Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,10 +1,10 @@
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
"""
|
| 3 |
-
Email Validator
|
| 4 |
-
|
| 5 |
"""
|
| 6 |
|
| 7 |
-
|
| 8 |
import socket
|
| 9 |
import time
|
| 10 |
import uuid
|
|
@@ -13,13 +13,11 @@ import dns.resolver
|
|
| 13 |
import smtplib
|
| 14 |
import logging
|
| 15 |
|
| 16 |
-
app = Flask(__name__)
|
| 17 |
logging.basicConfig(level=logging.INFO)
|
| 18 |
|
| 19 |
FROM_ADDRESS = 'verifier@yourdomain.com'
|
| 20 |
SMTP_TIMEOUT = 15
|
| 21 |
RATE_LIMIT_SECONDS = 1.0
|
| 22 |
-
MAX_BATCH_SIZE = 100
|
| 23 |
|
| 24 |
# Domain cache
|
| 25 |
domain_cache = {}
|
|
@@ -80,11 +78,11 @@ def smtp_probe_recipient(mx_host: str, from_addr: str, recipient: str):
|
|
| 80 |
pass
|
| 81 |
return False, None, None, 'SMTPRecipientsRefused'
|
| 82 |
|
| 83 |
-
except (smtplib.SMTPConnectError, smtplib.SMTPServerDisconnected, socket.timeout, ConnectionRefusedError, OSError)
|
| 84 |
-
return False, None, None,
|
| 85 |
|
| 86 |
-
except Exception
|
| 87 |
-
return False, None, None,
|
| 88 |
|
| 89 |
def decide_outcome(email, mx_host, rcpt_code, catch_all_flag, probe_error):
|
| 90 |
"""Determine if email is valid or invalid"""
|
|
@@ -94,39 +92,33 @@ def decide_outcome(email, mx_host, rcpt_code, catch_all_flag, probe_error):
|
|
| 94 |
if not mx_host:
|
| 95 |
return 'INVALID', 'UNDELIVERABLE_NO_MX'
|
| 96 |
|
| 97 |
-
if probe_error:
|
| 98 |
return 'INVALID', 'PROBE_FAILED_NETWORK'
|
| 99 |
|
| 100 |
-
if rcpt_code is None:
|
| 101 |
-
return 'INVALID', 'PROBE_FAILED_NETWORK'
|
| 102 |
-
|
| 103 |
-
# 5xx = definitive rejection
|
| 104 |
if 500 <= rcpt_code <= 599:
|
| 105 |
return 'INVALID', 'MAILBOX_DOES_NOT_EXIST'
|
| 106 |
|
| 107 |
-
# 2xx = accepted
|
| 108 |
if 200 <= rcpt_code <= 299:
|
| 109 |
if catch_all_flag:
|
| 110 |
return 'VALID', 'UNKNOWN_CATCHALL'
|
| 111 |
else:
|
| 112 |
return 'VALID', 'MAILBOX_EXISTS'
|
| 113 |
|
| 114 |
-
# 4xx = temporary error
|
| 115 |
if 400 <= rcpt_code <= 499:
|
| 116 |
return 'INVALID', 'TEMPORARY_ERROR'
|
| 117 |
|
| 118 |
return 'INVALID', 'UNKNOWN_STATUS'
|
| 119 |
|
| 120 |
-
def
|
| 121 |
-
"""
|
| 122 |
email = email.strip().lower()
|
| 123 |
|
| 124 |
if not is_valid_format(email):
|
| 125 |
return {
|
| 126 |
-
'
|
| 127 |
-
'
|
| 128 |
-
'
|
| 129 |
-
'
|
| 130 |
}
|
| 131 |
|
| 132 |
domain = email.split('@')[1].lower()
|
|
@@ -140,10 +132,10 @@ def probe_email(email: str, domain_cache: Dict):
|
|
| 140 |
|
| 141 |
if not hosts:
|
| 142 |
return {
|
| 143 |
-
'
|
| 144 |
-
'
|
| 145 |
-
'
|
| 146 |
-
'
|
| 147 |
}
|
| 148 |
|
| 149 |
mx_host = hosts[0]
|
|
@@ -158,8 +150,6 @@ def probe_email(email: str, domain_cache: Dict):
|
|
| 158 |
catch_all = False
|
| 159 |
if ok and code and 200 <= code <= 299:
|
| 160 |
catch_all = True
|
| 161 |
-
elif not ok and err and 'NETWORK_ERROR' in str(err):
|
| 162 |
-
catch_all = None
|
| 163 |
|
| 164 |
domain_cache[domain]['catch_all'] = catch_all
|
| 165 |
time.sleep(RATE_LIMIT_SECONDS)
|
