Spaces:
Running
Running
File size: 12,284 Bytes
24f95f0 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 | """
Domain Classifier for Janus Self-Improvement System.
Classifies queries into specific domains to enable:
1. Domain-specific model routing
2. Specialized prompt selection
3. Targeted knowledge retrieval
4. Expert system activation
"""
import re
import logging
from typing import Dict, List, Tuple, Optional
from dataclasses import dataclass
from enum import Enum
logger = logging.getLogger(__name__)
# Import existing domain packs
try:
from app.domain_packs import registry as domain_registry
DOMAIN_PACKS_AVAILABLE = True
except ImportError:
DOMAIN_PACKS_AVAILABLE = False
logger.warning("Domain packs not available, using fallback classification")
class DomainType(Enum):
"""Supported domain types for classification."""
FINANCE = "finance"
TECHNOLOGY = "technology"
HEALTHCARE = "healthcare"
POLICY = "policy"
SCIENCE = "science"
GEOPOLITICS = "geopolitics"
ENERGY = "energy"
CRITICAL_THINKING = "critical_thinking"
EMOTIONAL_INTELLIGENCE = "emotional_intelligence"
PHILOSOPHY = "philosophy"
BUSINESS = "business"
EDUCATION = "education"
GENERAL = "general"
@dataclass
class DomainClassification:
"""Result of domain classification."""
domain: DomainType
confidence: float
keywords_found: List[str]
reasoning: str
class DomainClassifier:
"""
Classifies user queries into specific domains for improved routing
and specialized processing.
"""
def __init__(self):
"""Initialize the domain classifier."""
self.domain_keywords = self._load_domain_keywords()
self.compiled_patterns = self._compile_domain_patterns()
def _load_domain_keywords(self) -> Dict[DomainType, List[str]]:
"""Load domain-specific keywords from domain packs or fallback."""
if DOMAIN_PACKS_AVAILABLE:
return self._load_from_domain_packs()
else:
return self._load_fallback_keywords()
def _load_from_domain_packs(self) -> Dict[DomainType, List[str]]:
"""Load keywords from existing domain packs."""
keywords = {}
# Map domain pack names to our DomainType enum
domain_mapping = {
"finance": DomainType.FINANCE,
"technology": DomainType.TECHNOLOGY,
"healthcare": DomainType.HEALTHCARE,
"policy": DomainType.POLICY,
"science": DomainType.SCIENCE,
"geopolitics": DomainType.GEOPOLITICS,
"energy": DomainType.ENERGY,
"critical_thinking": DomainType.CRITICAL_THINKING,
"emotional_intelligence": DomainType.EMOTIONAL_INTELLIGENCE,
"philosophy": DomainType.PHILOSOPHY,
"business": DomainType.BUSINESS,
"education": DomainType.EDUCATION,
}
try:
registry = domain_registry.get_registry()
except Exception as e:
logger.warning(f"Could not get domain registry: {e}")
registry = None
for pack_name, domain_type in domain_mapping.items():
try:
if registry is not None:
pack = registry.get_pack(pack_name)
else:
pack = None
if pack and hasattr(pack, "get_keywords"):
keywords[domain_type] = pack.get_keywords()
elif pack and hasattr(pack, "DOMAIN_KEYWORDS"):
keywords[domain_type] = pack.DOMAIN_KEYWORDS
else:
# Fallback to empty list if no keywords method
keywords[domain_type] = []
except Exception as e:
logger.warning(f"Could not load keywords for {pack_name}: {e}")
keywords[domain_type] = []
# Always include general domain
keywords[DomainType.GENERAL] = []
return keywords
def _load_fallback_keywords(self) -> Dict[DomainType, List[str]]:
"""Load fallback keyword definitions."""
