Spaces:
Sleeping
Sleeping
Update utils/spacy.py
Browse files- utils/spacy.py +461 -246
utils/spacy.py
CHANGED
|
@@ -1,246 +1,461 @@
|
|
| 1 |
-
import spacy
|
| 2 |
-
import logging
|
| 3 |
-
import json
|
| 4 |
-
from utils.fileTotext import extract_text_based_on_format
|
| 5 |
-
import re
|
| 6 |
-
|
| 7 |
-
def is_valid_email(email):
|
| 8 |
-
email_regex = r'^[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+$'
|
| 9 |
-
return re.match(email_regex, email) is not None
|
| 10 |
-
|
| 11 |
-
def is_valid_contact(contact):
|
| 12 |
-
patterns = [
|
| 13 |
-
r'^\+91[\s\.\-\/]?\(?0?\)?[\s\-\.\/]?\d{5}[\s\-\.\/]?\d{5}$', # +91 with optional 0 and separators
|
| 14 |
-
r'^\+91[\s\.\-\/]?\d{5}[\s\-\.\/]?\d{5}$', # +91 with 10 digits separated
|
| 15 |
-
r'^\d{5}[\s\-\.\/]?\d{5}$', # Local format without country code
|
| 16 |
-
r'^\+91[\s\.\-\/]?\d{10}$', # +91 with 10 digits together
|
| 17 |
-
r'^\d{10}$', # 10 digits together
|
| 18 |
-
r'^\+91[\s\.\-\/]?\(?\d{5}\)?[\s\-\.\/]?\d{5}[\s\-\.\/]?\d{5}$' # +91 with varying separators
|
| 19 |
-
r'\+1\s\(\d{3}\)\s\d{3}-\d{4} ', # USA/Canada Intl +1 (XXX) XXX-XXXX
|
| 20 |
-
r'\(\d{3}\)\s\d{3}-\d{4} ', # USA/Canada STD (XXX) XXX-XXXX
|
| 21 |
-
r'\(\d{3}\)\s\d{3}\s\d{4} ', # USA/Canada (XXX) XXX XXXX
|
| 22 |
-
r'\(\d{3}\)\s\d{3}\s\d{3} ', # USA/Canada (XXX) XXX XXX
|
| 23 |
-
r'\+1\d{10} ', # +1 XXXXXXXXXX
|
| 24 |
-
r'\d{10} ', # XXXXXXXXXX
|
| 25 |
-
r'\+44\s\d{4}\s\d{6} ', # UK Intl +44 XXXX XXXXXX
|
| 26 |
-
r'\+44\s\d{3}\s\d{3}\s\d{4} ', # UK Intl +44 XXX XXX XXXX
|
| 27 |
-
r'0\d{4}\s\d{6} ', # UK STD 0XXXX XXXXXX
|
| 28 |
-
r'0\d{3}\s\d{3}\s\d{4} ', # UK STD 0XXX XXX XXXX
|
| 29 |
-
r'\+44\d{10} ', # +44 XXXXXXXXXX
|
| 30 |
-
r'0\d{10} ', # 0XXXXXXXXXX
|
| 31 |
-
r'\+61\s\d\s\d{4}\s\d{4} ', # Australia Intl +61 X XXXX XXXX
|
| 32 |
-
r'0\d\s\d{4}\s\d{4} ', # Australia STD 0X XXXX XXXX
|
| 33 |
-
r'\+61\d{9} ', # +61 XXXXXXXXX
|
| 34 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 35 |
-
r'\+91\s\d{5}-\d{5} ', # India Intl +91 XXXXX-XXXXX
|
| 36 |
-
r'\+91\s\d{4}-\d{6} ', # India Intl +91 XXXX-XXXXXX
|
| 37 |
-
r'\+91\s\d{10} ', # India Intl +91 XXXXXXXXXX
|
| 38 |
-
r'0\d{2}-\d{7} ', # India STD 0XX-XXXXXXX
|
| 39 |
-
r'\+91\d{10} ', # +91 XXXXXXXXXX
|
| 40 |
-
r'\+49\s\d{4}\s\d{8} ', # Germany Intl +49 XXXX XXXXXXXX
|
| 41 |
-
r'\+49\s\d{3}\s\d{7} ', # Germany Intl +49 XXX XXXXXXX
|
| 42 |
-
r'0\d{3}\s\d{8} ', # Germany STD 0XXX XXXXXXXX
|
| 43 |
-
r'\+49\d{12} ', # +49 XXXXXXXXXXXX
|
| 44 |
-
r'\+49\d{10} ', # +49 XXXXXXXXXX
|
| 45 |
-
r'0\d{11} ', # 0XXXXXXXXXXX
|
| 46 |
-
r'\+86\s\d{3}\s\d{4}\s\d{4} ', # China Intl +86 XXX XXXX XXXX
|
| 47 |
-
r'0\d{3}\s\d{4}\s\d{4} ', # China STD 0XXX XXXX XXXX
|
| 48 |
-
r'\+86\d{11} ', # +86 XXXXXXXXXXX
|
| 49 |
-
r'\+81\s\d\s\d{4}\s\d{4} ', # Japan Intl +81 X XXXX XXXX
|
| 50 |
-
r'\+81\s\d{2}\s\d{4}\s\d{4} ', # Japan Intl +81 XX XXXX XXXX
|
| 51 |
-
r'0\d\s\d{4}\s\d{4} ', # Japan STD 0X XXXX XXXX
|
| 52 |
-
r'\+81\d{10} ', # +81 XXXXXXXXXX
|
| 53 |
-
r'\+81\d{9} ', # +81 XXXXXXXXX
|
| 54 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 55 |
-
r'\+55\s\d{2}\s\d{5}-\d{4} ', # Brazil Intl +55 XX XXXXX-XXXX
|
| 56 |
-
r'\+55\s\d{2}\s\d{4}-\d{4} ', # Brazil Intl +55 XX XXXX-XXXX
|
| 57 |
-
r'0\d{2}\s\d{4}\s\d{4} ', # Brazil STD 0XX XXXX XXXX
|
| 58 |
-
r'\+55\d{11} ', # +55 XXXXXXXXXXX
|
| 59 |
-
r'\+55\d{10} ', # +55 XXXXXXXXXX
|
| 60 |
-
r'0\d{10} ', # 0XXXXXXXXXX
|
| 61 |
-
r'\+33\s\d\s\d{2}\s\d{2}\s\d{2}\s\d{2} ', # France Intl +33 X XX XX XX XX
|
| 62 |
-
r'0\d\s\d{2}\s\d{2}\s\d{2}\s\d{2} ', # France STD 0X XX XX XX XX
|
| 63 |
-
r'\+33\d{9} ', # +33 XXXXXXXXX
|
