Merge branch 'main' into kevin

docs/modules/usage/configuration-options.md

@@ -385,6 +385,11 @@ To use these with the docker command, pass in `-e SANDBOX_<option>`. Example: `-
   - Default: `false`
   - Description: Use host network
 
+- `runtime_binding_address`
+  - Type: `str`
+  - Default: `127.0.0.1`
+  - Description: The binding address for the runtime ports.  It specifies which network interface on the host machine Docker should bind the runtime ports to.
+
 ### Linting and Plugins
 - `enable_auto_lint`
   - Type: `bool`

frontend/src/api/open-hands.ts

@@ -174,7 +174,7 @@ class OpenHands {
     code: string,
   ): Promise<GitHubAccessTokenResponse> {
     const { data } = await openHands.post<GitHubAccessTokenResponse>(
-      "/api/github/callback",
+      "/api/keycloak/callback",
       {
         code,
       },

frontend/src/i18n/declaration.ts

@@ -257,6 +257,7 @@ export enum I18nKey {
   STATUS$ERROR_LLM_AUTHENTICATION = "STATUS$ERROR_LLM_AUTHENTICATION",
   STATUS$ERROR_LLM_SERVICE_UNAVAILABLE = "STATUS$ERROR_LLM_SERVICE_UNAVAILABLE",
   STATUS$ERROR_LLM_INTERNAL_SERVER_ERROR = "STATUS$ERROR_LLM_INTERNAL_SERVER_ERROR",
+  STATUS$ERROR_LLM_OUT_OF_CREDITS = "STATUS$ERROR_LLM_OUT_OF_CREDITS",
   STATUS$ERROR_RUNTIME_DISCONNECTED = "STATUS$ERROR_RUNTIME_DISCONNECTED",
   STATUS$LLM_RETRY = "STATUS$LLM_RETRY",
   AGENT_ERROR$BAD_ACTION = "AGENT_ERROR$BAD_ACTION",
@@ -279,6 +280,7 @@ export enum I18nKey {
   ACTION_MESSAGE$EDIT = "ACTION_MESSAGE$EDIT",
   ACTION_MESSAGE$WRITE = "ACTION_MESSAGE$WRITE",
   ACTION_MESSAGE$BROWSE = "ACTION_MESSAGE$BROWSE",
+  ACTION_MESSAGE$THINK = "ACTION_MESSAGE$THINK",
   OBSERVATION_MESSAGE$RUN = "OBSERVATION_MESSAGE$RUN",
   OBSERVATION_MESSAGE$RUN_IPYTHON = "OBSERVATION_MESSAGE$RUN_IPYTHON",
   OBSERVATION_MESSAGE$READ = "OBSERVATION_MESSAGE$READ",
@@ -306,11 +308,8 @@ export enum I18nKey {
   STATUS$WAITING_FOR_CLIENT = "STATUS$WAITING_FOR_CLIENT",
   SUGGESTIONS$WHAT_TO_BUILD = "SUGGESTIONS$WHAT_TO_BUILD",
   SETTINGS_FORM$ENABLE_DEFAULT_CONDENSER_SWITCH_LABEL = "SETTINGS_FORM$ENABLE_DEFAULT_CONDENSER_SWITCH_LABEL",
-  BUTTON$ENABLE_SOUND = "BUTTON$ENABLE_SOUND",
-  BUTTON$DISABLE_SOUND = "BUTTON$DISABLE_SOUND",
   BUTTON$MARK_HELPFUL = "BUTTON$MARK_HELPFUL",
   BUTTON$MARK_NOT_HELPFUL = "BUTTON$MARK_NOT_HELPFUL",
-  NOTIFICATION$SOUND_ENABLED = "NOTIFICATION$SOUND_ENABLED",
-  NOTIFICATION$SOUND_DISABLED = "NOTIFICATION$SOUND_DISABLED",
   BUTTON$EXPORT_CONVERSATION = "BUTTON$EXPORT_CONVERSATION",
+  BILLING$CLICK_TO_TOP_UP = "BILLING$CLICK_TO_TOP_UP",
 }

