 |
+
|
| 236 |
+
precision = round((total_anonymized / total_entities * 100) if total_entities > 0 else 0, 1)
|
| 237 |
+
|
| 238 |
+
report += f"""## 📊 آمار نهایی
|
| 239 |
+
|
| 240 |
+
- **کل موجودیتهای شناسایی شده**: {total_entities:,}
|
| 241 |
+
- **کل موجودیتهای ناشناس شده**: {total_anonymized:,}
|
| 242 |
+
- **موجودیتهای جا مانده**: {total_entities - total_anonymized:,}
|
| 243 |
+
- **دقت (Precision)**: {precision}%
|
| 244 |
+
- **پوشش (Recall)**: {precision}%
|
| 245 |
+
- **امتیاز F1**: {precision}%
|
| 246 |
+
"""
|
| 247 |
+
|
| 248 |
+
return report
|
| 249 |
+
|
| 250 |
+
def create_interface():
|
| 251 |
+
evaluator = AnonymizationEvaluator()
|
| 252 |
+
|
| 253 |
+
def process_file(csv_file):
|
| 254 |
+
if csv_file is None:
|
| 255 |
+
return "لطفاً یک فایل CSV آپلود کنید."
|
| 256 |
+
return evaluator.evaluate_csv(csv_file)
|
| 257 |
+
|
| 258 |
+
with gr.Blocks(title="ارزیاب متریکهای ناشناسسازی") as demo:
|
| 259 |
+
gr.Markdown("# ارزیاب متریکهای ناشناسسازی متن فارسی")
|
| 260 |
+
|
| 261 |
+
file_input = gr.File(label="آپلود فایل CSV", file_types=[".csv"])
|
| 262 |
+
analyze_btn = gr.Button("محاسبه متریکها", variant="primary")
|
| 263 |
+
output = gr.Markdown(value="فایل CSV خود را آپلود کنید.")
|
| 264 |
+
|
| 265 |
+
analyze_btn.click(fn=process_file, inputs=[file_input], outputs=[output])
|
| 266 |
+
|
| 267 |
+
return demo
|
| 268 |
+
|
| 269 |
+
if __name__ == "__main__":
|
| 270 |
+
app = create_interface()
|
| 271 |
+
app.launch()
|