Feat/img gen tool and display

backend/app/core/db.py

@@ -74,7 +74,7 @@ def init_db(session: Session) -> None:
                 description=skill_info.description,
                 managed=True,
                 owner_id=user.id,
-                display_name=skill_info.display_name
+                display_name=skill_info.display_name,
             )
             session.add(new_skill)  # Prepare new skill for addition to the database
 
@@ -99,11 +99,7 @@ def init_modelprovider_model_db(session: Session) -> None:
         (1, 'Ollama', 'fakeurl', 'fakeapikey', 'string', 'string fake'),
         (2, 'Siliconflow', 'fakeurl', 'fakeapikey', 'string', 'siliconflow'),
         (3, 'zhipuai', 'https://open.bigmodel.cn/api/paas/v4', 'fakeapikey', 'zhipuai', '智谱AI')
-    ON CONFLICT (id) DO UPDATE
-        SET base_url = EXCLUDED.base_url,
-            api_key = EXCLUDED.api_key,
-            icon = EXCLUDED.icon,
-            description = EXCLUDED.description;
+    ON CONFLICT (id) DO NOTHING;
     """
 
     # Insert Models data

backend/app/core/tools/siliconflow/siliconflow.py

@@ -0,0 +1,51 @@
+import requests
+import json
+import requests
+from langchain.pydantic_v1 import BaseModel, Field
+from langchain.tools import StructuredTool
+
+
+class Text2ImageInput(BaseModel):
+    """Input for the text2img tool."""
+
+    prompt: str = Field(description="the prompt for generating image ")
+
+
+def text2img(
+    prompt: str,
+):
+    """
+    invoke tools
+    """
+
+    try:
+        # request URL
+        url = "https://api.siliconflow.cn/v1/image/generations"
+
+        payload = {
+            # "model": "black-forest-labs/FLUX.1-schnell",
+            "model": "stabilityai/stable-diffusion-3-medium",
+            "prompt": prompt,
+            "image_size": "1024x1024",
+        }
+        headers = {
+            "accept": "application/json",
+            "content-type": "application/json",
+            "authorization": "Bearer sk-uaxgsvfwwwpeuguzhsjpqigwopyhblsiesbptxnuxaoefqrb",
+        }
+
+        response = requests.post(url, json=payload, headers=headers)
+
+        return response.json()
+
+    except Exception as e:
+        return json.dumps(f"Openweather API Key is invalid. {e}")
+
+
+siliconflow = StructuredTool.from_function(
+    func=text2img,
+    name="Image Generation",
+    description="Image Generation is a tool that can generate images from text prompts.",
+    args_schema=Text2ImageInput,
+    return_direct=True,
+)

backend/app/core/workflow/init_graph.py

@@ -2,12 +2,11 @@
 from functools import lru_cache
 from typing import Any, Dict, Set
 
-from langchain.pydantic_v1 import BaseModel
 from langchain.tools import BaseTool
 from langchain_core.messages import AIMessage, AnyMessage
 from langchain_core.runnables import RunnableLambda
 from langgraph.checkpoint.base import BaseCheckpointSaver
-from langgraph.graph import END, StateGraph
+from langgraph.graph import END, START, StateGraph
 from langgraph.graph.graph import CompiledGraph
 from langgraph.prebuilt import ToolNode
 
@@ -200,31 +199,54 @@ def initialize_graph(
             source_node = next(node for node in nodes if node["id"] == edge["source"])
             target_node = next(node for node in nodes if node["id"] == edge["target"])
 
+            if source_node["type"] == "start":
+                if edge["type"] == "default":
+                    graph_builder.add_edge(START, edge["target"])
+                else:
+                    raise ValueError("Start node can only have normal edge.")
             if source_node["type"] == "llm":
                 if target_node["type"] == "tool":
-                    conditional_edges[source_node["id"]]["call_tools"][
-                        target_node["id"]
-                    ] = target_node["id"]
+                    if edge["type"] == "default":
+                        graph_builder.add_edge(edge["source"], edge["target"])
+                    else:
+                        conditional_edges[source_node["id"]]["call_tools"][
+                            target_node["id"]
+                        ] = target_node["id"]
                 elif target_node["type"] == "end":
-                    conditional_edges[source_node["id"]]["default"][END] = END
+                    if edge["type"] == "default":
+                        graph_builder.add_edge(edge["source"], END)
+                    else:
+                        conditional_edges[source_node["id"]]["default"][END] = END
+                elif target_node["type"] == "llm":
+                    if edge["type"] == "default":
+                        graph_builder.add_edge(edge["source"], edge["target"])
+                    else:
+                        conditional_edges[source_node["id"]]["default"][
+                            target_node["id"]
+                        ] = target_node["id"]
                 else:
-                    conditional_edges[source_node["id"]]["default"][
-                        target_node["id"]
-                    ] = target_node["id"]
+                    if edge["type"] == "default":
+                        graph_builder.add_edge(edge["source"], edge["target"])
+                    else:
+                        conditional_edges[source_node["id"]]["default"][
+                            target_node["id"]
+                        ] = target_node["id"]
 