| 64 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 65 |
-
r'\+7\s\d{3}\s\d{3}-\d{2}-\d{2} ', # Russia Intl +7 XXX XXX-XX-XX
|
| 66 |
-
r'8\s\d{3}\s\d{3}-\d{2}-\d{2} ', # Russia STD 8 XXX XXX-XX-XX
|
| 67 |
-
r'\+7\d{10} ', # +7 XXXXXXXXXX
|
| 68 |
-
r'8\d{10} ', # 8 XXXXXXXXXX
|
| 69 |
-
r'\+27\s\d{2}\s\d{3}\s\d{4} ', # South Africa Intl +27 XX XXX XXXX
|
| 70 |
-
r'0\d{2}\s\d{3}\s\d{4} ', # South Africa STD 0XX XXX XXXX
|
| 71 |
-
r'\+27\d{9} ', # +27 XXXXXXXXX
|
| 72 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 73 |
-
r'\+52\s\d{3}\s\d{3}\s\d{4} ', # Mexico Intl +52 XXX XXX XXXX
|
| 74 |
-
r'\+52\s\d{2}\s\d{4}\s\d{4} ', # Mexico Intl +52 XX XXXX XXXX
|
| 75 |
-
r'01\s\d{3}\s\d{4} ', # Mexico STD 01 XXX XXXX
|
| 76 |
-
r'\+52\d{10} ', # +52 XXXXXXXXXX
|
| 77 |
-
r'01\d{7} ', # 01 XXXXXXX
|
| 78 |
-
r'\+234\s\d{3}\s\d{3}\s\d{4} ', # Nigeria Intl +234 XXX XXX XXXX
|
| 79 |
-
r'0\d{3}\s\d{3}\s\d{4} ', # Nigeria STD 0XXX XXX XXXX
|
| 80 |
-
r'\+234\d{10} ', # +234 XXXXXXXXXX
|
| 81 |
-
r'0\d{10} ', # 0XXXXXXXXXX
|
| 82 |
-
r'\+971\s\d\s\d{3}\s\d{4} ', # UAE Intl +971 X XXX XXXX
|
| 83 |
-
r'0\d\s\d{3}\s\d{4} ', # UAE STD 0X XXX XXXX
|
| 84 |
-
r'\+971\d{8} ', # +971 XXXXXXXX
|
| 85 |
-
r'0\d{8} ', # 0XXXXXXXX
|
| 86 |
-
r'\+54\s9\s\d{3}\s\d{3}\s\d{4} ', # Argentina Intl +54 9 XXX XXX XXXX
|
| 87 |
-
r'\+54\s\d{1}\s\d{4}\s\d{4} ', # Argentina Intl +54 X XXXX XXXX
|
| 88 |
-
r'0\d{3}\s\d{4} ', # Argentina STD 0XXX XXXX
|
| 89 |
-
r'\+54\d{10} ', # +54 9 XXXXXXXXXX
|
| 90 |
-
r'\+54\d{9} ', # +54 XXXXXXXXX
|
| 91 |
-
r'0\d{7} ', # 0XXXXXXX
|
| 92 |
-
r'\+966\s\d\s\d{3}\s\d{4} ', # Saudi Intl +966 X XXX XXXX
|
| 93 |
-
r'0\d\s\d{3}\s\d{4} ', # Saudi STD 0X XXX XXXX
|
| 94 |
-
r'\+966\d{8} ', # +966 XXXXXXXX
|
| 95 |
-
r'0\d{8} ', # 0XXXXXXXX
|
| 96 |
-
r'\+1\d{10} ', # +1 XXXXXXXXXX
|
| 97 |
-
r'\+1\s\d{3}\s\d{3}\s\d{4} ', # +1 XXX XXX XXXX
|
| 98 |
-
r'\d{5}\s\d{5} ', # XXXXX XXXXX
|
| 99 |
-
r'\d{10} ', # XXXXXXXXXX
|
| 100 |
-
r'\+44\d{10} ', # +44 XXXXXXXXXX
|
| 101 |
-
r'0\d{10} ', # 0XXXXXXXXXX
|
| 102 |
-
r'\+61\d{9} ', # +61 XXXXXXXXX
|
| 103 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 104 |
-
r'\+91\d{10} ', # +91 XXXXXXXXXX
|
| 105 |
-
r'\+49\d{12} ', # +49 XXXXXXXXXXXX
|
| 106 |
-
r'\+49\d{10} ', # +49 XXXXXXXXXX
|
| 107 |
-
r'0\d{11} ', # 0XXXXXXXXXXX
|
| 108 |
-
r'\+86\d{11} ', # +86 XXXXXXXXXXX
|
| 109 |
-
r'\+81\d{10} ', # +81 XXXXXXXXXX
|
| 110 |
-
r'\+81\d{9} ', # +81 XXXXXXXXX
|
| 111 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 112 |
-
r'\+55\d{11} ', # +55 XXXXXXXXXXX
|
| 113 |
-
r'\+55\d{10} ', # +55 XXXXXXXXXX
|
| 114 |
-
r'0\d{10} ', # 0XXXXXXXXXX
|
| 115 |
-
r'\+33\d{9} ', # +33 XXXXXXXXX
|
| 116 |
-
r'0\d{9} ', # 0XXXXXXXXX
|
| 117 |
-
r'\+7\d{10} ', # +7 XXXXXXXXXX
|
| 118 |
-
r'8\d{10} ', # 8 XXXXXXXXXX
|
| 119 |
-
r'\+27\d{9} ', # +27 XXXXXXXXX
|
| 120 |
-
r'0\d{9} ', # 0XXXXXXXXX (South Africa STD)
|
| 121 |
-
r'\+52\d{10} ', # +52 XXXXXXXXXX
|
| 122 |
-
r'01\d{7} ', # 01 XXXXXXX
|
| 123 |
-
r'\+234\d{10} ', # +234 XXXXXXXXXX
|
| 124 |
-
r'0\d{10} ', # 0XXXXXXXXXX
|
| 125 |
-
r'\+971\d{8} ', # +971 XXXXXXXX
|
| 126 |
-
r'0\d{8} ', # 0XXXXXXXX
|
| 127 |
-
r'\+54\s9\s\d{10} ', # +54 9 XXXXXXXXXX
|
| 128 |
-
r'\+54\d{9} ', # +54 XXXXXXXXX
|
| 129 |
-
r'0\d{7} ', # 0XXXXXXX
|
| 130 |
-
r'\+966\d{8} ', # +966 XXXXXXXX
|
| 131 |
-
r'0\d{8}' # 0XXXXXXXX
|
| 132 |
-
]
|
| 133 |
-
|
| 134 |
-
# Check if the contact matches any of the patterns
|
| 135 |
-
return any(re.match(pattern, contact) for pattern in patterns) is not None
|
| 136 |
-
|
| 137 |
-
#
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
}
|
| 150 |
-
|
| 151 |
-
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
