Spaces:
Running
Running
File size: 8,017 Bytes
09801ca | 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 | """
🤖 Agentic AutoML - Shared Memory Manager
Implements the shared state architecture:
- Immutable storage per stage
- Version tracking
- Artifact management
- State history for debugging
"""
from dataclasses import dataclass, field
from typing import Any, Dict, List, Optional, Tuple
from datetime import datetime
from collections import defaultdict
import copy
import json
import logging
logger = logging.getLogger(__name__)
@dataclass
class StateEntry:
"""A single entry in the shared state"""
key: str
value: Any
stage: str
version: int
timestamp: datetime = field(default_factory=datetime.now)
def to_dict(self) -> Dict:
return {
"key": self.key,
"stage": self.stage,
"version": self.version,
"timestamp": self.timestamp.isoformat(),
"value_type": type(self.value).__name__
}
@dataclass
class Artifact:
"""An artifact produced by an agent (model, chart, report)"""
id: str
type: str # model, chart, report, features, etc.
producer: str # Agent that created it
data: Any
metadata: Dict[str, Any] = field(default_factory=dict)
timestamp: datetime = field(default_factory=datetime.now)
class AgentMemory:
"""
Centralized shared memory for all agents.
Features:
- Immutable state per stage (each write creates new version)
- Full history for debugging and replay
- Artifact storage for models, charts, etc.
- Thread-safe operations (TODO: add locks for production)
"""
def __init__(self):
# Current state (latest version of each key)
self._state: Dict[str, Any] = {}
# State history (key -> list of StateEntry)
self._history: Dict[str, List[StateEntry]] = defaultdict(list)
# Artifacts storage
self._artifacts: Dict[str, Artifact] = {}
# Logs for debugging
self._logs: List[Dict] = []
# Version counters
self._versions: Dict[str, int] = defaultdict(int)
# Pipeline metadata
self.pipeline_id: str = ""
self.created_at: datetime = datetime.now()
# =========================================================================
# STATE MANAGEMENT
# =========================================================================
def get(self, key: str, default: Any = None) -> Any:
"""Get current value for a key"""
return self._state.get(key, default)
def set(self, key: str, value: Any, stage: str):
"""
Set a value (creates new version, doesn't overwrite history)
"""
self._versions[key] += 1
version = self._versions[key]
# Create entry
entry = StateEntry(
key=key,
value=copy.deepcopy(value), # Deep copy to ensure immutability
stage=stage,
version=version
)
# Store in history
self._history[key].append(entry)
# Update current state
self._state[key] = value
logger.debug(f"📝 Memory[{key}] = {type(value).__name__} (v{version} by {stage})")
def get_history(self, key: str) -> List[StateEntry]:
"""Get full history for a key"""
return self._history.get(key, [])
def get_version(self, key: str, version: int) -> Optional[Any]:
"""Get a specific version of a value"""
history = self._history.get(key, [])
for entry in history:
if entry.version == version:
return entry.value
return None
def get_by_stage(self, key: str, stage: str) -> Optional[Any]:
"""Get value as set by a specific stage"""
history = self._history.get(key, [])
for entry in reversed(history): # Latest first
if entry.stage == stage:
return entry.value
return None
# =========================================================================
# ARTIFACT MANAGEMENT
# =========================================================================
def store_artifact(self, artifact_id: str, artifact_type: str,
producer: str, data: Any, metadata: Dict = None):
"""Store an artifact"""
artifact = Artifact(
id=artifact_id,
type=artifact_type,
producer=producer,
data=data,
metadata=metadata or {}
)
self._artifacts[artifact_id] = artifact
logger.info(f"📦 Artifact stored: {artifact_type} by {producer}")
def get_artifact(self, artifact_id: str) -> Optional[Artifact]:
"""Get an artifact by ID"""
return self._artifacts.get(artifact_id)
def get_artifacts(self, artifact_type: str) -> List[Artifact]:
"""Get all artifacts of a specific type"""
return [a for a in self._artifacts.values() if a.type == artifact_type]
def get_latest_artifact(self, artifact_type: str) -> Optional[Artifact]:
"""Get the most recent artifact of a type"""
artifacts = self.get_artifacts(artifact_type)
if artifacts:
return max(artifacts, key=lambda a: a.timestamp)
return None
# =========================================================================
# LOGGING
# =========================================================================
def log(self, agent: str, message: str, level: str = "info", data: Dict = None):
"""Add a log entry"""
entry = {
"timestamp": datetime.now().isoformat(),
"agent": agent,
"level": level,
"message": message,
"data": data or {}
}
self._logs.append(entry)
def get_logs(self, agent: str = None, level: str = None) -> List[Dict]:
"""Get logs, optionally filtered"""
logs = self._logs
if agent:
logs = [l for l in logs if l["agent"] == agent]
if level:
logs = [l for l in logs if l["level"] == level]
return logs
# =========================================================================
# CONVENIENCE METHODS
# =========================================================================
@property
def dataset(self):
"""Get the current dataset"""
return self.get("dataset")
@property
def features(self):
"""Get the current feature matrix"""
return self.get("features")
@property
def target(self):
"""Get the current target variable"""
return self.get("target")
@property
def best_model(self):
"""Get the best model artifact"""
return self.get_latest_artifact("model")
@property
def metrics(self) -> Dict[str, float]:
"""Get the current metrics"""
return self.get("metrics", {})
# =========================================================================
# SERIALIZATION
# =========================================================================
def get_state_summary(self) -> Dict:
"""Get a summary of current state"""
return {
"pipeline_id": self.pipeline_id,
"created_at": self.created_at.isoformat(),
"keys": list(self._state.keys()),
"artifact_count": len(self._artifacts),
"log_count": len(self._logs),
"versions": dict(self._versions)
}
def export_logs(self) -> str:
"""Export logs as JSON string"""
return json.dumps(self._logs, indent=2, default=str)
def clear(self):
"""Clear all state (for new pipeline)"""
self._state.clear()
self._history.clear()
self._artifacts.clear()
self._logs.clear()
self._versions.clear()
self.pipeline_id = ""
self.created_at = datetime.now()
|