|
@@ -177,135 +167,113 @@ def probe_email(email: str, domain_cache: Dict):
|
|
| 177 |
probe_error=(not ok or err is not None)
|
| 178 |
)
|
| 179 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 180 |
return {
|
| 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 |
-
'smtp_code': result['smtp_code']
|
| 233 |
-
}), 200
|
| 234 |
|
| 235 |
-
|
| 236 |
-
|
| 237 |
-
|
| 238 |
-
|
| 239 |
-
|
| 240 |
-
|
| 241 |
-
|
| 242 |
-
@app.route('/batch-validate', methods=['POST'])
|
| 243 |
-
def batch_validate():
|
| 244 |
-
"""Validate multiple emails"""
|
| 245 |
-
try:
|
| 246 |
-
data = request.get_json()
|
| 247 |
|
| 248 |
-
|
| 249 |
-
|
| 250 |
-
'error': 'Missing emails parameter',
|
| 251 |
-
'example': {'emails': ['test@gmail.com', 'user@yahoo.com']}
|
| 252 |
-
}), 400
|
| 253 |
|
| 254 |
-
|
| 255 |
-
|
| 256 |
-
|
| 257 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 258 |
|
| 259 |
-
|
| 260 |
-
|
|
|
|
|
|
|
| 261 |
|
| 262 |
-
|
| 263 |
-
|
|
|
|
|
|
|
| 264 |
|
| 265 |
-
|
| 266 |
-
|
| 267 |
-
|
| 268 |
-
|
| 269 |
-
|
| 270 |
-
|
| 271 |
-
|
| 272 |
-
|
| 273 |
-
|
| 274 |
-
|
| 275 |
-
is_valid = result['status'] == 'VALID'
|
| 276 |
-
|
| 277 |
-
results.append({
|
| 278 |
-
'email': result['email'],
|
| 279 |
-
'valid': is_valid,
|
| 280 |
-
'status': result['status'],
|
| 281 |
-
'reason': result['reason']
|
| 282 |
-
})
|
| 283 |
-
|
| 284 |
-
if is_valid:
|
| 285 |
-
valid_count += 1
|
| 286 |
-
|
| 287 |
-
return jsonify({
|
| 288 |
-
'total': len(emails),
|
| 289 |
-
'valid_count': valid_count,
|
| 290 |
-
'invalid_count': len(emails) - valid_count,
|
| 291 |
-
'results': results
|
| 292 |
-
}), 200
|
| 293 |
-
|
| 294 |
-
except Exception as e:
|
| 295 |
-
logging.error(f"Error in /batch-validate: {str(e)}")
|
| 296 |
-
return jsonify({
|
| 297 |
-
'error': 'Internal server error',
|
| 298 |
-
'message': str(e)
|
| 299 |
-
}), 500
|
| 300 |
-
|
| 301 |
-
@app.route('/health', methods=['GET'])
|
| 302 |
-
def health():
|
| 303 |
-
"""Health check endpoint"""
|
| 304 |
-
return jsonify({
|
| 305 |
-
'status': 'healthy',
|
| 306 |
-
'service': 'Email Validator API',
|
| 307 |
-
'version': '2.0'
|
| 308 |
-
}), 200
|
| 309 |
|
| 310 |
-
if __name__ ==
|
| 311 |
-
|
|
|
|
| 1 |
#!/usr/bin/env python3
|
| 2 |
"""
|
| 3 |
+
Email Validator - Hugging Face Spaces (Gradio Version)
|
| 4 |
+
Works perfectly on Hugging Face because Gradio is native
|
| 5 |
"""
|
| 6 |
|
| 7 |
+
import gradio as gr
|
| 8 |
import socket
|
| 9 |
import time
|
| 10 |
import uuid
|
|
|
|
| 13 |
import smtplib
|
| 14 |
import logging
|
| 15 |
|
|
|
|
| 16 |
logging.basicConfig(level=logging.INFO)
|
| 17 |
|
| 18 |
FROM_ADDRESS = 'verifier@yourdomain.com'
|
| 19 |
SMTP_TIMEOUT = 15
|
| 20 |
RATE_LIMIT_SECONDS = 1.0
|
|
|
|
| 21 |
|
| 22 |
# Domain cache
|
| 23 |
domain_cache = {}
|
|
|
|
| 78 |
pass
|
| 79 |
return False, None, None, 'SMTPRecipientsRefused'
|
| 80 |
|
| 81 |
+
except (smtplib.SMTPConnectError, smtplib.SMTPServerDisconnected, socket.timeout, ConnectionRefusedError, OSError):
