Initial commit: LocalPilot：本地模型运行时，复用 llama-server 并提供 Ollama 兼容 provider

localpilot/types.py

@@ -0,0 +1,50 @@
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Any
+
+
+@dataclass(slots=True)
+class ChatParams:
+    max_tokens: int = 512
+    temperature: float = 0.7
+    top_p: float = 0.95
+    top_k: int = 40
+    min_p: float = 0.05
+    repeat_penalty: float = 1.05
+    seed: int | None = None
+    stop: list[str] = field(default_factory=list)
+    enable_thinking: bool | None = None
+    extra: dict[str, Any] = field(default_factory=dict)
+    response_format: str | dict[str, Any] | None = None
+    tools: list[dict[str, Any]] = field(default_factory=list)
+    tool_choice: Any = None
+    keep_alive: str | int | float | None = None
+
+    def as_generation_kwargs(self) -> dict[str, Any]:
+        result: dict[str, Any] = {
+            "max_tokens": self.max_tokens,
+            "temperature": self.temperature,
+            "top_p": self.top_p,
+            "top_k": self.top_k,
+            "min_p": self.min_p,
+            "repeat_penalty": self.repeat_penalty,
+        }
+        if self.seed is not None:
+            result["seed"] = self.seed
+        if self.stop:
+            result["stop"] = self.stop
+        return result
+
+
+@dataclass(slots=True)
+class ChatChunk:
+    text: str = ""
+    done: bool = False
+    finish_reason: str | None = None
+    metadata: dict[str, Any] = field(default_factory=dict)
+    delta: dict[str, Any] = field(default_factory=dict)
+
+
+class BackendError(RuntimeError):
+    """An expected model backend failure with a user-facing message."""