Merge branch 'master' into kpczerwinski/open-1186-agent-wrapper

.vscode/all-projects.code-workspace

@@ -0,0 +1,47 @@
+{
+  "folders": [
+    {
+      "name": "autogpt",
+      "path": "../autogpt"
+    },
+    {
+      "name": "benchmark",
+      "path": "../benchmark"
+    },
+    {
+      "name": "docs",
+      "path": "../docs"
+    },
+    {
+      "name": "forge",
+      "path": "../forge"
+    },
+    {
+      "name": "frontend",
+      "path": "../frontend"
+    },
+    {
+      "name": "autogpt_server",
+      "path": "../rnd/autogpt_server"
+    },
+    {
+      "name": "autogpt_builder",
+      "path": "../rnd/autogpt_builder"
+    },
+    {
+      "name": "[root]",
+      "path": ".."
+    }
+  ],
+  "settings": {},
+  "extensions": {
+    "recommendations": [
+      "charliermarsh.ruff",
+      "dart-code.flutter",
+      "ms-python.black-formatter",
+      "ms-python.vscode-pylance",
+      "prisma.prisma",
+      "qwtel.sqlite-viewer"
+    ]
+  }
+}

autogpt/autogpt/agent_factory/profile_generator.py

@@ -19,7 +19,7 @@
 
 
 class AgentProfileGeneratorConfiguration(SystemConfiguration):
-    model_classification: LanguageModelClassification = UserConfigurable(
+    llm_classification: LanguageModelClassification = UserConfigurable(
         default=LanguageModelClassification.SMART_MODEL
     )
     _example_call: object = {
@@ -148,21 +148,21 @@ class AgentProfileGenerator(PromptStrategy):
 
     def __init__(
         self,
-        model_classification: LanguageModelClassification,
+        llm_classification: LanguageModelClassification,
         system_prompt: str,
         user_prompt_template: str,
         create_agent_function: dict,
     ):
-        self._model_classification = model_classification
+        self._llm_classification = llm_classification
         self._system_prompt_message = system_prompt
         self._user_prompt_template = user_prompt_template
         self._create_agent_function = CompletionModelFunction.model_validate(
             create_agent_function
         )
 
     @property
-    def model_classification(self) -> LanguageModelClassification:
-        return self._model_classification
+    def llm_classification(self) -> LanguageModelClassification:
+        return self._llm_classification
 
     def build_prompt(self, user_objective: str = "", **kwargs) -> ChatPrompt:
         system_message = ChatMessage.system(self._system_prompt_message)

autogpt/autogpt/agents/agent.py

@@ -119,7 +119,7 @@ def __init__(
                 lambda x: self.llm_provider.count_tokens(x, self.llm.name),
                 llm_provider,
                 ActionHistoryConfiguration(
-                    model_name=app_config.fast_llm, max_tokens=self.send_token_limit
+                    llm_name=app_config.fast_llm, max_tokens=self.send_token_limit
                 ),
             )
             .run_after(WatchdogComponent)

autogpt/autogpt/agents/prompt_strategies/one_shot.py

@@ -100,7 +100,7 @@ def __init__(
         self.logger = logger
 
     @property
-    def model_classification(self) -> LanguageModelClassification:
+    def llm_classification(self) -> LanguageModelClassification:
         return LanguageModelClassification.FAST_MODEL  # FIXME: dynamic switching
 
     def build_prompt(

docs/content/forge/components/built-in-components.md

@@ -40,7 +40,7 @@ Necessary for saving and loading agent's state (preserving session).
 
 | Config variable  | Details                                | Type  | Default                            |
 | ---------------- | -------------------------------------- | ----- | ---------------------------------- |
-| `storage_path`     | Path to agent files, e.g. state        | `str` | `agents/{agent_id}/`[^1]           |
+| `storage_path`   | Path to agent files, e.g. state        | `str` | `agents/{agent_id}/`[^1]           |
 | `workspace_path` | Path to files that agent has access to | `str` | `agents/{agent_id}/workspace/`[^1] |
 