frontend/src/services/actions.ts

@@ -59,6 +59,28 @@ const messageActions = {
       store.dispatch(appendJupyterInput(message.args.code));
     }
   },
+  [ActionType.FINISH]: (message: ActionMessage) => {
+    store.dispatch(addAssistantMessage(message.args.final_thought));
+    let successPrediction = "";
+    if (message.args.task_completed === "partial") {
+      successPrediction =
+        "The agent thinks that the task was **completed partially**.";
+    } else if (message.args.task_completed === "false") {
+      successPrediction =
+        "The agent thinks that the task was **not completed**.";
+    } else if (message.args.task_completed === "true") {
+      successPrediction =
+        "The agent thinks that the task was **completed successfully**.";
+    }
+    if (successPrediction) {
+      // if final_thought is not empty, add a new line before the success prediction
+      if (message.args.final_thought) {
+        store.dispatch(addAssistantMessage(`\n${successPrediction}`));
+      } else {
+        store.dispatch(addAssistantMessage(successPrediction));
+      }
+    }
+  },
 };
 
 export function handleActionMessage(message: ActionMessage) {

frontend/src/services/observations.ts

@@ -89,13 +89,24 @@ export function handleObservationMessage(message: ObservationMessage) {
         );
         break;
       case "read":
+        store.dispatch(
+          addAssistantObservation({
+            ...baseObservation,
+            observation,
+            extras: {
+              path: String(message.extras.path || ""),
+            },
+          }),
+        );
+        break;
       case "edit":
         store.dispatch(
           addAssistantObservation({
             ...baseObservation,
             observation,
             extras: {
               path: String(message.extras.path || ""),
+              diff: String(message.extras.diff || ""),
             },
           }),
         );

frontend/src/state/chat-slice.ts

@@ -173,9 +173,14 @@ export const chatSlice = createSlice({
           causeMessage.content
         }\n\nOutput:\n\`\`\`\n${content.trim() || "[Command finished execution with no output]"}\n\`\`\``;
         causeMessage.content = content; // Observation content includes the action
-      } else if (observationID === "read" || observationID === "edit") {
-        const { content } = observation.payload;
-        causeMessage.content = `\`\`\`${observationID === "edit" ? "diff" : "python"}\n${content}\n\`\`\``; // Content is already truncated by the ACI
+      } else if (observationID === "read") {
+        causeMessage.content = `\`\`\`\n${observation.payload.content}\n\`\`\``; // Content is already truncated by the ACI
+      } else if (observationID === "edit") {
+        if (causeMessage.success) {
+          causeMessage.content = `\`\`\`diff\n${observation.payload.extras.diff}\n\`\`\``; // Content is already truncated by the ACI
+        } else {
+          causeMessage.content = observation.payload.content;
+        }
       } else if (observationID === "browse") {
         let content = `**URL:** ${observation.payload.extras.url}\n`;
         if (observation.payload.extras.error) {

frontend/src/types/core/actions.ts

@@ -51,6 +51,8 @@ export interface ThinkAction extends OpenHandsActionEvent<"think"> {
 export interface FinishAction extends OpenHandsActionEvent<"finish"> {
   source: "agent";
   args: {
+    final_thought: string;
+    task_completed: "success" | "failure" | "partial";
     outputs: Record<string, unknown>;
     thought: string;
   };

frontend/src/types/core/observations.ts

@@ -70,6 +70,7 @@ export interface EditObservation extends OpenHandsObservationEvent<"edit"> {
   source: "agent";
   extras: {
     path: string;
+    diff: string;
   };
 }
 