# Import from query_classifier as fallback
try:
from app.services.query_classifier import DOMAIN_KEYWORDS
keywords = {}
for domain_str, word_list in DOMAIN_KEYWORDS.items():
try:
domain_type = DomainType(domain_str)
keywords[domain_type] = word_list
except ValueError:
# Skip domains not in our enum
continue
keywords[DomainType.GENERAL] = []
return keywords
except ImportError:
# Ultimate fallback - minimal keyword sets
logger.warning("Using ultimate fallback domain keywords")
return {
DomainType.FINANCE: [
"stock",
"market",
"investment",
"trading",
"finance",
"economic",
],
DomainType.TECHNOLOGY: [
"ai",
"software",
"technology",
"programming",
"computer",
],
DomainType.HEALTHCARE: [
"health",
"medical",
"disease",
"treatment",
"healthcare",
],
DomainType.POLICY: [
"policy",
"government",
"law",
"regulation",
"politics",
],
DomainType.SCIENCE: [
"science",
"research",
"study",
"experiment",
"theory",
],
DomainType.GEOPOLITICS: [
"war",
"conflict",
"country",
"international",
"diplomacy",
],
DomainType.ENERGY: [
"energy",
"power",
"electricity",
"oil",
"gas",
"renewable",
],
DomainType.CRITICAL_THINKING: [
"analyze",
"evaluate",
"critique",
"assess",
"reason",
],
DomainType.EMOTIONAL_INTELLIGENCE: [
"feel",
"emotion",
"relationship",
"communication",
],
DomainType.PHILOSOPHY: [
"ethics",
"meaning",
"purpose",
"existence",
"consciousness",
],
DomainType.BUSINESS: [
"business",
"company",
"management",
"strategy",
"marketing",
],
DomainType.EDUCATION: [
"learn",
"education",
"study",
"teach",
"student",
],
DomainType.GENERAL: [],
}
def _compile_domain_patterns(self) -> Dict[DomainType, List[re.Pattern]]:
"""Compile regex patterns for each domain."""
patterns = {}
for domain, keywords in self.domain_keywords.items():
domain_patterns = []
for keyword in keywords:
# Create word boundary pattern for accurate matching
pattern = re.compile(rf"\b{re.escape(keyword)}\b", re.IGNORECASE)
domain_patterns.append(pattern)
patterns[domain] = domain_patterns
return patterns
def classify(self, query: str) -> DomainClassification:
"""
Classify a query into a domain.
Args:
query: User query string
Returns:
DomainClassification with domain, confidence, and details
"""
query_lower = query.lower().strip()
# Score each domain
domain_scores = {}
matched_keywords = {}
for domain, patterns in self.compiled_patterns.items():
score = 0
matched = []
for pattern in patterns:
matches = pattern.findall(query_lower)
if matches:
score += len(matches)
matched.extend(matches)
if score > 0:
domain_scores[domain] = score
matched_keywords[domain] = list(set(matched)) # Remove duplicates
# Determine domain
if not domain_scores:
# No domain keywords found - default to general
domain_scores[DomainType.GENERAL] = 1
matched_keywords[DomainType.GENERAL] = []
detected_domain = DomainType.GENERAL
confidence = 0.5 # Low confidence for default
reasoning = "No domain-specific keywords found"
else:
# Find domain with highest score
detected_domain = max(domain_scores, key=domain_scores.get)
max_score = domain_scores[detected_domain]
# Calculate confidence based on score relative to query length
query_word_count = max(len(query.split()), 1)
raw_confidence = min(max_score / (query_word_count * 0.5), 1.0) # Normalize
# Boost confidence if we found multiple keyword matches
keyword_bonus = min(len(matched_keywords[detected_domain]) * 0.1, 0.3)
confidence = min(raw_confidence + keyword_bonus, 0.95)
# Ensure minimum confidence for detected domain
confidence = max(confidence, 0.6)
reasoning = f"Found {max_score} keyword matches for {detected_domain.value}"
return DomainClassification(
domain=detected_domain,
confidence=confidence,
keywords_found=matched_keywords.get(detected_domain, []),
reasoning=reasoning,
)
def get_domain_confidence(self, query: str, domain: DomainType) -> float:
"""
Get confidence score for a specific domain.
Args:
query: User query string
domain: Domain to check confidence for
Returns:
Confidence score between 0.0 and 1.0
"""
classification = self.classify(query)
if classification.domain == domain:
return classification.confidence
else:
# Return inverse confidence for other domains
return max(0.0, 1.0 - classification.confidence)
def get_top_domains(
self, query: str, top_n: int = 3
) -> List[Tuple[DomainType, float]]:
"""
Get top N domain classifications for a query.
Args:
query: User query string
top_n: Number of top domains to return
Returns:
List of (domain, confidence) tuples sorted by confidence
"""
query_lower = query.lower().strip()
domain_scores = {}
# Score each domain
for domain, patterns in self.compiled_patterns.items():
score = 0
for pattern in patterns:
matches = pattern.findall(query_lower)
score += len(matches)
if score > 0:
domain_scores[domain] = score
if not domain_scores:
return [(DomainType.GENERAL, 0.5)]
# Sort by score and normalize to confidence
sorted_domains = sorted(domain_scores.items(), key=lambda x: x[1], reverse=True)
max_score = sorted_domains[0][1] if sorted_domains else 1
results = []
for domain, score in sorted_domains[:top_n]:
# Normalize score to 0-1 range with minimum confidence
confidence = max(0.5, min(score / (max_score * 0.5), 0.95))
results.append((domain, confidence))
return results
# Global instance for easy access
domain_classifier = DomainClassifier()
|