 [^1] This option is set dynamically during component construction as opposed to by default inside the configuration model, `{agent_id}` is replaced with the agent's unique identifier.
@@ -84,7 +84,7 @@ Keeps track of agent's actions and their outcomes. Provides their summary to the
 
 | Config variable        | Details                                                 | Type        | Default            |
 | ---------------------- | ------------------------------------------------------- | ----------- | ------------------ |
-| `model_name`           | Name of the llm model used to compress the history      | `ModelName` | `"gpt-3.5-turbo"`  |
+| `llm_name`             | Name of the llm model used to compress the history      | `ModelName` | `"gpt-3.5-turbo"`  |
 | `max_tokens`           | Maximum number of tokens to use for the history summary | `int`       | `1024`             |
 | `spacy_language_model` | Language model used for summary chunking using spacy    | `str`       | `"en_core_web_sm"` |
 | `full_message_count`   | Number of cycles to include unsummarized in the prompt  | `int`       | `4`                |
@@ -178,7 +178,7 @@ Allows agent to read websites using Selenium.
 
 | Config variable               | Details                                     | Type                                          | Default                                                                                                                      |
 | ----------------------------- | ------------------------------------------- | --------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------- |
-| `model_name`                  | Name of the llm model used to read websites | `ModelName`                                   | `"gpt-3.5-turbo"`                                                                                                            |
+| `llm_name`                    | Name of the llm model used to read websites | `ModelName`                                   | `"gpt-3.5-turbo"`                                                                                                            |
 | `web_browser`                 | Web browser used by Selenium                | `"chrome" \| "firefox" \| "safari" \| "edge"` | `"chrome"`                                                                                                                   |
 | `headless`                    | Run browser in headless mode                | `bool`                                        | `True`                                                                                                                       |
 | `user_agent`                  | User agent used by the browser              | `str`                                         | `"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36"` |

forge/forge/components/README.md

@@ -116,7 +116,7 @@ You can set sensitive variables in the `.json` file as well but it's recommended
         "github_username": null
     },
     "ActionHistoryConfiguration": {
-        "model_name": "gpt-3.5-turbo",
+        "llm_name": "gpt-3.5-turbo",
         "max_tokens": 1024,
         "spacy_language_model": "en_core_web_sm"
     },
@@ -129,7 +129,7 @@ You can set sensitive variables in the `.json` file as well but it's recommended
         "duckduckgo_max_attempts": 3
     },
     "WebSeleniumConfiguration": {
-        "model_name": "gpt-3.5-turbo",
+        "llm_name": "gpt-3.5-turbo",
         "web_browser": "chrome",
         "headless": true,
         "user_agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.97 Safari/537.36",

forge/forge/components/action_history/action_history.py

@@ -16,7 +16,7 @@
 
 
 class ActionHistoryConfiguration(BaseModel):
-    model_name: ModelName = OpenAIModelName.GPT3
+    llm_name: ModelName = OpenAIModelName.GPT3
     """Name of the llm model used to compress the history"""
     max_tokens: int = 1024
     """Maximum number of tokens to use up with generated history messages"""
@@ -97,7 +97,7 @@ def after_parse(self, result: AnyProposal) -> None:
     async def after_execute(self, result: ActionResult) -> None:
         self.event_history.register_result(result)
         await self.event_history.handle_compression(
-            self.llm_provider, self.config.model_name, self.config.spacy_language_model
+            self.llm_provider, self.config.llm_name, self.config.spacy_language_model
         )
 
     @staticmethod

forge/forge/components/web/selenium.py

@@ -55,7 +55,7 @@ class BrowsingError(CommandExecutionError):
 
 
 class WebSeleniumConfiguration(BaseModel):
-    model_name: ModelName = OpenAIModelName.GPT3
+    llm_name: ModelName = OpenAIModelName.GPT3
     """Name of the llm model used to read websites"""
     web_browser: Literal["chrome", "firefox", "safari", "edge"] = "chrome"
     """Web browser used by Selenium"""
@@ -164,7 +164,7 @@ async def read_webpage(
             elif get_raw_content:
                 if (
                     output_tokens := self.llm_provider.count_tokens(
-                        text, self.config.model_name
+                        text, self.config.llm_name
                     )
                 ) > MAX_RAW_CONTENT_LENGTH:
                     oversize_factor = round(output_tokens / MAX_RAW_CONTENT_LENGTH, 1)
@@ -382,7 +382,7 @@ async def summarize_webpage(
                 text,
                 topics_of_interest=topics_of_interest,
                 llm_provider=self.llm_provider,
-                model_name=self.config.model_name,
+                model_name=self.config.llm_name,
                 spacy_model=self.config.browse_spacy_language_model,
             )
             return "\n".join(f"* {i}" for i in information)
@@ -391,7 +391,7 @@ async def summarize_webpage(
                 text,
                 question=question,
                 llm_provider=self.llm_provider,
-                model_name=self.config.model_name,
+                model_name=self.config.llm_name,
                 spacy_model=self.config.browse_spacy_language_model,
             )
             return result