|
| 82 |
+
return False, None, None, 'NETWORK_ERROR'
|
| 83 |
|
| 84 |
+
except Exception:
|
| 85 |
+
return False, None, None, 'ERROR'
|
| 86 |
|
| 87 |
def decide_outcome(email, mx_host, rcpt_code, catch_all_flag, probe_error):
|
| 88 |
"""Determine if email is valid or invalid"""
|
|
|
|
| 92 |
if not mx_host:
|
| 93 |
return 'INVALID', 'UNDELIVERABLE_NO_MX'
|
| 94 |
|
| 95 |
+
if probe_error or rcpt_code is None:
|
| 96 |
return 'INVALID', 'PROBE_FAILED_NETWORK'
|
| 97 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 98 |
if 500 <= rcpt_code <= 599:
|
| 99 |
return 'INVALID', 'MAILBOX_DOES_NOT_EXIST'
|
| 100 |
|
|
|
|
| 101 |
if 200 <= rcpt_code <= 299:
|
| 102 |
if catch_all_flag:
|
| 103 |
return 'VALID', 'UNKNOWN_CATCHALL'
|
| 104 |
else:
|
| 105 |
return 'VALID', 'MAILBOX_EXISTS'
|
| 106 |
|
|
|
|
| 107 |
if 400 <= rcpt_code <= 499:
|
| 108 |
return 'INVALID', 'TEMPORARY_ERROR'
|
| 109 |
|
| 110 |
return 'INVALID', 'UNKNOWN_STATUS'
|
| 111 |
|
| 112 |
+
def validate_email_function(email: str) -> Dict:
|
| 113 |
+
"""Validate a single email"""
|
| 114 |
email = email.strip().lower()
|
| 115 |
|
| 116 |
if not is_valid_format(email):
|
| 117 |
return {
|
| 118 |
+
'Email': email,
|
| 119 |
+
'Status': '❌ INVALID',
|
| 120 |
+
'Reason': 'Invalid format',
|
| 121 |
+
'Valid': 'No'
|
| 122 |
}
|
| 123 |
|
| 124 |
domain = email.split('@')[1].lower()
|
|
|
|
| 132 |
|
| 133 |
if not hosts:
|
| 134 |
return {
|
| 135 |
+
'Email': email,
|
| 136 |
+
'Status': '❌ INVALID',
|
| 137 |
+
'Reason': 'No MX records',
|
| 138 |
+
'Valid': 'No'
|
| 139 |
}
|
| 140 |
|
| 141 |
mx_host = hosts[0]
|
|
|
|
| 150 |
catch_all = False
|
| 151 |
if ok and code and 200 <= code <= 299:
|
| 152 |
catch_all = True
|
|
|
|
|
|
|
| 153 |
|
| 154 |
domain_cache[domain]['catch_all'] = catch_all
|
| 155 |
time.sleep(RATE_LIMIT_SECONDS)
|
|
|
|
| 167 |
probe_error=(not ok or err is not None)
|
| 168 |
)
|
| 169 |
|
| 170 |
+
is_valid = status == 'VALID'
|
| 171 |
+
status_display = '✅ VALID' if is_valid else '❌ INVALID'
|
| 172 |
+
valid_text = 'Yes' if is_valid else 'No'
|
| 173 |
+
|
| 174 |
return {
|
| 175 |
+
'Email': email,
|
| 176 |
+
'Status': status_display,
|
| 177 |
+
'Reason': reason,
|
| 178 |
+
'Valid': valid_text
|
| 179 |
}
|
| 180 |
|
| 181 |
+
def validate_batch(emails_text: str) -> str:
|
| 182 |
+
"""Validate multiple emails from text input"""
|
| 183 |
+
emails = [e.strip() for e in emails_text.split('\n') if e.strip()]
|
| 184 |
+
|
| 185 |
+
if not emails:
|
| 186 |
+
return "Please enter at least one email"
|
| 187 |
+
|
| 188 |
+
results = []
|
| 189 |
+
valid_count = 0
|
| 190 |
+
|
| 191 |
+
for email in emails:
|
| 192 |
+
result = validate_email_function(email)
|
| 193 |
+
results.append(result)
|
| 194 |
+
if result['Valid'] == 'Yes':
|
| 195 |
+
valid_count += 1
|
| 196 |
+
|
| 197 |
+
# Format output
|
| 198 |
+
output = f"**Total: {len(emails)} | Valid: {valid_count} | Invalid: {len(emails) - valid_count}**\n\n"