}
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
| 189 |
-
|
| 190 |
-
|
| 191 |
-
|
| 192 |
-
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
|
| 199 |
-
|
| 200 |
-
|
| 201 |
-
|
| 202 |
-
|
| 203 |
-
|
| 204 |
-
|
| 205 |
-
|
| 206 |
-
|
| 207 |
-
|
| 208 |
-
|
| 209 |
-
|
| 210 |
-
|
| 211 |
-
|
| 212 |
-
|
| 213 |
-
|
| 214 |
-
|
| 215 |
-
|
| 216 |
-
|
| 217 |
-
|
| 218 |
-
|
| 219 |
-
|
| 220 |
-
|
| 221 |
-
|
| 222 |
-
|
| 223 |
-
|
| 224 |
-
|
| 225 |
-
|
| 226 |
-
|
| 227 |
-
|
| 228 |
-
|
| 229 |
-
|
| 230 |
-
|
| 231 |
-
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
|
| 235 |
-
|
| 236 |
-
|
| 237 |
-
|
| 238 |
-
|
| 239 |
-
|
| 240 |
-
|
| 241 |
-
|
| 242 |
-
|
| 243 |
-
|
| 244 |
-
|
| 245 |
-
|
| 246 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import spacy
|
| 2 |
+
import logging
|
| 3 |
+
import json
|
| 4 |
+
from utils.fileTotext import extract_text_based_on_format
|
| 5 |
+
import re
|
| 6 |
+
|
| 7 |
+
def is_valid_email(email):
|
| 8 |
+
email_regex = r'^[a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+$'
|
| 9 |
+
return re.match(email_regex, email) is not None
|
| 10 |
+
|
| 11 |
+
def is_valid_contact(contact):
|
| 12 |
+
patterns = [
|
| 13 |
+
r'^\+91[\s\.\-\/]?\(?0?\)?[\s\-\.\/]?\d{5}[\s\-\.\/]?\d{5}$', # +91 with optional 0 and separators
|
| 14 |
+
r'^\+91[\s\.\-\/]?\d{5}[\s\-\.\/]?\d{5}$', # +91 with 10 digits separated
|
| 15 |
+
r'^\d{5}[\s\-\.\/]?\d{5}$', # Local format without country code
|
| 16 |
+
r'^\+91[\s\.\-\/]?\d{10}$', # +91 with 10 digits together
|
| 17 |
+
r'^\d{10}$', # 10 digits together
|
| 18 |
+
r'^\+91[\s\.\-\/]?\(?\d{5}\)?[\s\-\.\/]?\d{5}[\s\-\.\/]?\d{5}$' # +91 with varying separators
|
| 19 |
+
r'\+1\s\(\d{3}\)\s\d{3}-\d{4} ', # USA/Canada Intl +1 (XXX) XXX-XXXX
|
| 20 |
+
r'\(\d{3}\)\s\d{3}-\d{4} ', # USA/Canada STD (XXX) XXX-XXXX
|
| 21 |
+
r'\(\d{3}\)\s\d{3}\s\d{4} ', # USA/Canada (XXX) XXX XXXX
|
| 22 |
+
r'\(\d{3}\)\s\d{3}\s\d{3} ', # USA/Canada (XXX) XXX XXX
|
| 23 |
+
r'\+1\d{10} ', # +1 XXXXXXXXXX
|
| 24 |
+
r'\d{10} ', # XXXXXXXXXX
|
| 25 |
+
r'\+44\s\d{4}\s\d{6} ', # UK Intl +44 XXXX XXXXXX
|
| 26 |
+
r'\+44\s\d{3}\s\d{3}\s\d{4} ', # UK Intl +44 XXX XXX XXXX
|
| 27 |
+
r'0\d{4}\s\d{6} ', # UK STD 0XXXX XXXXXX
|
| 28 |
+
r'0\d{3}\s\d{3}\s\d{4} ', # UK STD 0XXX XXX XXXX
|
| 29 |
+
r'\+44\d{10} ', # +44 XXXXXXXXXX
|
| 30 |
+
r'0\d{10} ', # 0XXXXXXXXXX
|
| 31 |
+
r'\+61\s\d\s\d{4}\s\d{4} ', # Australia Intl +61 X XXXX XXXX
|
| 32 |
+
r'0\d\s\d{4}\s\d{4} ', # Australia STD 0X XXXX XXXX
|
| 33 |
+
r'\+61\d{9} ', # +61 XXXXXXXXX
|
| 34 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 35 |
+
r'\+91\s\d{5}-\d{5} ', # India Intl +91 XXXXX-XXXXX
|
| 36 |
+
r'\+91\s\d{4}-\d{6} ', # India Intl +91 XXXX-XXXXXX
|
| 37 |
+
r'\+91\s\d{10} ', # India Intl +91 XXXXXXXXXX
|
| 38 |
+
r'0\d{2}-\d{7} ', # India STD 0XX-XXXXXXX
|
| 39 |
+
r'\+91\d{10} ', # +91 XXXXXXXXXX
|
| 40 |
+
r'\+49\s\d{4}\s\d{8} ', # Germany Intl +49 XXXX XXXXXXXX
|
| 41 |
+
r'\+49\s\d{3}\s\d{7} ', # Germany Intl +49 XXX XXXXXXX
|
| 42 |
+
r'0\d{3}\s\d{8} ', # Germany STD 0XXX XXXXXXXX
|
| 43 |
+
r'\+49\d{12} ', # +49 XXXXXXXXXXXX
|
| 44 |
+
r'\+49\d{10} ', # +49 XXXXXXXXXX
|
| 45 |
+
r'0\d{11} ', # 0XXXXXXXXXXX
|
| 46 |
+
r'\+86\s\d{3}\s\d{4}\s\d{4} ', # China Intl +86 XXX XXXX XXXX
|
| 47 |
+
r'0\d{3}\s\d{4}\s\d{4} ', # China STD 0XXX XXXX XXXX
|
| 48 |
+
r'\+86\d{11} ', # +86 XXXXXXXXXXX
|
| 49 |
+
r'\+81\s\d\s\d{4}\s\d{4} ', # Japan Intl +81 X XXXX XXXX
|
| 50 |
+
r'\+81\s\d{2}\s\d{4}\s\d{4} ', # Japan Intl +81 XX XXXX XXXX
|
| 51 |
+
r'0\d\s\d{4}\s\d{4} ', # Japan STD 0X XXXX XXXX
|
| 52 |
+
