Spaces:
Runtime error
Runtime error
File size: 17,041 Bytes
dcb2a99 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 |
"""Unified reasoning engine that combines multiple reasoning strategies."""
import logging
from typing import Dict, Any, List, Optional, Set, Union, Type
import json
from dataclasses import dataclass, field
from enum import Enum
from datetime import datetime
import asyncio
from collections import defaultdict
from .base import ReasoningStrategy
from .chain_of_thought import ChainOfThoughtStrategy
from .tree_of_thoughts import TreeOfThoughtsStrategy
from .meta_learning import MetaLearningStrategy
from .recursive import RecursiveReasoning
from .analogical import AnalogicalReasoning
from .local_llm import LocalLLMStrategy
from .agentic import (
TaskDecompositionStrategy,
ResourceManagementStrategy,
ContextualPlanningStrategy,
AdaptiveExecutionStrategy,
FeedbackIntegrationStrategy
)
class StrategyType(str, Enum):
"""Types of reasoning strategies."""
CHAIN_OF_THOUGHT = "chain_of_thought"
TREE_OF_THOUGHTS = "tree_of_thoughts"
META_LEARNING = "meta_learning"
RECURSIVE = "recursive"
ANALOGICAL = "analogical"
TASK_DECOMPOSITION = "task_decomposition"
RESOURCE_MANAGEMENT = "resource_management"
CONTEXTUAL_PLANNING = "contextual_planning"
ADAPTIVE_EXECUTION = "adaptive_execution"
FEEDBACK_INTEGRATION = "feedback_integration"
LOCAL_LLM = "local_llm"
@dataclass
class StrategyResult:
"""Result from a reasoning strategy."""
strategy_type: StrategyType
success: bool
answer: Optional[str]
confidence: float
reasoning_trace: List[Dict[str, Any]]
metadata: Dict[str, Any]
performance_metrics: Dict[str, Any]
timestamp: datetime = field(default_factory=datetime.now)
@dataclass
class UnifiedResult:
"""Combined result from multiple strategies."""
success: bool
answer: str
confidence: float
strategy_results: Dict[StrategyType, StrategyResult]
synthesis_method: str
meta_insights: List[str]
performance_metrics: Dict[str, Any]
timestamp: datetime = field(default_factory=datetime.now)
class UnifiedReasoningEngine:
"""
Advanced unified reasoning engine that:
1. Combines multiple reasoning strategies
2. Dynamically selects and weights strategies
3. Synthesizes results from different approaches
4. Learns from experience
5. Adapts to different types of tasks
"""
def __init__(self,
min_confidence: float = 0.7,
strategy_weights: Optional[Dict[StrategyType, float]] = None,
parallel_threshold: int = 3,
learning_rate: float = 0.1):
self.min_confidence = min_confidence
self.parallel_threshold = parallel_threshold
self.learning_rate = learning_rate
# Initialize strategies
self.strategies: Dict[StrategyType, ReasoningStrategy] = {
StrategyType.CHAIN_OF_THOUGHT: ChainOfThoughtStrategy(),
StrategyType.TREE_OF_THOUGHTS: TreeOfThoughtsStrategy(),
StrategyType.META_LEARNING: MetaLearningStrategy(),
StrategyType.RECURSIVE: RecursiveReasoning(),
StrategyType.ANALOGICAL: AnalogicalReasoning(),
StrategyType.TASK_DECOMPOSITION: TaskDecompositionStrategy(),
StrategyType.RESOURCE_MANAGEMENT: ResourceManagementStrategy(),
StrategyType.CONTEXTUAL_PLANNING: ContextualPlanningStrategy(),
StrategyType.ADAPTIVE_EXECUTION: AdaptiveExecutionStrategy(),
StrategyType.FEEDBACK_INTEGRATION: FeedbackIntegrationStrategy(),
StrategyType.LOCAL_LLM: LocalLLMStrategy() # Add local LLM strategy
}
# Strategy weights with higher weight for LOCAL_LLM
self.strategy_weights = strategy_weights or {
**{strategy_type: 1.0 for strategy_type in StrategyType},
StrategyType.LOCAL_LLM: 2.0 # Higher weight for local LLM
}
# Performance tracking
self.strategy_performance: Dict[StrategyType, List[float]] = defaultdict(list)
self.task_type_performance: Dict[str, Dict[StrategyType, float]] = defaultdict(lambda: defaultdict(float))
self.synthesis_performance: Dict[str, List[float]] = defaultdict(list)
async def reason(self, query: str, context: Dict[str, Any]) -> UnifiedResult:
"""Main reasoning method combining multiple strategies."""