             elif source_node["type"] == "tool" and target_node["type"] == "llm":
                 # Tool to LLM edge
                 graph_builder.add_edge(edge["source"], edge["target"])
 
         # Add conditional edges
+
         for llm_id, conditions in conditional_edges.items():
             edges_dict = {
                 "default": next(iter(conditions["default"].values()), END),
                 **conditions["call_tools"],
             }
             if conditions["call_human"]:
                 edges_dict["call_human"] = next(iter(conditions["call_human"].values()))
-            graph_builder.add_conditional_edges(llm_id, should_continue, edges_dict)
+            if edges_dict != {"default": END}:
+                graph_builder.add_conditional_edges(llm_id, should_continue, edges_dict)
 
         # Set entry point
         graph_builder.set_entry_point(metadata["entry_point"])

backend/app/core/workflow/node.py

@@ -253,7 +253,7 @@ async def work(self, state: TeamState, config: RunnableConfig) -> ReturnTeamStat
                         "If you are unable to perform the task, that's OK, you can ask human for help, or just say that you are unable to perform the task."
                         "Execute what you can to make progress. "
                         "And your role is:" + self.system_prompt + "\n"
-                        "And your name is:" + self.agent_name + "\n"
+                        "And your name is:" + self.agent_name + "please remember your name\n"
                         "Stay true to your role and use your tools if necessary.\n\n",
                     ),
                     (

backend/app/initial_data.py

@@ -15,7 +15,7 @@
 def init() -> None:
     with Session(engine) as session:
         init_db(session)
-        # init_modelprovider_model_db(session)
+        init_modelprovider_model_db(session)
 
 
 def main() -> None:

web/src/components/Playground/ChatMain.tsx

@@ -359,6 +359,7 @@ const ChatMain = ({ isPlayground }: { isPlayground?: boolean }) => {
             key={index}
             message={message}
             onResume={onResumeHandler}
+            isPlayground = {isPlayground}
           />
         ))}
       </Box>

web/src/components/Playground/MessageBox.tsx

@@ -39,9 +39,10 @@ import Markdown from "../Markdown/Markdown";
 interface MessageBoxProps {
   message: ChatResponse;
   onResume: (decision: InterruptDecision, toolMessage: string | null) => void;
+  isPlayground?: boolean;
 }
 
-const MessageBox = ({ message, onResume }: MessageBoxProps) => {
+const MessageBox = ({ message, onResume, isPlayground }: MessageBoxProps) => {
   const { type, name, next, content, tool_calls, tool_output, documents } =
     message;
   const [decision, setDecision] = useState<InterruptDecision | null>(null);
@@ -89,10 +90,12 @@ const MessageBox = ({ message, onResume }: MessageBoxProps) => {
     <VStack spacing={0} my={4} onMouseEnter={onOpen} onMouseLeave={onClose}>
       <Box
         w="full"
-        ml={10}
-        mr={10}
-        pl={10}
-        pr={10}
+
+        ml={isPlayground ? "10" : "0"}
+        mr={isPlayground ? "10" : "0"}
+        pl={isPlayground ? "10" : "0"}
+        pr={isPlayground ? "10" : "0"}
+
         display="flex"
         alignItems="center"
         justifyContent={type === "human" ? "flex-end" : "flex-start"}

web/src/components/WorkFlow/nodes/nodeConfig.ts

@@ -63,7 +63,7 @@ export const nodeConfig: Record<string, NodeConfigItem> = {
       targets: ["left", "right"],
     },
     initialData: {
-      tools: ["open-weather"],
+      tools: ["Open Weather"],
     },
   },
   questionClassifier: {