r'\+81\d{10} ', # +81 XXXXXXXXXX
|
| 53 |
+
r'\+81\d{9} ', # +81 XXXXXXXXX
|
| 54 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 55 |
+
r'\+55\s\d{2}\s\d{5}-\d{4} ', # Brazil Intl +55 XX XXXXX-XXXX
|
| 56 |
+
r'\+55\s\d{2}\s\d{4}-\d{4} ', # Brazil Intl +55 XX XXXX-XXXX
|
| 57 |
+
r'0\d{2}\s\d{4}\s\d{4} ', # Brazil STD 0XX XXXX XXXX
|
| 58 |
+
r'\+55\d{11} ', # +55 XXXXXXXXXXX
|
| 59 |
+
r'\+55\d{10} ', # +55 XXXXXXXXXX
|
| 60 |
+
r'0\d{10} ', # 0XXXXXXXXXX
|
| 61 |
+
r'\+33\s\d\s\d{2}\s\d{2}\s\d{2}\s\d{2} ', # France Intl +33 X XX XX XX XX
|
| 62 |
+
r'0\d\s\d{2}\s\d{2}\s\d{2}\s\d{2} ', # France STD 0X XX XX XX XX
|
| 63 |
+
r'\+33\d{9} ', # +33 XXXXXXXXX
|
| 64 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 65 |
+
r'\+7\s\d{3}\s\d{3}-\d{2}-\d{2} ', # Russia Intl +7 XXX XXX-XX-XX
|
| 66 |
+
r'8\s\d{3}\s\d{3}-\d{2}-\d{2} ', # Russia STD 8 XXX XXX-XX-XX
|
| 67 |
+
r'\+7\d{10} ', # +7 XXXXXXXXXX
|
| 68 |
+
r'8\d{10} ', # 8 XXXXXXXXXX
|
| 69 |
+
r'\+27\s\d{2}\s\d{3}\s\d{4} ', # South Africa Intl +27 XX XXX XXXX
|
| 70 |
+
r'0\d{2}\s\d{3}\s\d{4} ', # South Africa STD 0XX XXX XXXX
|
| 71 |
+
r'\+27\d{9} ', # +27 XXXXXXXXX
|
| 72 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 73 |
+
r'\+52\s\d{3}\s\d{3}\s\d{4} ', # Mexico Intl +52 XXX XXX XXXX
|
| 74 |
+
r'\+52\s\d{2}\s\d{4}\s\d{4} ', # Mexico Intl +52 XX XXXX XXXX
|
| 75 |
+
r'01\s\d{3}\s\d{4} ', # Mexico STD 01 XXX XXXX
|
| 76 |
+
r'\+52\d{10} ', # +52 XXXXXXXXXX
|
| 77 |
+
r'01\d{7} ', # 01 XXXXXXX
|
| 78 |
+
r'\+234\s\d{3}\s\d{3}\s\d{4} ', # Nigeria Intl +234 XXX XXX XXXX
|
| 79 |
+
r'0\d{3}\s\d{3}\s\d{4} ', # Nigeria STD 0XXX XXX XXXX
|
| 80 |
+
r'\+234\d{10} ', # +234 XXXXXXXXXX
|
| 81 |
+
r'0\d{10} ', # 0XXXXXXXXXX
|
| 82 |
+
r'\+971\s\d\s\d{3}\s\d{4} ', # UAE Intl +971 X XXX XXXX
|
| 83 |
+
r'0\d\s\d{3}\s\d{4} ', # UAE STD 0X XXX XXXX
|
| 84 |
+
r'\+971\d{8} ', # +971 XXXXXXXX
|
| 85 |
+
r'0\d{8} ', # 0XXXXXXXX
|
| 86 |
+
r'\+54\s9\s\d{3}\s\d{3}\s\d{4} ', # Argentina Intl +54 9 XXX XXX XXXX
|
| 87 |
+
r'\+54\s\d{1}\s\d{4}\s\d{4} ', # Argentina Intl +54 X XXXX XXXX
|
| 88 |
+
r'0\d{3}\s\d{4} ', # Argentina STD 0XXX XXXX
|
| 89 |
+
r'\+54\d{10} ', # +54 9 XXXXXXXXXX
|
| 90 |
+
r'\+54\d{9} ', # +54 XXXXXXXXX
|
| 91 |
+
r'0\d{7} ', # 0XXXXXXX
|
| 92 |
+
r'\+966\s\d\s\d{3}\s\d{4} ', # Saudi Intl +966 X XXX XXXX
|
| 93 |
+
r'0\d\s\d{3}\s\d{4} ', # Saudi STD 0X XXX XXXX
|
| 94 |
+
r'\+966\d{8} ', # +966 XXXXXXXX
|
| 95 |
+
r'0\d{8} ', # 0XXXXXXXX
|
| 96 |
+
r'\+1\d{10} ', # +1 XXXXXXXXXX
|
| 97 |
+
r'\+1\s\d{3}\s\d{3}\s\d{4} ', # +1 XXX XXX XXXX
|
| 98 |
+
r'\d{5}\s\d{5} ', # XXXXX XXXXX
|
| 99 |
+
r'\d{10} ', # XXXXXXXXXX
|
| 100 |
+
r'\+44\d{10} ', # +44 XXXXXXXXXX
|
| 101 |
+
r'0\d{10} ', # 0XXXXXXXXXX
|
| 102 |
+
r'\+61\d{9} ', # +61 XXXXXXXXX
|
| 103 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 104 |
+
r'\+91\d{10} ', # +91 XXXXXXXXXX
|
| 105 |
+
r'\+49\d{12} ', # +49 XXXXXXXXXXXX
|
| 106 |
+
r'\+49\d{10} ', # +49 XXXXXXXXXX
|
| 107 |
+
r'0\d{11} ', # 0XXXXXXXXXXX
|
| 108 |
+
r'\+86\d{11} ', # +86 XXXXXXXXXXX
|
| 109 |
+
r'\+81\d{10} ', # +81 XXXXXXXXXX
|
| 110 |
+
r'\+81\d{9} ', # +81 XXXXXXXXX
|
| 111 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 112 |
+
r'\+55\d{11} ', # +55 XXXXXXXXXXX
|
| 113 |
+
r'\+55\d{10} ', # +55 XXXXXXXXXX
|
| 114 |
+
r'0\d{10} ', # 0XXXXXXXXXX
|
| 115 |
+
r'\+33\d{9} ', # +33 XXXXXXXXX
|
| 116 |
+
r'0\d{9} ', # 0XXXXXXXXX
|
| 117 |
+
r'\+7\d{10} ', # +7 XXXXXXXXXX
|
| 118 |
+
r'8\d{10} ', # 8 XXXXXXXXXX
|
| 119 |
+
r'\+27\d{9} ', # +27 XXXXXXXXX
|
| 120 |
+
r'0\d{9} ', # 0XXXXXXXXX (South Africa STD)
|
| 121 |
+
r'\+52\d{10} ', # +52 XXXXXXXXXX
|
| 122 |
+
r'01\d{7} ', # 01 XXXXXXX
|
| 123 |