microagents/knowledge/docker.md

@@ -0,0 +1,56 @@
+---
+name: docker
+type: knowledge
+version: 1.0.0
+agent: CodeActAgent
+triggers:
+- docker
+- container
+---
+
+# Docker Installation and Usage Guide
+
+## Installation on Debian/Ubuntu Systems
+
+To install Docker on a Debian/Ubuntu system, follow these steps:
+
+```bash
+# Update package index
+sudo apt-get update
+
+# Install prerequisites
+sudo apt-get install -y apt-transport-https ca-certificates curl gnupg lsb-release
+
+# Add Docker's official GPG key
+curl -fsSL https://download.docker.com/linux/debian/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg
+
+# Set up the stable repository
+echo "deb [arch=amd64 signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/debian $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
+
+# Update package index again
+sudo apt-get update
+
+# Install Docker Engine
+sudo apt-get install -y docker-ce docker-ce-cli containerd.io
+```
+
+## Starting Docker in Container Environments
+
+If you're in a container environment without systemd (like this workspace), start Docker with:
+
+```bash
+# Start Docker daemon in the background
+sudo dockerd > /tmp/docker.log 2>&1 &
+
+# Wait for Docker to initialize
+sleep 5
+```
+
+## Verifying Docker Installation
+
+To verify Docker is working correctly, run the hello-world container:
+
+```bash
+sudo docker run hello-world
+```
+

microagents/knowledge/kubernetes.md

@@ -0,0 +1,50 @@
+---
+name: kubernetes
+type: knowledge
+version: 1.0.0
+agent: CodeActAgent
+triggers:
+- kubernetes
+- k8s
+- kube
+---
+
+# Kubernetes Local Development with KIND
+
+## KIND Installation and Setup
+
+KIND (Kubernetes IN Docker) is a tool for running local Kubernetes clusters using Docker containers as nodes. It's designed for testing Kubernetes applications locally.
+
+IMPORTANT: Before you proceed with installation, make sure you have docker installed locally.
+
+### Installation
+
+To install KIND on a Debian/Ubuntu system:
+
+```bash
+# Download KIND binary
+curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.22.0/kind-linux-amd64
+# Make it executable
+chmod +x ./kind
+# Move to a directory in your PATH
+sudo mv ./kind /usr/local/bin/
+```
+
+To install kubectl:
+
+```bash
+# Download kubectl
+curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"
+# Make it executable
+chmod +x kubectl
+# Move to a directory in your PATH
+sudo mv ./kubectl /usr/local/bin/
+```
+
+### Creating a Cluster
+
+Create a basic KIND cluster:
+
+```bash
+kind create cluster
+```

openhands/agenthub/codeact_agent/function_calling.py

@@ -108,7 +108,10 @@ def response_to_actions(response: ModelResponse) -> list[Action]:
             # AgentFinishAction
             # ================================================
             elif tool_call.function.name == FinishTool['function']['name']:
-                action = AgentFinishAction()
+                action = AgentFinishAction(
+                    final_thought=arguments.get('message', ''),
+                    task_completed=arguments.get('task_completed', None),
+                )
 
             # ================================================
             # LLMBasedFileEditTool (LLM-based file editor, deprecated)

openhands/agenthub/codeact_agent/tools/finish.py

@@ -1,11 +1,39 @@
 from litellm import ChatCompletionToolParam, ChatCompletionToolParamFunctionChunk
 
-_FINISH_DESCRIPTION = """Finish the interaction when the task is complete OR if the assistant cannot proceed further with the task."""
+_FINISH_DESCRIPTION = """Signals the completion of the current task or conversation.
+
+Use this tool when:
+- You have successfully completed the user's requested task
+- You cannot proceed further due to technical limitations or missing information
+
+The message should include:
+- A clear summary of actions taken and their results
+- Any next steps for the user
+- Explanation if you're unable to complete the task
+- Any follow-up questions if more information is needed
+
+The task_completed field should be set to True if you believed you have completed the task, and False otherwise.
+"""
 
 FinishTool = ChatCompletionToolParam(
     type='function',
     function=ChatCompletionToolParamFunctionChunk(
         name='finish',
         description=_FINISH_DESCRIPTION,
+        parameters={
+            'type': 'object',
+            'required': ['message', 'task_completed'],
+            'properties': {
+                'message': {
+                    'type': 'string',
+                    'description': 'Final message to send to the user',
+                },
+                'task_completed': {
+                    'type': 'string',
+                    'enum': ['true', 'false', 'partial'],
+                    'description': 'Whether you have completed the task.',
+                },
+            },
+        },
     ),
 )