|
| 199 |
+
output += "| Email | Status | Reason | Valid |\n"
|
| 200 |
+
output += "|-------|--------|--------|-------|\n"
|
| 201 |
+
|
| 202 |
+
for r in results:
|
| 203 |
+
output += f"| {r['Email']} | {r['Status']} | {r['Reason']} | {r['Valid']} |\n"
|
| 204 |
+
|
| 205 |
+
return output
|
| 206 |
|
| 207 |
+
# Create Gradio Interface
|
| 208 |
+
with gr.Blocks(title="Email Validator") as demo:
|
| 209 |
+
gr.Markdown("# 📧 Email Validator API")
|
| 210 |
+
gr.Markdown("Validate emails using DNS MX records and SMTP probing")
|
| 211 |
+
|
| 212 |
+
with gr.Tab("Single Email"):
|
| 213 |
+
with gr.Row():
|
| 214 |
+
email_input = gr.Textbox(
|
| 215 |
+
label="Email Address",
|
| 216 |
+
placeholder="Enter email to validate",
|
| 217 |
+
value="test@gmail.com"
|
| 218 |
+
)
|
| 219 |
|
| 220 |
+
validate_btn = gr.Button("Validate", variant="primary", size="lg")
|
| 221 |
|
| 222 |
+
with gr.Row():
|
| 223 |
+
email_output = gr.Textbox(label="Email", interactive=False)
|
| 224 |
+
status_output = gr.Textbox(label="Status", interactive=False)
|
| 225 |
|
| 226 |
+
with gr.Row():
|
| 227 |
+
reason_output = gr.Textbox(label="Reason", interactive=False)
|
| 228 |
+
valid_output = gr.Textbox(label="Valid", interactive=False)
|
| 229 |
|
| 230 |
+
validate_btn.click(
|
| 231 |
+
fn=validate_email_function,
|
| 232 |
+
inputs=email_input,
|
| 233 |
+
outputs=[email_output, status_output, reason_output, valid_output]
|
| 234 |
+
)
|
|
|
|
|
|
|
| 235 |
|
| 236 |
+
with gr.Tab("Batch Validation"):
|
| 237 |
+
with gr.Row():
|
| 238 |
+
batch_input = gr.Textbox(
|
| 239 |
+
label="Emails (one per line)",
|
| 240 |
+
placeholder="test1@gmail.com\ntest2@yahoo.com\ninvalid@test.com",
|
| 241 |
+
lines=10
|
| 242 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 243 |
|
| 244 |
+
batch_btn = gr.Button("Validate Batch", variant="primary", size="lg")
|
| 245 |
+
batch_output = gr.Markdown(label="Results")
|
|
|
|
|
|
|
|
|
|
| 246 |
|
| 247 |
+
batch_btn.click(
|
| 248 |
+
fn=validate_batch,
|
| 249 |
+
inputs=batch_input,
|
| 250 |
+
outputs=batch_output
|
| 251 |
+
)
|
| 252 |
+
|
| 253 |
+
with gr.Tab("API Docs"):
|
| 254 |
+
gr.Markdown("""
|
| 255 |
+
## API Endpoints
|
| 256 |
|
| 257 |
+
### Single Email Validation
|
| 258 |
+
- **URL:** `/api/validate`
|
| 259 |
+
- **Method:** POST
|
| 260 |
+
- **Payload:** `{"email": "test@gmail.com"}`
|
| 261 |
|
| 262 |
+
### Batch Validation
|
| 263 |
+
- **URL:** `/api/batch`
|
| 264 |
+
- **Method:** POST
|
| 265 |
+
- **Payload:** `{"emails": ["test1@gmail.com", "test2@yahoo.com"]}`
|
| 266 |
|
| 267 |
+
## Response
|
| 268 |
+
```
|
| 269 |
+
{
|
| 270 |
+
"Email": "test@gmail.com",
|
| 271 |
+
"Status": "✅ VALID",
|
| 272 |
+
"Reason": "MAILBOX_EXISTS",
|
| 273 |
+
"Valid": "Yes"
|
| 274 |
+
}
|
| 275 |
+
```
|
| 276 |
+
""")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 277 |
|
| 278 |
+
if __name__ == "__main__":
|
| 279 |
+
demo.launch(server_name="0.0.0.0", server_port=7860, share=False)
|