+
r'\+234\d{10} ', # +234 XXXXXXXXXX
|
| 124 |
+
r'0\d{10} ', # 0XXXXXXXXXX
|
| 125 |
+
r'\+971\d{8} ', # +971 XXXXXXXX
|
| 126 |
+
r'0\d{8} ', # 0XXXXXXXX
|
| 127 |
+
r'\+54\s9\s\d{10} ', # +54 9 XXXXXXXXXX
|
| 128 |
+
r'\+54\d{9} ', # +54 XXXXXXXXX
|
| 129 |
+
r'0\d{7} ', # 0XXXXXXX
|
| 130 |
+
r'\+966\d{8} ', # +966 XXXXXXXX
|
| 131 |
+
r'0\d{8}' # 0XXXXXXXX
|
| 132 |
+
]
|
| 133 |
+
|
| 134 |
+
# Check if the contact matches any of the patterns
|
| 135 |
+
return any(re.match(pattern, contact) for pattern in patterns) is not None
|
| 136 |
+
|
| 137 |
+
# Extracting the Contact Details through Regular Expression
|
| 138 |
+
def extract_contact_details(text):
|
| 139 |
+
# Regex patterns
|
| 140 |
+
# Phone numbers with at least 5 digits in any segment
|
| 141 |
+
combined_phone_regex = re.compile(r'''
|
| 142 |
+
(?:
|
| 143 |
+
#(?:(?:\+91[-.\s]?)?\d{5}[-.\s]?\d{5})|(?:\+?\d{1,3})?[-.\s()]?\d{5,}[-.\s()]?\d{5,}[-.\s()]?\d{1,9} | /^[\.-)( ]*([0-9]{3})[\.-)( ]*([0-9]{3})[\.-)( ]*([0-9]{4})$/ |
|
| 144 |
+
\+1\s\(\d{3}\)\s\d{3}-\d{4} | # USA/Canada Intl +1 (XXX) XXX-XXXX
|
| 145 |
+
\(\d{3}\)\s\d{3}-\d{4} | # USA/Canada STD (XXX) XXX-XXXX
|
| 146 |
+
\(\d{3}\)\s\d{3}\s\d{4} | # USA/Canada (XXX) XXX XXXX
|
| 147 |
+
\(\d{3}\)\s\d{3}\s\d{3} | # USA/Canada (XXX) XXX XXX
|
| 148 |
+
\+1\d{10} | # +1 XXXXXXXXXX
|
| 149 |
+
\d{10} | # XXXXXXXXXX
|
| 150 |
+
\+44\s\d{4}\s\d{6} | # UK Intl +44 XXXX XXXXXX
|
| 151 |
+
\+44\s\d{3}\s\d{3}\s\d{4} | # UK Intl +44 XXX XXX XXXX
|
| 152 |
+
0\d{4}\s\d{6} | # UK STD 0XXXX XXXXXX
|
| 153 |
+
0\d{3}\s\d{3}\s\d{4} | # UK STD 0XXX XXX XXXX
|
| 154 |
+
\+44\d{10} | # +44 XXXXXXXXXX
|
| 155 |
+
0\d{10} | # 0XXXXXXXXXX
|
| 156 |
+
\+61\s\d\s\d{4}\s\d{4} | # Australia Intl +61 X XXXX XXXX
|
| 157 |
+
0\d\s\d{4}\s\d{4} | # Australia STD 0X XXXX XXXX
|
| 158 |
+
\+61\d{9} | # +61 XXXXXXXXX
|
| 159 |
+
0\d{9} | # 0XXXXXXXXX
|
| 160 |
+
\+91\s\d{5}-\d{5} | # India Intl +91 XXXXX-XXXXX
|
| 161 |
+
\+91\s\d{4}-\d{6} | # India Intl +91 XXXX-XXXXXX
|
| 162 |
+
\+91\s\d{10} | # India Intl +91 XXXXXXXXXX
|
| 163 |
+
\+91\s\d{3}\s\d{3}\s\d{4} | # India Intl +91 XXX XXX XXXX
|
| 164 |
+
\+91\s\d{3}-\d{3}-\d{4} | # India Intl +91 XXX-XXX-XXXX
|
| 165 |
+
\+91\s\d{2}\s\d{4}\s\d{4} | # India Intl +91 XX XXXX XXXX
|
| 166 |
+
\+91\s\d{2}-\d{4}-\d{4} | # India Intl +91 XX-XXXX-XXXX
|
| 167 |
+
\+91\s\d{5}\s\d{5} | # India Intl +91 XXXXX XXXXX
|
| 168 |
+
\d{5}\s\d{5} | # India XXXXX XXXXX
|
| 169 |
+
\d{5}-\d{5} | # India XXXXX-XXXXX
|
| 170 |
+
0\d{2}-\d{7} | # India STD 0XX-XXXXXXX
|
| 171 |
+
\+91\d{10} | # +91 XXXXXXXXXX
|
| 172 |
+
\d{10} | # XXXXXXXXXX # Here is the regex to handle all possible combination of the contact
|
| 173 |
+
\d{6}-\d{4} | # XXXXXX-XXXX
|
| 174 |
+
\d{4}-\d{6} | # XXXX-XXXXXX
|
| 175 |
+
\d{3}\s\d{3}\s\d{4} | # XXX XXX XXXX
|
| 176 |
+
\d{3}-\d{3}-\d{4} | # XXX-XXX-XXXX
|
| 177 |
+
\d{4}\s\d{3}\s\d{3} | # XXXX XXX XXX
|
| 178 |
+
\d{4}-\d{3}-\d{3} | # XXXX-XXX-XXX #-----
|
| 179 |
+
\+49\s\d{4}\s\d{8} | # Germany Intl +49 XXXX XXXXXXXX
|
| 180 |
+
\+49\s\d{3}\s\d{7} | # Germany Intl +49 XXX XXXXXXX
|
| 181 |
+
0\d{3}\s\d{8} | # Germany STD 0XXX XXXXXXXX
|
| 182 |
+
\+49\d{12} | # +49 XXXXXXXXXXXX
|
| 183 |
+
\+49\d{10} | # +49 XXXXXXXXXX
|
| 184 |
+
0\d{11} | # 0XXXXXXXXXXX
|
| 185 |
+
\+86\s\d{3}\s\d{4}\s\d{4} | # China Intl +86 XXX XXXX XXXX
|
| 186 |
+
0\d{3}\s\d{4}\s\d{4} | # China STD 0XXX XXXX XXXX
|
| 187 |
+
\+86\d{11} | # +86 XXXXXXXXXXX
|
| 188 |
+
\+81\s\d\s\d{4}\s\d{4} | # Japan Intl +81 X XXXX XXXX
|
| 189 |
+
\+81\s\d{2}\s\d{4}\s\d{4} | # Japan Intl +81 XX XXXX XXXX
|
| 190 |
+
0\d\s\d{4}\s\d{4} | # Japan STD 0X XXXX XXXX
|
| 191 |
+
