File size: 12,876 Bytes
9abace2 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 | """
Layer 1: Protocol Definitions
Part of SOVEREIGN PYTHON LLM ENGINE
Typed protocols for all major system components.
Protocols define contracts without implementation.
"""
from typing import Protocol, AsyncIterator, Any, runtime_checkable
import numpy as np
# ==========================================
# Retrieval Protocols
# ==========================================
@runtime_checkable
class Retriever(Protocol):
"""
Protocol for all retrieval sources (RAG, search, database, etc.)
Implementations:
- WikipediaRetriever
- GitHubRetriever
- VectorStoreRetriever
- SQLRetriever
"""
async def retrieve(self, query: str) -> str:
"""
Retrieve relevant content for query.
Args:
query: User query string
Returns:
Retrieved content (may be concatenated from multiple sources)
Raises:
RetrievalError: If retrieval fails
"""
...
@runtime_checkable
class BatchRetriever(Protocol):
"""Retriever that supports batch queries"""
async def retrieve_batch(self, queries: list[str]) -> list[str]:
"""Retrieve content for multiple queries concurrently"""
...
# ==========================================
# Tool Execution Protocols
# ==========================================
@runtime_checkable
class Tool(Protocol):
"""
Protocol for executable tools (code execution, API calls, etc.)
All tools must:
- Accept structured parameters (dict)
- Return structured results (dict)
- Be async
- Handle errors gracefully
"""
name: str
description: str
parameters_schema: dict[str, Any] # JSON schema
async def execute(self, params: dict[str, Any]) -> dict[str, Any]:
"""
Execute tool with given parameters.
Args:
params: Tool parameters (validated against schema)
Returns:
Tool execution results
Raises:
ToolExecutionError: If execution fails
"""
...
@runtime_checkable
class SandboxedTool(Protocol):
"""Tool that runs in isolated sandbox (e.g., code execution)"""
timeout: float # Execution timeout in seconds
async def execute_sandboxed(
self,
params: dict[str, Any]
) -> dict[str, Any]:
"""Execute in isolated environment"""
...
# ==========================================
# Model Inference Protocols
# ==========================================
@runtime_checkable
class Model(Protocol):
"""
Protocol for LLM inference backends.
Implementations:
- LlamaAPIBackend
- OpenAIBackend
- AnthropicBackend
- LocalTransformerBackend
"""
model_id: str
async def generate(
self,
messages: list[dict[str, str]],
temperature: float = 0.0,
max_tokens: int | None = None,
stream: bool = False
) -> str | AsyncIterator[str]:
"""
Generate completion from messages.
Args:
messages: List of {role, content} dicts
temperature: Sampling temperature [0.0, 2.0]
max_tokens: Max tokens to generate (None = model default)
stream: If True, return AsyncIterator of chunks
Returns:
Complete response string, or AsyncIterator of chunks if stream=True
Raises:
ModelError: If generation fails
"""
...
@runtime_checkable
class StructuredOutputModel(Protocol):
"""Model that supports structured output (JSON schema enforcement)"""
async def generate_structured(
self,
messages: list[dict[str, str]],
response_schema: dict[str, Any], # JSON schema
temperature: float = 0.0
) -> dict[str, Any]:
"""
Generate structured output matching schema.
Args:
messages: Conversation history
response_schema: JSON schema to enforce
temperature: Sampling temperature
Returns:
Validated structured output
Raises:
SchemaValidationError: If output doesn't match schema
"""
...
@runtime_checkable
class ToolCallingModel(Protocol):
"""Model that supports native tool calling"""
async def generate_with_tools(
self,
messages: list[dict[str, str]],
tools: list[dict[str, Any]], # Tool definitions
temperature: float = 0.0
) -> dict[str, Any]:
"""
Generate with tool calling support.
Returns:
{
"content": str,
"tool_calls": [{"name": str, "arguments": dict}]
}
"""
...
# ==========================================
# Storage Protocols
# ==========================================
@runtime_checkable
class KeyValueStore(Protocol):
"""Key-value storage interface"""
async def get(self, key: str) -> bytes | None:
"""Get value for key, None if not found"""
...
async def put(self, key: str, value: bytes) -> None:
"""Store key-value pair"""
...
async def delete(self, key: str) -> None:
"""Delete key"""
...
async def exists(self, key: str) -> bool:
"""Check if key exists"""
...
@runtime_checkable
class VectorStore(Protocol):
"""Vector database interface for embeddings"""
dimension: int # Embedding dimension
async def add(
self,
vectors: np.ndarray, # [n, dimension]
metadata: list[dict[str, Any]]
) -> list[str]:
"""
Add vectors with metadata.
Returns:
List of assigned IDs
"""
...
async def search(
self,
query_vector: np.ndarray, # [dimension]
k: int = 5
) -> list[dict[str, Any]]:
"""
Search for k nearest neighbors.
Returns:
List of {id, distance, metadata} dicts
"""
...
