feat: raise configurable chat tool call limit

backend/wisegrade/chat_api.py

@@ -10,6 +10,7 @@ from sqlalchemy.orm import Session, sessionmaker
 
 from .actions import action_dict, confirm, prepare, require_scope
 from .agent_tools import execute, tool_specs
+from .config import get_settings
 from .db import get_db
 from .llm import ModelError
 from .models import AgentRun, ChatMessage, Conversation, PendingAction, Project, ScoringBatch, now
@@ -113,7 +114,10 @@ def _run_agent(factory, conversation_id, run_id, user_text):
                 "project_id": conversation.project_id}, ensure_ascii=False)}]
             messages.extend({"role": m.role, "content": m.content[:4000]} for m in history)
             yield _event("run", {"id": run.id, "status": "running"})
-            for round_number in range(4):
+            max_tool_calls = get_settings().chat_max_tool_calls
+            attempted_calls = 0
+            # One extra model response lets the agent answer after its last allowed tool call.
+            for _ in range(max_tool_calls + 1):
                 reply = client.tool_call(messages, tool_specs())
                 calls = reply.get("tool_calls") or []
                 if not calls:
@@ -121,16 +125,17 @@ def _run_agent(factory, conversation_id, run_id, user_text):
                     if not isinstance(content, str) or not content.strip():
                         raise ModelError("对话模型未返回文字或工具调用")
                     db.add(ChatMessage(conversation_id=conversation_id, role="assistant", content=content,
-                                       citations=citations[:20], tool_calls=used[:8]))
+                                       citations=citations[:20], tool_calls=used))
                     run.status, run.finished_at = "completed", now()
                     conversation.updated_at = now()
                     db.commit()
                     completed = True
-                    yield _event("message", {"content": content, "citations": citations[:20], "tool_calls": used[:8]})
+                    yield _event("message", {"content": content, "citations": citations[:20], "tool_calls": used})
                     yield _event("done", {"run_id": run.id})
                     return
-                if len(calls) + len(used) > 8:
-                    raise ModelError("单轮工具调用超过上限 8 次")
+                if attempted_calls + len(calls) > max_tool_calls:
+                    raise ModelError(f"本次消息工具调用超过上限 {max_tool_calls} 次")
+                attempted_calls += len(calls)
                 messages.append({"role": "assistant", "content": reply.get("content"), "tool_calls": calls})
                 for call in calls:
                     name = call.get("function", {}).get("name", "")
@@ -155,7 +160,7 @@ def _run_agent(factory, conversation_id, run_id, user_text):
                         messages.append({"role": "tool", "tool_call_id": call_id,
                                          "content": json.dumps({"error": error[:500]}, ensure_ascii=False)})
                         yield _event("tool", {"name": name, "status": "failed", "error": error[:500]})
-            raise ModelError("模型工具调用轮数超过上限 4 次")
+            raise ModelError(f"本次消息模型调用超过上限 {max_tool_calls + 1} 轮")
         except Exception as exc:
             run.status, run.error, run.finished_at = "failed", str(exc)[:500], now()
             db.commit()

backend/wisegrade/config.py

@@ -1,6 +1,7 @@
 from functools import lru_cache
 from pathlib import Path
 
+from pydantic import Field
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
 
@@ -23,6 +24,7 @@ class Settings(BaseSettings):
     model_timeout_seconds: int = 120
     model_concurrency: int = 2
     model_retries: int = 2
+    chat_max_tool_calls: int = Field(default=32, ge=1, le=64)
 
 
 @lru_cache