feat : BaseToolParser.feed 增加可选 token_ids 参数

- format_chunk ABC 改为 (token, **kwargs)，body/token_ids 通过 kw 传入
- ProtocolHandler._handle_stream 逐 token encode 并透传
- Anthropic builder 用 **kwargs 吸收不使用的参数，零变更
- 新增 3 个 token_ids 参数测试

astrai/inference/api/anthropic.py

@@ -73,7 +73,7 @@ class AnthropicResponseBuilder(ResponseBuilder):
             ),
         ]
 
-    def format_chunk(self, token: str, body: str) -> List[str]:
+    def format_chunk(self, token: str, **kwargs) -> List[str]:
         return [
             sse_event(
                 {

astrai/inference/api/openai.py

@@ -113,9 +113,10 @@ class OpenAIResponseBuilder(ResponseBuilder):
             )
         ]
 
-    def format_chunk(self, token: str, body: str) -> List[str]:
+    def format_chunk(self, token: str, **kwargs) -> List[str]:
+        body = kwargs.get("body", "")
         if self._parser is not None:
-            return self._format_tool_chunk(body)
+            return self._format_tool_chunk(body, **kwargs)
 
         return [
             sse_event(
@@ -135,8 +136,12 @@ class OpenAIResponseBuilder(ResponseBuilder):
             )
         ]
 
-    def _format_tool_chunk(self, body: str) -> List[str]:
-        deltas = self._parser.feed(body)
+    def _format_tool_chunk(self, body: str, **kwargs) -> List[str]:
+        deltas = self._parser.feed(
+            body,
+            current_token_ids=kwargs.get("current_token_ids"),
+            delta_token_ids=kwargs.get("delta_token_ids"),
+        )
         events: List[str] = []
         for d in deltas:
             if "content" in d:

astrai/inference/api/protocol.py

@@ -78,11 +78,13 @@ class ResponseBuilder(ABC):
         """SSE events that open the stream."""
 
     @abstractmethod
-    def format_chunk(self, token: str, body: str) -> List[str]:
+    def format_chunk(self, token: str, **kwargs) -> List[str]:
         """SSE events for a single generated token.
 
-        Receives the current token and the full accumulated *body* so
-        that tool-call parsers can re-parse the complete text each step.
+        ``body`` (the full accumulated text so far) is always provided
+        as a keyword argument. Additional keyword arguments such as
+        ``current_token_ids`` and ``delta_token_ids`` may be included
+        for tool parsers that need token-level information.
         Returns a list of SSE event strings (may be empty).
         """
 
@@ -142,15 +144,24 @@ class ProtocolHandler:
             body = ""
             yielded = ""
             matched = None
+            token_ids: List[int] = []
             async for token in agen:
                 body += token
 
+                new_ids = self.engine.tokenizer.encode(token)
+                token_ids.extend(new_ids)
+
                 matched = checker.check(body)
                 if matched:
                     break
 
                 ctx.completion_tokens += 1
-                for event in self.builder.format_chunk(token, body):
+                for event in self.builder.format_chunk(
+                    token,
+                    body=body,
+                    current_token_ids=token_ids,
+                    delta_token_ids=new_ids,
+                ):
                     yield event
                 yielded += token
 

astrai/inference/api/tool_parser.py

@@ -2,6 +2,9 @@
 
 Patterned after vLLM's ToolParser abstraction. Each parser knows how to
 detect and incrementally extract tool calls from raw generated text.
+
+Subclasses may optionally consume ``token_ids`` for token-level parsing
+(e.g. Harmony / VLM-style parsers).
 """
 
 import re
@@ -17,6 +20,14 @@ class BaseToolParser(ABC):
 
     Maintains streaming state internally so that each call to :meth:`feed`
     can diff against previously emitted content.
+
+    Parameters
+    ----------
+    tools : list of dict, optional
+        Tool definitions from the request.
+    tool_choice : str
+        ``"auto"`` / ``"required"`` / ``"none"`` or a named tool choice
+        dict.
     """
 
     def __init__(self, tools: Optional[List[Dict]] = None, tool_choice: str = "auto"):
@@ -24,7 +35,12 @@ class BaseToolParser(ABC):
         self.tool_choice = tool_choice
 
     @abstractmethod
-    def feed(self, body: str) -> List[Dict]:
+    def feed(
+        self,
+        body: str,
+        current_token_ids: Optional[List[int]] = None,
+        delta_token_ids: Optional[List[int]] = None,
+    ) -> List[Dict]:
         """Feed the *full* accumulated text each step.
 
         Returns a list of delta dicts to emit. Each delta is one of:
@@ -33,6 +49,15 @@ class BaseToolParser(ABC):
         - ``{"tool_calls": [...]}``    — tool-call delta (OpenAI format)
 
         Returns an empty list when nothing new should be emitted.
+
+        Parameters
+        ----------
+        body : str
+            The complete accumulated generated text so far.
+        current_token_ids : list of int, optional
+            All token IDs decoded into *body* (cumulative).
+        delta_token_ids : list of int, optional
+            Only the token IDs for this chunk.
         """
 
     @abstractmethod
@@ -199,7 +224,12 @@ class SimpleJsonToolParser(BaseToolParser):
 
     # -------------------------------------------------------------- feed
 
-    def feed(self, body: str) -> List[Dict]:
+    def feed(
+        self,
+        body: str,
+        current_token_ids: Optional[List[int]] = None,
+        delta_token_ids: Optional[List[int]] = None,
+    ) -> List[Dict]:
         deltas: List[Dict] = []
 
         completed = _find_tool_calls(body)