Spaces:
Sleeping
Sleeping
File size: 8,165 Bytes
f1b19d3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 |
"""
Email Intent Classifier Tool - Classify email intents using NLP
"""
import logging
from typing import Dict, Any, List
import re
import sys
import os
# Add parent directory to path for imports
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
logger = logging.getLogger(__name__)
class EmailIntentClassifier:
"""
Rule-based email intent classifier with confidence scoring
"""
# Define intent patterns (keywords and phrases)
INTENT_PATTERNS = {
"inquiry": [
r'\b(question|wondering|curious|clarification|information|details|help)\b',
r'\b(what|when|where|who|why|how)\b.*\?',
r'\b(could you|can you|would you).*\b(explain|tell|provide|share)\b'
],
"complaint": [
r'\b(complaint|issue|problem|disappointed|frustrated|unhappy|angry)\b',
r'\b(not working|broken|failed|error|mistake)\b',
r'\b(terrible|awful|worst|horrible|unacceptable)\b'
],
"request": [
r'\b(please|kindly|request|need|require|would like)\b',
r'\b(send|provide|share|give|deliver|forward)\b.*\b(me|us)\b',
r'\b(need|want|looking for)\b'
],
"feedback": [
r'\b(feedback|suggestion|recommend|improve|enhancement)\b',
r'\b(think|believe|feel|opinion)\b.*\b(should|could|would)\b',
r'\b(great|excellent|good|nice|appreciate|love)\b'
],
"meeting": [
r'\b(meeting|schedule|appointment|call|discuss|conference)\b',
r'\b(available|availability|free time|calendar)\b',
r'\b(reschedule|postpone|cancel|confirm)\b'
],
"order": [
r'\b(order|purchase|buy|payment|invoice|receipt)\b',
r'\b(shipping|delivery|tracking|status)\b',
r'\b(product|item|package)\b'
],
"urgent": [
r'\b(urgent|asap|immediately|critical|emergency|priority)\b',
r'\b(time-sensitive|deadline|due)\b',
r'!!+|\bIMPORTANT\b'
],
"follow_up": [
r'\b(follow up|following up|checking in|reminder)\b',
r'\b(haven\'t heard|waiting for|still pending)\b',
r'\b(previous|earlier|sent|mentioned)\b.*\b(email|message)\b'
],
"thank_you": [
r'\b(thank|thanks|grateful|appreciate|gratitude)\b',
r'\b(wonderful|excellent|helpful)\b.*\b(work|help|support)\b'
],
"application": [
r'\b(apply|application|position|job|role|opportunity)\b',
r'\b(resume|cv|cover letter|portfolio)\b',
r'\b(interested in|applying for)\b'
]
}
def classify(self, email_text: str) -> Dict[str, Any]:
"""
Classify email intent with confidence scores.
Args:
email_text: Email text to classify
Returns:
Dictionary with primary intent, confidence, and secondary intents
"""
if not email_text or not email_text.strip():
raise ValueError("Email text cannot be empty")
# Convert to lowercase for matching
text_lower = email_text.lower()
# Calculate scores for each intent
intent_scores = {}
for intent, patterns in self.INTENT_PATTERNS.items():
score = 0
matches = 0
for pattern in patterns:
found = re.findall(pattern, text_lower, re.IGNORECASE)
if found:
matches += len(found)
score += len(found)
# Normalize score
if score > 0:
intent_scores[intent] = min(score / 3.0, 1.0) # Cap at 1.0
# If no patterns matched, classify as "general"
if not intent_scores:
return {
"intent": "general",
"confidence": 0.5,
"secondary_intents": [],
"explanation": "No specific intent patterns detected"
}
# Sort by score
sorted_intents = sorted(intent_scores.items(), key=lambda x: x[1], reverse=True)
# Get primary intent
primary_intent = sorted_intents[0][0]
primary_confidence = sorted_intents[0][1]
# Get secondary intents (top 3)
secondary_intents = [
{"intent": intent, "confidence": round(score, 3)}
for intent, score in sorted_intents[1:4]
]
return {
"intent": primary_intent,
"confidence": round(primary_confidence, 3),
"secondary_intents": secondary_intents,
"explanation": f"Detected {primary_intent} intent based on keyword analysis"
}
def classify_email_intent(email_text: str) -> Dict[str, Any]:
"""
Classify the intent of an email.
Args:
email_text: Email text to classify
Returns:
Dictionary with classification results
"""
try:
classifier = EmailIntentClassifier()
result = classifier.classify(email_text)
# Add metadata
result["email_length"] = len(email_text)
result["word_count"] = len(email_text.split())
return result
except Exception as e:
logger.error(f"Error classifying email intent: {e}")
raise
def classify_batch_emails(emails: List[str]) -> Dict[str, Any]:
"""
Classify multiple emails at once.
Args:
emails: List of email text strings
Returns:
Dictionary with batch classification results
"""
try:
classifier = EmailIntentClassifier()
results = []
for idx, email_text in enumerate(emails):
try:
result = classifier.classify(email_text)
result["email_index"] = idx
results.append(result)
except Exception as e:
logger.error(f"Error classifying email {idx}: {e}")
results.append({
"email_index": idx,
"error": str(e),
"intent": "error",
"confidence": 0.0
})
# Aggregate statistics
intent_distribution = {}
for result in results:
intent = result.get("intent", "unknown")
intent_distribution[intent] = intent_distribution.get(intent, 0) + 1
return {
"total_emails": len(emails),
"results": results,
"intent_distribution": intent_distribution
}
except Exception as e:
logger.error(f"Error in batch email classification: {e}")
raise
def extract_email_features(email_text: str) -> Dict[str, Any]:
"""
Extract features from an email for analysis.
Args:
email_text: Email text
Returns:
Dictionary with extracted features
"""
try:
features = {
"length": len(email_text),
"word_count": len(email_text.split()),
"sentence_count": len(re.split(r'[.!?]+', email_text)),
"has_greeting": bool(re.search(r'\b(hi|hello|dear|hey)\b', email_text.lower())),
"has_closing": bool(re.search(r'\b(regards|sincerely|thanks|best)\b', email_text.lower())),
"question_count": len(re.findall(r'\?', email_text)),
"exclamation_count": len(re.findall(r'!', email_text)),
"has_url": bool(re.search(r'https?://', email_text)),
"has_email_address": bool(re.search(r'\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Z|a-z]{2,}\b', email_text))
}
return features
except Exception as e:
logger.error(f"Error extracting email features: {e}")
raise
|