Spaces:
Sleeping
Sleeping
File size: 11,342 Bytes
f088210 | 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 | """Enhanced RAG-Powered Analytics Agent with Community Insights.
Integrates retrieval-augmented generation for context-aware analytics,
versioning hooks per SCHEMA_VERSIONING.md, and rollback capabilities.
Enhanced with patterns from bziiit/AGENT_ANALYSE_RAG and fahmiaziz/rag-sql-agent.
"""
import os
import json
import logging
from datetime import datetime
from typing import Dict, List, Optional, Any, Tuple
from dataclasses import dataclass, asdict
import sys
# Import RAG pipeline for context retrieval
from smolagents.rag_pipeline import RAGPipeline, RetrievalConfig
from schema_registry import SchemaRegistry
logger = logging.getLogger(__name__)
@dataclass
class AnalyticsConfig:
"""Configuration for RAG-powered analytics."""
rag_enabled: bool = True
top_k_docs: int = 5
similarity_threshold: float = 0.7
enable_versioning: bool = True
schema_version: str = "1.0.0"
rollback_enabled: bool = True
audit_trail: bool = True
@dataclass
class AnalyticsQuery:
"""Structured analytics query with RAG context."""
query_id: str
query_text: str
context_docs: List[Dict[str, Any]]
filters: Dict[str, Any]
aggregations: List[str]
timestamp: str
schema_version: str
@dataclass
class AnalyticsResult:
"""Enhanced analytics result with citations and provenance."""
query_id: str
data: Dict[str, Any]
citations: List[Dict[str, Any]]
confidence_score: float
schema_version: str
rollback_checkpoint: Optional[str]
timestamp: str
class RAGAnalyticsAgent:
"""Enhanced Analytics Agent with RAG, versioning, and rollback."""
def __init__(
self,
config: Optional[AnalyticsConfig] = None,
rag_pipeline: Optional[RAGPipeline] = None,
schema_registry: Optional[SchemaRegistry] = None
):
self.config = config or AnalyticsConfig()
self.rag_pipeline = rag_pipeline
self.schema_registry = schema_registry
self.checkpoints: Dict[str, Dict] = {}
logger.info(f"Initialized RAGAnalyticsAgent with config: {self.config}")
def create_checkpoint(self, checkpoint_id: str, state: Dict[str, Any]) -> None:
"""Create rollback checkpoint per SCHEMA_VERSIONING.md."""
self.checkpoints[checkpoint_id] = {
"state": state,
"timestamp": datetime.utcnow().isoformat(),
"schema_version": self.config.schema_version
}
logger.info(f"Created checkpoint: {checkpoint_id}")
def rollback_to_checkpoint(self, checkpoint_id: str) -> Optional[Dict[str, Any]]:
"""Rollback to a previous checkpoint."""
if checkpoint_id in self.checkpoints:
checkpoint = self.checkpoints[checkpoint_id]
logger.info(f"Rolling back to checkpoint: {checkpoint_id}")
return checkpoint["state"]
logger.warning(f"Checkpoint not found: {checkpoint_id}")
return None
def retrieve_context(
self,
query_text: str,
top_k: Optional[int] = None
) -> List[Dict[str, Any]]:
"""Retrieve relevant context using RAG pipeline."""
if not self.config.rag_enabled or not self.rag_pipeline:
return []
top_k = top_k or self.config.top_k_docs
try:
results = self.rag_pipeline.retrieve(
query=query_text,
top_k=top_k,
threshold=self.config.similarity_threshold
)
context_docs = [
{
"content": r.get("content", ""),
"source": r.get("metadata", {}).get("source", "unknown"),
"score": r.get("score", 0.0),
"metadata": r.get("metadata", {})
}
for r in results
]
logger.info(f"Retrieved {len(context_docs)} context documents")
return context_docs
except Exception as e:
logger.error(f"Context retrieval failed: {e}")
return []
def validate_schema(
self,
data: Dict[str, Any],
schema_version: Optional[str] = None
) -> Tuple[bool, List[str]]:
"""Validate data against schema version."""
if not self.config.enable_versioning or not self.schema_registry:
return True, []
schema_version = schema_version or self.config.schema_version
try:
is_valid = self.schema_registry.validate(data, schema_version)
errors = [] if is_valid else ["Schema validation failed"]
return is_valid, errors
except Exception as e:
logger.error(f"Schema validation error: {e}")
return False, [str(e)]
def analyze(
self,
query_text: str,
data_source: Dict[str, Any],
filters: Optional[Dict[str, Any]] = None,
aggregations: Optional[List[str]] = None
) -> AnalyticsResult:
"""Perform RAG-enhanced analytics with versioning and rollback."""