\+81\d{10} | # +81 XXXXXXXXXX
|
| 192 |
+
\+81\d{9} | # +81 XXXXXXXXX
|
| 193 |
+
0\d{9} | # 0XXXXXXXXX
|
| 194 |
+
\+55\s\d{2}\s\d{5}-\d{4} | # Brazil Intl +55 XX XXXXX-XXXX
|
| 195 |
+
\+55\s\d{2}\s\d{4}-\d{4} | # Brazil Intl +55 XX XXXX-XXXX
|
| 196 |
+
0\d{2}\s\d{4}\s\d{4} | # Brazil STD 0XX XXXX XXXX
|
| 197 |
+
\+55\d{11} | # +55 XXXXXXXXXXX
|
| 198 |
+
\+55\d{10} | # +55 XXXXXXXXXX
|
| 199 |
+
0\d{10} | # 0XXXXXXXXXX
|
| 200 |
+
\+33\s\d\s\d{2}\s\d{2}\s\d{2}\s\d{2} | # France Intl +33 X XX XX XX XX
|
| 201 |
+
0\d\s\d{2}\s\d{2}\s\d{2}\s\d{2} | # France STD 0X XX XX XX XX
|
| 202 |
+
\+33\d{9} | # +33 XXXXXXXXX
|
| 203 |
+
0\d{9} | # 0XXXXXXXXX
|
| 204 |
+
\+7\s\d{3}\s\d{3}-\d{2}-\d{2} | # Russia Intl +7 XXX XXX-XX-XX
|
| 205 |
+
8\s\d{3}\s\d{3}-\d{2}-\d{2} | # Russia STD 8 XXX XXX-XX-XX
|
| 206 |
+
\+7\d{10} | # +7 XXXXXXXXXX
|
| 207 |
+
8\d{10} | # 8 XXXXXXXXXX
|
| 208 |
+
\+27\s\d{2}\s\d{3}\s\d{4} | # South Africa Intl +27 XX XXX XXXX
|
| 209 |
+
0\d{2}\s\d{3}\s\d{4} | # South Africa STD 0XX XXX XXXX
|
| 210 |
+
\+27\d{9} | # +27 XXXXXXXXX
|
| 211 |
+
0\d{9} | # 0XXXXXXXXX
|
| 212 |
+
\+52\s\d{3}\s\d{3}\s\d{4} | # Mexico Intl +52 XXX XXX XXXX
|
| 213 |
+
\+52\s\d{2}\s\d{4}\s\d{4} | # Mexico Intl +52 XX XXXX XXXX
|
| 214 |
+
01\s\d{3}\s\d{4} | # Mexico STD 01 XXX XXXX
|
| 215 |
+
\+52\d{10} | # +52 XXXXXXXXXX
|
| 216 |
+
01\d{7} | # 01 XXXXXXX
|
| 217 |
+
\+234\s\d{3}\s\d{3}\s\d{4} | # Nigeria Intl +234 XXX XXX XXXX
|
| 218 |
+
0\d{3}\s\d{3}\s\d{4} | # Nigeria STD 0XXX XXX XXXX
|
| 219 |
+
\+234\d{10} | # +234 XXXXXXXXXX
|
| 220 |
+
0\d{10} | # 0XXXXXXXXXX
|
| 221 |
+
\+971\s\d\s\d{3}\s\d{4} | # UAE Intl +971 X XXX XXXX
|
| 222 |
+
0\d\s\d{3}\s\d{4} | # UAE STD 0X XXX XXXX
|
| 223 |
+
\+971\d{8} | # +971 XXXXXXXX
|
| 224 |
+
0\d{8} | # 0XXXXXXXX
|
| 225 |
+
\+54\s9\s\d{3}\s\d{3}\s\d{4} | # Argentina Intl +54 9 XXX XXX XXXX
|
| 226 |
+
\+54\s\d{1}\s\d{4}\s\d{4} | # Argentina Intl +54 X XXXX XXXX
|
| 227 |
+
0\d{3}\s\d{4} | # Argentina STD 0XXX XXXX
|
| 228 |
+
\+54\d{10} | # +54 9 XXXXXXXXXX
|
| 229 |
+
\+54\d{9} | # +54 XXXXXXXXX
|
| 230 |
+
0\d{7} | # 0XXXXXXX
|
| 231 |
+
\+966\s\d\s\d{3}\s\d{4} | # Saudi Intl +966 X XXX XXXX
|
| 232 |
+
0\d\s\d{3}\s\d{4} | # Saudi STD 0X XXX XXXX
|
| 233 |
+
\+966\d{8} | # +966 XXXXXXXX
|
| 234 |
+
0\d{8} | # 0XXXXXXXX
|
| 235 |
+
\+1\d{10} | # +1 XXXXXXXXXX
|
| 236 |
+
\+1\s\d{3}\s\d{3}\s\d{4} | # +1 XXX XXX XXXX
|
| 237 |
+
\d{5}\s\d{5} | # XXXXX XXXXX
|
| 238 |
+
\d{10} | # XXXXXXXXXX
|
| 239 |
+
\+44\d{10} | # +44 XXXXXXXXXX
|
| 240 |
+
0\d{10} | # 0XXXXXXXXXX
|
| 241 |
+
\+61\d{9} | # +61 XXXXXXXXX
|
| 242 |
+
0\d{9} | # 0XXXXXXXXX
|
| 243 |
+
\+91\d{10} | # +91 XXXXXXXXXX
|
| 244 |
+
\+49\d{12} | # +49 XXXXXXXXXXXX
|
| 245 |
+
\+49\d{10} | # +49 XXXXXXXXXX
|
| 246 |
+
0\d{11} | # 0XXXXXXXXXXX
|
| 247 |
+
\+86\d{11} | # +86 XXXXXXXXXXX
|
| 248 |
+
\+81\d{10} | # +81 XXXXXXXXXX
|
| 249 |
+
\+81\d{9} | # +81 XXXXXXXXX
|
| 250 |
+
0\d{9} | # 0XXXXXXXXX
|
| 251 |
+
\+55\d{11} | # +55 XXXXXXXXXXX
|
| 252 |
+
\+55\d{10} | # +55 XXXXXXXXXX
|
| 253 |
+
0\d{10} | # 0XXXXXXXXXX
|
| 254 |
+
\+33\d{9} | # +33 XXXXXXXXX
|
| 255 |
+
0\d{9} | # 0XXXXXXXXX
|
| 256 |
+
\+7\d{10} | # +7 XXXXXXXXXX
|
| 257 |
+
8\d{10} | # 8 XXXXXXXXXX
|
| 258 |
+
\+27\d{9} | # +27 XXXXXXXXX
|
| 259 |
+
0\d{9} | # 0XXXXXXXXX (South Africa STD)
|
| 260 |
+
\+52\d{10} | # +52 XXXXXXXXXX
|
| 261 |
+
01\d{7} | # 01 XXXXXXX
|
| 262 |
+
\+234\d{10} | # +234 XXXXXXXXXX
|
| 263 |
+
0\d{10} | # 0XXXXXXXXXX
|
| 264 |
+
\+971\d{8} | # +971 XXXXXXXX
|
| 265 |
+
0\d{8} | # 0XXXXXXXX
|
| 266 |
+
\+54\s9\s\d{10} | # +54 9 XXXXXXXXXX
|
| 267 |
+
\+54\d{9} | # +54 XXXXXXXXX
|
| 268 |
+
0\d{7} | # 0XXXXXXX