openhands/controller/agent_controller.py

@@ -744,10 +744,13 @@ async def _step(self) -> None:
             except Exception as e:
                 # FIXME: this is a hack until a litellm fix is confirmed
                 # Check if this is a nested context window error
+                # We have to rely on string-matching because LiteLLM doesn't consistently
+                # wrap the failure in a ContextWindowExceededError
                 error_str = str(e).lower()
                 if (
                     'contextwindowexceedederror' in error_str
                     or 'prompt is too long' in error_str
+                    or 'input length and `max_tokens` exceed context limit' in error_str
                     or isinstance(e, ContextWindowExceededError)
                 ):
                     if self.agent.config.enable_history_truncation:

openhands/core/config/sandbox_config.py

@@ -17,6 +17,7 @@ class SandboxConfig(BaseModel):
         remote_runtime_api_timeout: The timeout for the remote runtime API requests.
         enable_auto_lint: Whether to enable auto-lint.
         use_host_network: Whether to use the host network.
+        runtime_binding_address: The binding address for the runtime ports.  It specifies which network interface on the host machine Docker should bind the runtime ports to.
         initialize_plugins: Whether to initialize plugins.
         force_rebuild_runtime: Whether to force rebuild the runtime image.
         runtime_extra_deps: The extra dependencies to install in the runtime image (typically used for evaluation).
@@ -65,6 +66,7 @@ class SandboxConfig(BaseModel):
         default=False
     )  # once enabled, OpenHands would lint files after editing
     use_host_network: bool = Field(default=False)
+    runtime_binding_address: str = Field(default='127.0.0.1')
     runtime_extra_build_args: list[str] | None = Field(default=None)
     initialize_plugins: bool = Field(default=True)
     force_rebuild_runtime: bool = Field(default=False)
@@ -75,7 +77,7 @@ class SandboxConfig(BaseModel):
     close_delay: int = Field(default=15)
     remote_runtime_resource_factor: int = Field(default=1)
     enable_gpu: bool = Field(default=False)
-    docker_runtime_kwargs: str | None = Field(default=None)
+    docker_runtime_kwargs: dict | None = Field(default=None)
     selected_repo: str | None = Field(default=None)
 
     # custom configs

openhands/events/action/agent.py

@@ -1,4 +1,5 @@
 from dataclasses import dataclass, field
+from enum import Enum
 from typing import Any
 
 from openhands.core.schema import ActionType
@@ -44,16 +45,26 @@ def __str__(self) -> str:
         return ret
 
 
+class AgentFinishTaskCompleted(Enum):
+    FALSE = 'false'
+    PARTIAL = 'partial'
+    TRUE = 'true'
+
+
 @dataclass
 class AgentFinishAction(Action):
     """An action where the agent finishes the task.
 
     Attributes:
-        outputs (dict): The outputs of the agent, for instance "content".
+        final_thought (str): The message to send to the user.
+        task_completed (enum): Whether the agent believes the task has been completed.
+        outputs (dict): The other outputs of the agent, for instance "content".
         thought (str): The agent's explanation of its actions.
         action (str): The action type, namely ActionType.FINISH.
     """
 
+    final_thought: str = ''
+    task_completed: AgentFinishTaskCompleted | None = None
     outputs: dict[str, Any] = field(default_factory=dict)
     thought: str = ''
     action: str = ActionType.FINISH

openhands/events/observation/files.py

@@ -62,7 +62,12 @@ class FileEditObservation(Observation):
     new_content: str | None = None
     observation: str = ObservationType.EDIT
     impl_source: FileEditSource = FileEditSource.LLM_BASED_EDIT
-    _diff_cache: str | None = None  # Cache for the diff visualization
+    diff: str | None = (
+        None  # The raw diff between old and new content, used in OH_ACI mode
+    )
+    _diff_cache: str | None = (
+        None  # Cache for the diff visualization, used in LLM-based editing mode
+    )
 
     @property
     def message(self) -> str:
@@ -126,7 +131,7 @@ def visualize_diff(
         n_context_lines: int = 2,
         change_applied: bool = True,
     ) -> str:
-        """Visualize the diff of the file edit.
+        """Visualize the diff of the file edit. Used in the LLM-based editing mode.
 
         Instead of showing the diff line by line, this function shows each hunk
         of changes as a separate entity.