query_id = f"analytics_{datetime.utcnow().timestamp()}"
# Create checkpoint before analysis
if self.config.rollback_enabled:
self.create_checkpoint(
f"{query_id}_pre",
{"data_source": data_source, "filters": filters}
)
# Retrieve context
context_docs = self.retrieve_context(query_text)
# Build query
query = AnalyticsQuery(
query_id=query_id,
query_text=query_text,
context_docs=context_docs,
filters=filters or {},
aggregations=aggregations or [],
timestamp=datetime.utcnow().isoformat(),
schema_version=self.config.schema_version
)
# Perform analysis (simplified - extend with actual analytics logic)
try:
analysis_data = self._execute_analysis(query, data_source)
# Validate result schema
is_valid, errors = self.validate_schema(analysis_data)
if not is_valid:
logger.error(f"Schema validation failed: {errors}")
# Attempt rollback
if self.config.rollback_enabled:
rollback_state = self.rollback_to_checkpoint(f"{query_id}_pre")
if rollback_state:
logger.info("Successfully rolled back to pre-analysis state")
# Build citations from context
citations = [
{
"source": doc["source"],
"relevance_score": doc["score"],
"excerpt": doc["content"][:200]
}
for doc in context_docs
]
# Calculate confidence based on context quality
avg_score = sum(d["score"] for d in context_docs) / len(context_docs) if context_docs else 0.5
result = AnalyticsResult(
query_id=query_id,
data=analysis_data,
citations=citations,
confidence_score=avg_score,
schema_version=self.config.schema_version,
rollback_checkpoint=f"{query_id}_pre" if self.config.rollback_enabled else None,
timestamp=datetime.utcnow().isoformat()
)
# Audit trail
if self.config.audit_trail:
self._log_audit(query, result)
return result
except Exception as e:
logger.error(f"Analysis failed: {e}")
# Return error result with rollback info
return AnalyticsResult(
query_id=query_id,
data={"error": str(e)},
citations=[],
confidence_score=0.0,
schema_version=self.config.schema_version,
rollback_checkpoint=f"{query_id}_pre" if self.config.rollback_enabled else None,
timestamp=datetime.utcnow().isoformat()
)
def _execute_analysis(
self,
query: AnalyticsQuery,
data_source: Dict[str, Any]
) -> Dict[str, Any]:
"""Execute the actual analysis logic (extend as needed)."""
# Simplified analytics - enhance with real logic
results = {
"query_id": query.query_id,
"metrics": {},
"insights": [],
"context_applied": len(query.context_docs) > 0
}
# Apply filters
filtered_data = data_source
for key, value in query.filters.items():
if key in filtered_data:
filtered_data = {k: v for k, v in filtered_data.items() if v == value}
# Apply aggregations
for agg in query.aggregations:
if agg == "count":
results["metrics"]["count"] = len(filtered_data)
elif agg == "sum":
results["metrics"]["sum"] = sum(v for v in filtered_data.values() if isinstance(v, (int, float)))
# Extract insights from RAG context
if query.context_docs:
for doc in query.context_docs[:3]: # Top 3 docs
results["insights"].append({
"source": doc["source"],
"insight": doc["content"][:150]
})
return results
def _log_audit(
self,
query: AnalyticsQuery,
result: AnalyticsResult
) -> None:
"""Log audit trail for compliance."""
audit_entry = {
"timestamp": datetime.utcnow().isoformat(),
"query_id": query.query_id,
"query_text": query.query_text,
"schema_version": query.schema_version,
"confidence_score": result.confidence_score,
"citations_count": len(result.citations)
}
logger.info(f"Audit: {json.dumps(audit_entry)}")
def get_version_info(self) -> Dict[str, Any]:
"""Get current version and rollback information."""
return {
"schema_version": self.config.schema_version,
"checkpoints": list(self.checkpoints.keys()),
"versioning_enabled": self.config.enable_versioning,
"rollback_enabled": self.config.rollback_enabled
}
# Example usage
if __name__ == "__main__":
logging.basicConfig(level=logging.INFO)
# Initialize agent
config = AnalyticsConfig(
rag_enabled=True,
top_k_docs=5,
enable_versioning=True,
rollback_enabled=True
)
agent = RAGAnalyticsAgent(config=config)
# Example query
result = agent.analyze(
query_text="What are the top engagement metrics for Q4?",
data_source={"engagement": 100, "reach": 500},
aggregations=["count", "sum"]
)
print(f"\nAnalysis Result:")
print(f"Query ID: {result.query_id}")
print(f"Confidence: {result.confidence_score}")
print(f"Citations: {len(result.citations)}")
print(f"Rollback checkpoint: {result.rollback_checkpoint}")
print(f"\nVersion Info: {json.dumps(agent.get_version_info(), indent=2)}")
|