|
| 269 |
+
\+966\d{8} | # +966 XXXXXXXX
|
| 270 |
+
0\d{8} # 0XXXXXXXX
|
| 271 |
+
\+\d{3}-\d{3}-\d{4}
|
| 272 |
+
)
|
| 273 |
+
''',re.VERBOSE)
|
| 274 |
+
|
| 275 |
+
# Email regex
|
| 276 |
+
email_regex = re.compile(r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b')
|
| 277 |
+
|
| 278 |
+
# URL and links regex, updated to avoid conflicts with email domains
|
| 279 |
+
link_regex = re.compile(r'\b(?:https?:\/\/)?(?:www\.)[a-zA-Z0-9-]+\.(?:com|co\.in|co|io|org|net|edu|gov|mil|int|uk|us|in|de|au|app|tech|xyz|info|biz|fr|dev)\b')
|
| 280 |
+
|
| 281 |
+
# Find all matches in the text
|
| 282 |
+
phone_numbers = [num for num in combined_phone_regex.findall(text) if len(num) >= 5]
|
| 283 |
+
|
| 284 |
+
emails = email_regex.findall(text)
|
| 285 |
+
|
| 286 |
+
links_RE = [link for link in link_regex.findall(text) if len(link)>=11]
|
| 287 |
+
|
| 288 |
+
# Remove profile links that might conflict with emails
|
| 289 |
+
links_RE = [link for link in links_RE if not any(email in link for email in emails)]
|
| 290 |
+
|
| 291 |
+
return {
|
| 292 |
+
"phone_numbers": phone_numbers,
|
| 293 |
+
"emails": emails,
|
| 294 |
+
"links_RE": links_RE
|
| 295 |
+
}
|
| 296 |
+
|
| 297 |
+
# preprocessing the data
|
| 298 |
+
def process_extracted_text(extracted_text):
|
| 299 |
+
# Load JSON data
|
| 300 |
+
data = json.dumps(extracted_text, indent=4)
|
| 301 |
+
data = json.loads(data)
|
| 302 |
+
|
| 303 |
+
# Create a single dictionary to hold combined results
|
| 304 |
+
combined_results = {
|
| 305 |
+
"phone_numbers": [],
|
| 306 |
+
"emails": [],
|
| 307 |
+
"links_RE": []
|
| 308 |
+
}
|
| 309 |
+
|
| 310 |
+
# Process each text entry
|
| 311 |
+
for filename, text in data.items():
|
| 312 |
+
contact_details = extract_contact_details(text)
|
| 313 |
+
# Extend combined results with the details from this file
|
| 314 |
+
combined_results["phone_numbers"].extend(contact_details["phone_numbers"])
|
| 315 |
+
combined_results["emails"].extend(contact_details["emails"])
|
| 316 |
+
combined_results["links_RE"].extend(contact_details["links_RE"])
|
| 317 |
+
|
| 318 |
+
# Convert the combined results to JSON
|
| 319 |
+
#combined_results_json = json.dumps(combined_results, indent=4)
|
| 320 |
+
combined_results_json = combined_results
|
| 321 |
+
|
| 322 |
+
# Print the final JSON results
|
| 323 |
+
print("Combined contact details in JSON format:")
|
| 324 |
+
print(combined_results_json)
|
| 325 |
+
|
| 326 |
+
return combined_results_json
|
| 327 |
+
|
| 328 |
+
# Function to parse resume with SpaCy
|
| 329 |
+
|
| 330 |
+
def Parser_from_model(file_path):
|
| 331 |
+
# Initialize result with lists instead of strings for consistency
|
| 332 |
+
result = {
|
| 333 |
+
"personal": {
|
| 334 |
+
"name": [],
|
| 335 |
+
"contact": [],
|
| 336 |
+
"email": [],
|
| 337 |
+
"location": [],
|
| 338 |
+
"link": [],
|
| 339 |
+
"invalid_email": [],
|
| 340 |
+
"invalid_contact": []
|
| 341 |
+
},
|
| 342 |
+
"professional": {
|
| 343 |
+
"technical_skills": [],
|
| 344 |
+
"non_technical_skills": [],
|
| 345 |
+
"tools": [],
|
| 346 |
+
"experience": [
|
| 347 |
+
{
|
| 348 |
+
"company": [],
|
| 349 |
+
"projects": [],
|
| 350 |
+
"role": [],
|
| 351 |
+
"years": [],
|
| 352 |
+
"project_experience": []
|
| 353 |
+
}
|
| 354 |
+
],
|
| 355 |
+
"education": [
|
| 356 |
+
{
|
| 357 |
+
"qualification": [],
|
| 358 |
+
"university": [],
|
| 359 |
+
"course": [],
|
| 360 |
+
"certificate": []
|
| 361 |
+
}
|
| 362 |
+
]
|
| 363 |
+
}
|
| 364 |
+
}
|
| 365 |
+
|
| 366 |
+
try:
|
| 367 |
+
nlp = spacy.load("Spacy_Models/ner_model_05_3")
|
| 368 |
+
logging.debug("Model loaded successfully.")