forge/forge/llm/prompting/base.py

@@ -10,7 +10,7 @@
 class PromptStrategy(abc.ABC):
     @property
     @abc.abstractmethod
-    def model_classification(self) -> LanguageModelClassification:
+    def llm_classification(self) -> LanguageModelClassification:
         ...
 
     @abc.abstractmethod

forge/forge/llm/providers/_openai_base.py

@@ -224,7 +224,7 @@ async def create_chat_completion(
                         tool_calls=tool_calls or None,
                     ),
                     parsed_result=parsed_result,
-                    model_info=self.CHAT_MODELS[model_name],
+                    llm_info=self.CHAT_MODELS[model_name],
                     prompt_tokens_used=t_input,
                     completion_tokens_used=t_output,
                 )
@@ -457,7 +457,7 @@ async def create_embedding(
 
         return EmbeddingModelResponse(
             embedding=embedding_parser(response.data[0].embedding),
-            model_info=self.EMBEDDING_MODELS[model_name],
+            llm_info=self.EMBEDDING_MODELS[model_name],
             prompt_tokens_used=response.usage.prompt_tokens,
         )
 

forge/forge/llm/providers/anthropic.py

@@ -309,7 +309,7 @@ async def create_chat_completion(
         return ChatModelResponse(
             response=assistant_msg,
             parsed_result=parsed_result,
-            model_info=ANTHROPIC_CHAT_MODELS[model_name],
+            llm_info=ANTHROPIC_CHAT_MODELS[model_name],
             prompt_tokens_used=t_input,
             completion_tokens_used=t_output,
         )

forge/forge/llm/providers/schema.py

@@ -186,7 +186,7 @@ class ModelResponse(BaseModel):
 
     prompt_tokens_used: int
     completion_tokens_used: int
-    model_info: ModelInfo
+    llm_info: ModelInfo
 
 
 class ModelProviderConfiguration(SystemConfiguration):

rnd/autogpt_builder/src/components/CustomNode.tsx

@@ -33,6 +33,7 @@ type CustomNodeData = {
 
 const CustomNode: FC<NodeProps<CustomNodeData>> = ({ data, id }) => {
   const [isPropertiesOpen, setIsPropertiesOpen] = useState(data.isPropertiesOpen || false);
+  const [isAdvancedOpen, setIsAdvancedOpen] = useState(false);
   const [keyValuePairs, setKeyValuePairs] = useState<{ key: string, value: string }[]>([]);
   const [newKey, setNewKey] = useState<string>('');
   const [newValue, setNewValue] = useState<string>('');
@@ -55,6 +56,16 @@ const CustomNode: FC<NodeProps<CustomNodeData>> = ({ data, id }) => {
     setIsPropertiesOpen(!isPropertiesOpen);
   };
 
