Update 2026-03-22 12:00:38

agent/__init__.py

@@ -0,0 +1,21 @@
+
+# ══════════════════════════════════════════════════════════════════════════════
+# FILE: agent/state.py
+# ══════════════════════════════════════════════════════════════════════════════
+
+from typing import TypedDict, Annotated, List, Optional
+from langchain_core.messages import BaseMessage
+import operator
+
+class AgentState(TypedDict):
+    messages: Annotated[List[BaseMessage], operator.add]
+    current_node: str
+    model_name: str
+    session_id: str
+    hf_token: str
+    iteration_count: int
+    should_end: bool
+    final_answer: Optional[str]
+    error: Optional[str]
+    conversation_history: List[dict]
+    pending_tool: Optional[dict]

app.py

@@ -10,8 +10,8 @@ app = Flask(__name__)
 app.secret_key = os.getenv("SECRET_KEY","lgsa-2025-dev-secret")
 
 AVAILABLE_MODELS = [
-    {"id":"meta-llama/Meta-Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","badge":"🦙 Recommended"},
-    {"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen 2.5 7B Instruct","badge":"⚡ Fast"},
+    {"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen 2.5 7B Instruct","badge":"⚡ Recommended"},
+    {"id":"meta-llama/Meta-Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","badge":"🦙 Advanced"},
     {"id":"mistralai/Mistral-7B-Instruct-v0.3","name":"Mistral 7B Instruct v0.3","badge":"🌀 Mistral"},
     {"id":"google/gemma-2-9b-it","name":"Gemma 2 9B Instruct","badge":"💎 Google"},
 ]
@@ -158,5 +158,4 @@ def api_reset():
     return jsonify({"status":"ok","session_id":sid})
 
 if __name__ == "__main__":
-    app.run(host="0.0.0.0", port=7860, debug=False, threaded=True)
-
+    app.run(host="0.0.0.0", port=7860, debug=False, threaded=True)

code.txt

@@ -382,7 +382,7 @@ def agent_node(state: AgentState) -> dict:
             user_msg = msg.content
         elif cname == "ToolMessage":
             tool_obs.append({"tool":getattr(msg,"name","tool"),"result":msg.content})
-    client = InferenceClient(api_key=state["hf_token"])
+    client = InferenceClient(api_key=state["hf_token"], provider="auto")
     messages = build_messages(user_msg, state.get("conversation_history",[]), tool_obs)
     full_text = call_llm_streaming(client, state["model_name"], messages,
                                    emit_token=lambda t: ev.emit(sid,{"type":"token","content":t}))
@@ -486,10 +486,10 @@ app = Flask(__name__)
 app.secret_key = os.getenv("SECRET_KEY","lgsa-2025-dev-secret")
 
 AVAILABLE_MODELS = [
-    {"id":"mistralai/Mistral-7B-Instruct-v0.3","name":"Mistral 7B Instruct v0.3","badge":"⚡ Strong"},
-    {"id":"HuggingFaceH4/zephyr-7b-beta","name":"Zephyr 7B Beta","badge":"💬 Chat-tuned"},
-    {"id":"microsoft/Phi-3-mini-4k-instruct","name":"Phi-3 Mini 4K Instruct","badge":"🚀 Fast"},
-    {"id":"meta-llama/Meta-Llama-3-8B-Instruct","name":"Llama 3 8B Instruct","badge":"🦙 Powerful"},
+    {"id":"meta-llama/Meta-Llama-3.1-8B-Instruct","name":"Llama 3.1 8B Instruct","badge":"🦙 Recommended"},
+    {"id":"Qwen/Qwen2.5-7B-Instruct","name":"Qwen 2.5 7B Instruct","badge":"⚡ Fast"},
+    {"id":"mistralai/Mistral-7B-Instruct-v0.3","name":"Mistral 7B Instruct v0.3","badge":"🌀 Mistral"},
+    {"id":"google/gemma-2-9b-it","name":"Gemma 2 9B Instruct","badge":"💎 Google"},
 ]
 
 _sessions: dict = {}