|
| 369 |
+
except Exception as e:
|
| 370 |
+
logging.error(f"Error loading model: {e}")
|
| 371 |
+
return {"error": "Model loading failed"}
|
| 372 |
+
|
| 373 |
+
try:
|
| 374 |
+
cleaned_text, hyperlinks = extract_text_based_on_format(file_path)
|
| 375 |
+
if not cleaned_text.strip():
|
| 376 |
+
logging.error("No text extracted from the file.")
|
| 377 |
+
return {"error": "Text extraction failed"}
|
| 378 |
+
except Exception as e:
|
| 379 |
+
logging.error(f"Error extracting text from file: {e}")
|
| 380 |
+
return {"error": "Text extraction failed"}
|
| 381 |
+
|
| 382 |
+
try:
|
| 383 |
+
doc = nlp(cleaned_text)
|
| 384 |
+
except Exception as e:
|
| 385 |
+
logging.error(f"Error processing text with SpaCy: {e}")
|
| 386 |
+
return {"error": "Text processing failed"}
|
| 387 |
+
|
| 388 |
+
# Extracting Contact Details
|
| 389 |
+
cont_data = process_extracted_text(cleaned_text)
|
| 390 |
+
|
| 391 |
+
# Initialize entities as a dictionary with lists
|
| 392 |
+
entities = {label: [] for label in [
|
| 393 |
+
'PERSON', 'EMAIL', 'CONTACT', 'LOCATION', 'SKILL', 'SOFT_SKILL',
|
| 394 |
+
'COMPANY', 'PROJECTS', 'JOB_TITLE', 'YEARS_EXPERIENCE', 'EXPERIENCE',
|
| 395 |
+
'QUALIFICATION', 'UNIVERSITY', 'COURSE', 'CERTIFICATE'
|
| 396 |
+
]}
|
| 397 |
+
|
| 398 |
+
# Process entities and avoid duplicates
|
| 399 |
+
for ent in doc.ents:
|
| 400 |
+
if ent.label_ in entities and ent.text not in entities[ent.label_]:
|
| 401 |
+
entities[ent.label_].append(ent.text)
|
| 402 |
+
|
| 403 |
+
# Helper function to handle None or unexpected types
|
| 404 |
+
def normalize_to_list(value):
|
| 405 |
+
if value is []:
|
| 406 |
+
return None
|
| 407 |
+
elif isinstance(value, (str, int, float)):
|
| 408 |
+
return [str(value)]
|
| 409 |
+
elif isinstance(value, list):
|
| 410 |
+
return value
|
| 411 |
+
else:
|
| 412 |
+
return [str(value)]
|
| 413 |
+
|
| 414 |
+
# Map entities to result JSON
|
| 415 |
+
result['personal']['name'] = normalize_to_list(entities.get('PERSON'))
|
| 416 |
+
result['personal']['email'] += cont_data['emails']
|
| 417 |
+
result['personal']['invalid_email'] = []
|
| 418 |
+
|
| 419 |
+
# Validate email and handle invalid ones
|
| 420 |
+
for email in entities.get('EMAIL', []):
|
| 421 |
+
if is_valid_email(email):
|
| 422 |
+
result['personal']['email'].append(email)
|
| 423 |
+
else:
|
| 424 |
+
result['personal']['invalid_email'].append(email)
|
| 425 |
+
|
| 426 |
+
# Validate contact and handle invalid ones
|
| 427 |
+
result['personal']['contact'] += cont_data['phone_numbers']
|
| 428 |
+
result['personal']['invalid_contact'] = []
|
| 429 |
+
for contact in entities.get('CONTACT', []):
|
| 430 |
+
if is_valid_contact(contact):
|
| 431 |
+
result['personal']['contact'].append(contact)
|
| 432 |
+
else:
|
| 433 |
+
result['personal']['invalid_contact'].append(contact)
|
| 434 |
+
|
| 435 |
+
result['personal']['location'] = normalize_to_list(entities.get('LOCATION'))
|
| 436 |
+
result['personal']['link'] = normalize_to_list(hyperlinks)
|
| 437 |
+
result['personal']['link'] += cont_data['links_RE']
|
| 438 |
+
|
| 439 |
+
e
|
| 440 |
+
|
| 441 |
+
result['professional']['technical_skills'] = normalize_to_list(entities.get('SKILL'))
|
| 442 |
+
result['professional']['non_technical_skills'] = normalize_to_list(entities.get('SOFT_SKILL'))
|
| 443 |
+
result['professional']['tools'] = None # Logic for tools can be added if needed
|
| 444 |
+
|
| 445 |
+
experience = result['professional']['experience'][0]
|
| 446 |
+
experience['company'] = normalize_to_list(entities.get('COMPANY'))
|
| 447 |
+
experience['projects'] = normalize_to_list(entities.get('PROJECTS'))
|
| 448 |
+
experience['role'] = normalize_to_list(entities.get('JOB_TITLE'))
|
| 449 |
+
experience['years'] = normalize_to_list(entities.get('YEARS_EXPERIENCE'))
|
| 450 |
+
experience['project_experience'] = normalize_to_list(entities.get('EXPERIENCE'))
|
| 451 |
+
|
| 452 |
+
education = result['professional']['education'][0]
|
| 453 |
+
education['qualification'] = normalize_to_list(entities.get('QUALIFICATION'))
|
| 454 |
+
education['university'] = normalize_to_list(entities.get('UNIVERSITY'))
|
| 455 |
+
education['course'] = normalize_to_list(entities.get('COURSE'))
|
| 456 |
+
education['certificate'] = normalize_to_list(entities.get('CERTIFICATE'))
|
| 457 |
+
|
| 458 |
+
print(result)
|
| 459 |
+
return result
|
| 460 |
+
|
| 461 |
+
|