try:
# Analyze task
task_analysis = await self._analyze_task(query, context)
# Select strategies
selected_strategies = await self._select_strategies(task_analysis, context)
# Execute strategies
strategy_results = await self._execute_strategies(
selected_strategies, query, context)
# Synthesize results
unified_result = await self._synthesize_results(
strategy_results, task_analysis, context)
# Learn from experience
self._update_performance(unified_result)
return unified_result
except Exception as e:
logging.error(f"Error in unified reasoning: {str(e)}")
return UnifiedResult(
success=False,
answer=f"Error: {str(e)}",
confidence=0.0,
strategy_results={},
synthesis_method="failed",
meta_insights=[f"Error occurred: {str(e)}"],
performance_metrics={}
)
async def _analyze_task(self, query: str, context: Dict[str, Any]) -> Dict[str, Any]:
"""Analyze the task to determine optimal strategy selection."""
prompt = f"""
Analyze reasoning task:
Query: {query}
Context: {json.dumps(context)}
Determine:
1. Task type and complexity
2. Required reasoning capabilities
3. Resource requirements
4. Success criteria
5. Risk factors
Format as:
[Analysis]
Type: ...
Complexity: ...
Capabilities: ...
Resources: ...
Criteria: ...
Risks: ...
"""
response = await context["groq_api"].predict(prompt)
return self._parse_task_analysis(response["answer"])
async def _select_strategies(self, task_analysis: Dict[str, Any], context: Dict[str, Any]) -> List[StrategyType]:
"""Select appropriate strategies based on task analysis."""
# Calculate strategy scores
scores: Dict[StrategyType, float] = {}
for strategy_type in StrategyType:
base_score = self.strategy_weights[strategy_type]
# Task type performance
task_type = task_analysis["type"]
type_score = self.task_type_performance[task_type][strategy_type]
# Recent performance
recent_performance = (
sum(self.strategy_performance[strategy_type][-5:]) / 5
if self.strategy_performance[strategy_type] else 0.5
)
# Resource match
resource_match = self._calculate_resource_match(
strategy_type, task_analysis["resources"])
# Capability match
capability_match = self._calculate_capability_match(
strategy_type, task_analysis["capabilities"])
# Combined score
scores[strategy_type] = (
0.3 * base_score +
0.2 * type_score +
0.2 * recent_performance +
0.15 * resource_match +
0.15 * capability_match
)
# Select top strategies
selected = sorted(
StrategyType,
key=lambda x: scores[x],
reverse=True
)[:self.parallel_threshold]
return selected
async def _execute_strategies(self,
strategies: List[StrategyType],
query: str,
context: Dict[str, Any]) -> Dict[StrategyType, StrategyResult]:
"""Execute selected strategies in parallel."""
async def execute_strategy(strategy_type: StrategyType) -> StrategyResult:
strategy = self.strategies[strategy_type]
start_time = datetime.now()
try:
result = await strategy.reason(query, context)
return StrategyResult(
strategy_type=strategy_type,
success=result.get("success", False),
answer=result.get("answer"),
confidence=result.get("confidence", 0.0),
reasoning_trace=result.get("reasoning_trace", []),
metadata=result.get("metadata", {}),
performance_metrics={
"execution_time": (datetime.now() - start_time).total_seconds(),
**result.get("performance_metrics", {})
}
)
except Exception as e:
logging.error(f"Error in strategy {strategy_type}: {str(e)}")
return StrategyResult(
strategy_type=strategy_type,
success=False,
answer=None,
confidence=0.0,
reasoning_trace=[{"error": str(e)}],
metadata={},
performance_metrics={"execution_time": (datetime.now() - start_time).total_seconds()}
)
# Execute strategies in parallel
tasks = [execute_strategy(strategy) for strategy in strategies]
results = await asyncio.gather(*tasks)
return {result.strategy_type: result for result in results}
async def _synthesize_results(self,
strategy_results: Dict[StrategyType, StrategyResult],
task_analysis: Dict[str, Any],
context: Dict[str, Any]) -> UnifiedResult:
"""Synthesize results from multiple strategies."""
prompt = f"""
Synthesize reasoning results:
Results: {json.dumps({str(k): self._strategy_result_to_dict(v)
for k, v in strategy_results.items()})}
Task Analysis: {json.dumps(task_analysis)}
Context: {json.dumps(context)}
Provide:
1. Optimal synthesis method
2. Combined answer
3. Confidence assessment
4. Meta-insights
5. Performance analysis
Format as:
[Synthesis]
Method: ...
Answer: ...
Confidence: ...
Insights: ...
Performance: ...