+  const toggleAdvancedSettings = () => {
+    setIsAdvancedOpen(!isAdvancedOpen);
+  };
+
+  const hasOptionalFields = () => {
+    return data.inputSchema && Object.keys(data.inputSchema.properties).some((key) => {
+      return !(data.inputSchema.required?.includes(key));
+    });
+  };
+
   const generateHandles = (schema: Schema, type: 'source' | 'target') => {
     if (!schema?.properties) return null;
     const keys = Object.keys(schema.properties);
@@ -395,27 +406,37 @@ const CustomNode: FC<NodeProps<CustomNodeData>> = ({ data, id }) => {
     <div className={`custom-node dark-theme ${data.status === 'RUNNING' ? 'running' : data.status === 'COMPLETED' ? 'completed' : data.status === 'FAILED' ? 'failed' :''}`}>
       <div className="node-header">
         <div className="node-title">{data.blockType || data.title}</div>
-        <Button onClick={toggleProperties} className="toggle-button">
-          &#9776;
-        </Button>
+        <div className="node-buttons">
+          <Button onClick={toggleProperties} className="toggle-button">
+            &#9776;
+          </Button>
+          {hasOptionalFields() && (
+            <Button onClick={toggleAdvancedSettings} className="toggle-button">
+              &#9881;
+            </Button>
+          )}
+        </div>
       </div>
       <div className="node-content">
         <div className="input-section">
           {data.inputSchema &&
-            Object.entries(data.inputSchema.properties).map(([key, schema]) => (
-              <div key={key}>
-                <div className="handle-container">
-                  <Handle
-                    type="target"
-                    position={Position.Left}
-                    id={key}
-                    style={{ background: '#555', borderRadius: '50%' }}
-                  />
-                  <span className="handle-label">{key}</span>
+            Object.entries(data.inputSchema.properties).map(([key, schema]) => {
+              const isRequired = data.inputSchema.required?.includes(key);
+              return (isRequired || isAdvancedOpen) && (
+                <div key={key}>
+                  <div className="handle-container">
+                    <Handle
+                      type="target"
+                      position={Position.Left}
+                      id={key}
+                      style={{ background: '#555', borderRadius: '50%' }}
+                    />
+                    <span className="handle-label">{key}</span>
+                  </div>
+                  {renderInputField(key, schema)}
                 </div>
-                {renderInputField(key, schema)}
-              </div>
-            ))}
+              );
+            })}
         </div>
         <div className="output-section">
           {data.outputSchema && generateHandles(data.outputSchema, 'source')}

rnd/autogpt_builder/src/components/Flow.tsx

@@ -137,6 +137,30 @@ const FlowEditor: React.FC<{ flowID?: string; className?: string }> = ({
     [setEdges, setNodes]
   );
 
+  const onEdgesDelete = useCallback(
+  (edgesToDelete: Edge[]) => {
+    setNodes((nds) =>
+      nds.map((node) => ({
+        ...node,
+        data: {
+          ...node.data,
+          connections: node.data.connections.filter(
+            (conn: any) =>
+              !edgesToDelete.some(
+                (edge) =>
+                  edge.source === conn.source &&
+                  edge.target === conn.target &&
+                  edge.sourceHandle === conn.sourceHandle &&
+                  edge.targetHandle === conn.targetHandle
+              )
+          ),
+        },
+      }))
+    );
+  },
+  [setNodes]
+);
+
   const addNode = (blockId: string, nodeType: string) => {
     const nodeSchema = availableNodes.find(node => node.id === blockId);
     if (!nodeSchema) {
@@ -186,7 +210,7 @@ const FlowEditor: React.FC<{ flowID?: string; className?: string }> = ({
           title: `${block.name} ${node.id}`,
           inputSchema: block.inputSchema,
           outputSchema: block.outputSchema,
-          hardcodedValues: {},
+          hardcodedValues: node.input_default,
           setHardcodedValues: (values: { [key: string]: any; }) => {
             setNodes((nds) => nds.map((node) => node.id === newNode.id
               ? { ...node, data: { ...node.data, hardcodedValues: values } }
@@ -243,23 +267,6 @@ const FlowEditor: React.FC<{ flowID?: string; className?: string }> = ({
 
     let inputData = getNestedData(blockSchema, node.data.hardcodedValues);
 
-    // Get data from connected nodes
-    const incomingEdges = allEdges.filter(edge => edge.target === node.id);
-    incomingEdges.forEach(edge => {
-      const sourceNode = allNodes.find(n => n.id === edge.source);
-      if (sourceNode && sourceNode.data.output_data) {
-        const outputKey = Object.keys(sourceNode.data.output_data)[0]; // Assuming single output
-        inputData[edge.targetHandle as string] = sourceNode.data.output_data[outputKey];
-      }
-    });
-
-    // Filter out any inputs that are not in the block's schema
-    Object.keys(inputData).forEach(key => {
-      if (!blockSchema.properties[key]) {
-        delete inputData[key];
-      }
-    });
-
     console.log(`Final prepared input for ${node.data.blockType} (${node.id}):`, inputData);
     return inputData;
   };
@@ -433,6 +440,7 @@ const FlowEditor: React.FC<{ flowID?: string; className?: string }> = ({
         onEdgesChange={onEdgesChange}
         onConnect={onConnect}
         nodeTypes={nodeTypes}
+        onEdgesDelete={onEdgesDelete}
       >
         <div style={{ position: 'absolute', right: 10, zIndex: 4 }}>
           <Input