@runtime_checkable
class TransactionalStore(Protocol):
"""Database with transaction support"""
async def begin_transaction(self) -> Any:
"""Begin transaction, return transaction handle"""
...
async def commit(self, txn: Any) -> None:
"""Commit transaction"""
...
async def rollback(self, txn: Any) -> None:
"""Rollback transaction"""
...
# ==========================================
# Agent Protocols
# ==========================================
@runtime_checkable
class Agent(Protocol):
"""
Protocol for autonomous agents.
Implementations:
- ReActAgent
- MCTSAgent
- ReasoningAgent
"""
agent_id: str
max_steps: int
async def run(self, task: str) -> str:
"""
Execute agent on task.
Args:
task: Task description
Returns:
Final answer/result
Raises:
AgentError: If execution fails
MaxStepsExceeded: If max_steps reached without answer
"""
...
@runtime_checkable
class ReflectiveAgent(Protocol):
"""Agent with self-reflection capability"""
async def run_with_reflection(
self,
task: str,
reflection_trigger: str = "ERROR"
) -> dict[str, Any]:
"""
Run with reflection on errors.
Returns:
{
"answer": str,
"reflections": list[str],
"steps": int
}
"""
...
# ==========================================
# Router Protocols
# ==========================================
@runtime_checkable
class Router(Protocol):
"""
Protocol for routing/dispatching queries.
Implementations:
- LLMRouter (LLM-based routing)
- RuleRouter (rule-based routing)
- HybridRouter (combination)
"""
async def route(self, query: str) -> str:
"""
Route query to appropriate destination.
Args:
query: User query
Returns:
Destination identifier (e.g., "vector_db", "sql_database")
"""
...
@runtime_checkable
class WeightedRouter(Protocol):
"""Router that returns routing weights (for ensemble)"""
async def route_weighted(
self,
query: str
) -> dict[str, float]:
"""
Route with weights for each destination.
Returns:
{destination: weight} where sum(weights) = 1.0
"""
...
# ==========================================
# MoE Expert Protocols
# ==========================================
@runtime_checkable
class Expert(Protocol):
"""
Protocol for MoE experts.
Each expert is a feed-forward network (typically SwiGLU).
"""
expert_id: int
hidden_dim: int
intermediate_dim: int
def forward(self, x: np.ndarray) -> np.ndarray:
"""
Forward pass through expert.
Args:
x: Input hidden state [hidden_dim]
Returns:
Output hidden state [hidden_dim]
"""
...
@runtime_checkable
class QuantumExpert(Protocol):
"""Expert with quantum token handling"""
def forward_quantum(
self,
x: np.ndarray,
quantum_state: Any # QuantumState from quantum_moe.py
) -> np.ndarray:
"""Forward pass with quantum token encoding"""
...
# ==========================================
# Gating Network Protocols
# ==========================================
@runtime_checkable
class GatingNetwork(Protocol):
"""
Protocol for MoE gating/routing.
Implementations:
- Top-K Gating
- Top-K with noise
- Learned routing
"""
num_experts: int
top_k: int
def gate(self, x: np.ndarray) -> tuple[np.ndarray, np.ndarray]:
"""
Compute gating weights.
Args:
x: Input hidden state
Returns:
(expert_indices, expert_weights)
- expert_indices: [top_k] indices of selected experts
- expert_weights: [top_k] routing weights
"""
...
# ==========================================
# Scanner Protocols
# ==========================================
@runtime_checkable
class CodeScanner(Protocol):
"""Protocol for code analysis/scanning"""
async def scan_file(self, file_path: str) -> dict[str, Any]:
"""
Scan single file.
Returns:
{
"classes": list[str],
"functions": list[str],
"imports": list[str],
...
}
"""
...
async def scan_directory(self, root: str) -> dict[str, Any]:
"""Scan entire directory recursively"""
...
@runtime_checkable
class DependencyAnalyzer(Protocol):
"""Analyze code dependencies"""
async def build_graph(self, root: str) -> dict[str, Any]:
"""
Build dependency graph.
Returns:
{
"nodes": list[str], # File paths
"edges": list[tuple[str, str]], # (source, target)
"forward": dict, # file -> dependencies
"reverse": dict # file -> dependents
}
"""
...
# ==========================================
# Ledger/Evidence Protocols
# ==========================================
@runtime_checkable
class EvidenceLedger(Protocol):
"""Protocol for append-only evidence logging"""
async def append(
self,
event_type: str,
data: bytes,
metadata: dict[str, Any]
) -> dict[str, Any]:
"""
Append evidence record.
Returns:
Record metadata (timestamp, hash, signature)
"""
...
async def verify_chain(self) -> bool:
"""Verify cryptographic chain integrity"""
...
# ==========================================
# Type Checking Helpers
# ==========================================
def is_retriever(obj: Any) -> bool:
"""Check if object implements Retriever protocol"""
return isinstance(obj, Retriever)
def is_model(obj: Any) -> bool:
"""Check if object implements Model protocol"""
return isinstance(obj, Model)
def is_agent(obj: Any) -> bool:
"""Check if object implements Agent protocol"""
return isinstance(obj, Agent)
|