"""
response = await context["groq_api"].predict(prompt)
synthesis = self._parse_synthesis(response["answer"])
return UnifiedResult(
success=synthesis["confidence"] >= self.min_confidence,
answer=synthesis["answer"],
confidence=synthesis["confidence"],
strategy_results=strategy_results,
synthesis_method=synthesis["method"],
meta_insights=synthesis["insights"],
performance_metrics=synthesis["performance"]
)
def _update_performance(self, result: UnifiedResult):
"""Update performance metrics and strategy weights."""
# Update strategy performance
for strategy_type, strategy_result in result.strategy_results.items():
self.strategy_performance[strategy_type].append(strategy_result.confidence)
# Update weights using exponential moving average
current_weight = self.strategy_weights[strategy_type]
performance = strategy_result.confidence
self.strategy_weights[strategy_type] = (
(1 - self.learning_rate) * current_weight +
self.learning_rate * performance
)
# Update synthesis performance
self.synthesis_performance[result.synthesis_method].append(result.confidence)
def _calculate_resource_match(self, strategy_type: StrategyType, required_resources: Dict[str, Any]) -> float:
"""Calculate how well a strategy matches required resources."""
# Implementation-specific resource matching logic
return 0.8 # Placeholder
def _calculate_capability_match(self, strategy_type: StrategyType, required_capabilities: List[str]) -> float:
"""Calculate how well a strategy matches required capabilities."""
# Implementation-specific capability matching logic
return 0.8 # Placeholder
def _parse_task_analysis(self, response: str) -> Dict[str, Any]:
"""Parse task analysis from response."""
analysis = {
"type": "",
"complexity": 0.0,
"capabilities": [],
"resources": {},
"criteria": [],
"risks": []
}
for line in response.split('\n'):
line = line.strip()
if line.startswith('Type:'):
analysis["type"] = line[5:].strip()
elif line.startswith('Complexity:'):
try:
analysis["complexity"] = float(line[11:].strip())
except:
pass
elif line.startswith('Capabilities:'):
analysis["capabilities"] = [c.strip() for c in line[13:].split(',')]
elif line.startswith('Resources:'):
try:
analysis["resources"] = json.loads(line[10:].strip())
except:
analysis["resources"] = {"raw": line[10:].strip()}
elif line.startswith('Criteria:'):
analysis["criteria"] = [c.strip() for c in line[9:].split(',')]
elif line.startswith('Risks:'):
analysis["risks"] = [r.strip() for r in line[7:].split(',')]
return analysis
def _parse_synthesis(self, response: str) -> Dict[str, Any]:
"""Parse synthesis result from response."""
synthesis = {
"method": "",
"answer": "",
"confidence": 0.0,
"insights": [],
"performance": {}
}
for line in response.split('\n'):
line = line.strip()
if line.startswith('Method:'):
synthesis["method"] = line[7:].strip()
elif line.startswith('Answer:'):
synthesis["answer"] = line[7:].strip()
elif line.startswith('Confidence:'):
try:
synthesis["confidence"] = float(line[11:].strip())
except:
pass
elif line.startswith('Insights:'):
synthesis["insights"] = [i.strip() for i in line[9:].split(',')]
elif line.startswith('Performance:'):
try:
synthesis["performance"] = json.loads(line[12:].strip())
except:
synthesis["performance"] = {"raw": line[12:].strip()}
return synthesis
def _strategy_result_to_dict(self, result: StrategyResult) -> Dict[str, Any]:
"""Convert strategy result to dictionary for serialization."""
return {
"strategy_type": result.strategy_type.value,
"success": result.success,
"answer": result.answer,
"confidence": result.confidence,
"reasoning_trace": result.reasoning_trace,
"metadata": result.metadata,
"performance_metrics": result.performance_metrics,
"timestamp": result.timestamp.isoformat()
}
def get_performance_metrics(self) -> Dict[str, Any]:
"""Get comprehensive performance metrics."""
return {
"strategy_weights": dict(self.strategy_weights),
"average_performance": {
strategy_type.value: sum(scores) / len(scores) if scores else 0
for strategy_type, scores in self.strategy_performance.items()
},
"synthesis_success": {
method: sum(scores) / len(scores) if scores else 0
for method, scores in self.synthesis_performance.items()
},
"task_type_performance": {
task_type: dict(strategy_scores)
for task_type, strategy_scores in self.task_type_performance.items()
}
}
def clear_performance_history(self):
"""Clear performance history and reset weights."""
self.strategy_performance.clear()
self.task_type_performance.clear()
self.synthesis_performance.clear()
self.strategy_weights = {
strategy_type: 1.0 for strategy_type in StrategyType
}
|