- Added logic to display a temporary chat message while waiting for the response from the agent.
- When a user sends a message, a temporary chat message is created and displayed.
- Implemented a method to remove the temporary chat message when the response is received or in case of an error.
- Updated the UI to invoke the method for adding a temporary chat message when the send button is pressed.
Feedback:
- This change improves the user experience by allowing them to see their own message immediately while waiting for the response from the agent.
- Added a boolean property `isWaitingForAgentResponse` to the `TaskViewModel` class to track whether a task is being created and waiting for a response from the agent.
- When a task is being created, we set `isWaitingForAgentResponse` to `true` and notify the listeners.
- When the task creation process is completed (successfully or not), `isWaitingForAgentResponse` is set to `false` and listeners are notified.
- Updated the `ChatView` class to listen to changes in `TaskViewModel.isWaitingForAgentResponse` to show the loading indicator conditionally.
A new property named _isWaitingForAgentResponse has been introduced to the ChatViewModel class to track the loading state when waiting for the agent's response. This boolean property is set to true when a chat message is being sent and reverts to false upon completion, whether successful or not. The notifyListeners() method is invoked to update the UI accordingly. This enhancement facilitates the display of a loading indicator, offering users visual feedback during the wait time for agent responses.
- Enhanced the `Chat` model to include an `artifacts` field to hold associated artifacts.
- Updated the `AgentMessageTile` widget to display the number of artifacts and added functionality to trigger artifact downloads upon button press.
- Introduced a method in `ChatViewModel` to leverage the `ChatService` for artifact downloads.
This commit incorporates significant enhancements to the SkillTreeViewModel, introducing the ability to track current benchmark runs and submit results to the leaderboard. A new list, `currentBenchmarkRuns`, is introduced to store each benchmark run object during a specific benchmark session. This list is reset to an empty state when initiating a new benchmark.
Changes made:
- Introduced `currentBenchmarkRuns` to track ongoing benchmark runs, ensuring real-time data availability.
- Enhanced `runBenchmark` method to populate `currentBenchmarkRuns` with benchmark run objects as the benchmark progresses.
- Implemented `submitToLeaderboard` method, accepting parameters `teamName`, `repoUrl`, and `agentGitCommitSha`, and updating each run object with this information. All runs share a common UUID generated at the beginning of the submission process.
These enhancements ensure that benchmark run data is readily available and organized, facilitating a streamlined process for submitting well-structured data to the leaderboard. It fosters a more interactive and informative user experience, offering insights into each benchmark run's progress and outcomes.
Refactored the TaskService and TaskViewModel classes to fully support paginated fetching of tasks. The solution fetches tasks page by page until all tasks have been retrieved, accommodating the API's pagination mechanism. Renamed `listAllTasks` to `fetchTasksPage` to better reflect its purpose and introduced a new method, `fetchAllTasks`, to handle the paginated fetching logic.
- Introduced a new `SkillTreeType` enum to represent different skill tree categories: General, Coding, Data, and Scrape/Synthesize.
- Extended the `SkillTreeType` enum to provide associated string values and JSON file names for each category.
- Refactored the `SkillTreeViewModel` to reload the skill tree data based on the selected category.
- Enhanced `SkillTreeView` by adding a positioned dropdown in the top-left corner to allow users to select and load different skill tree categories dynamically.
- Deprecated `ApiSettingsViewModel` in favor of enhancing `SettingsViewModel`.
- Moved the API Base URL field from `TaskView` to `SettingsView` to centralize configuration.
- Integrated `RestApiUtility` dependency into `SettingsViewModel` to ensure consistent URL management across the app.
Refactored the ChatViewModel's sendChatMessage method to utilize the continuousModeSteps setting from the SettingsViewModel. This allows the continuous chat mode to honor the user's preference from the settings. Also introduced an optional currentStep parameter with a default value, enabling better control over the progression of continuous mode.
This commit enriches the `SettingsViewModel` by integrating state persistence functionality and a sign-out method, thus ensuring a more robust and user-friendly settings feature.
Key Enhancements:
1. **State Persistence**:
- Leveraging the `shared_preferences` package, the app now stores and retrieves user preferences related to app settings, ensuring consistency across app restarts.
- Preferences like Dark Mode, Developer Mode, Base URL, and Continuous Mode Steps are persistently stored and loaded when the ViewModel is initialized.
2. **Sign Out Method**:
- A `signOut` method has been introduced in the ViewModel, utilizing the `AuthService` to facilitate user sign-out processes.
- This addition allows for seamless integration of sign-out functionality within the settings interface, granting users the ability to easily terminate sessions.
3. **SettingsView Enhancement**:
- The `SettingsView` has been adapted to incorporate a UI element invoking the `signOut` method, enhancing user interaction within the settings environment.
This enhancement not only bolsters the resilience and usability of the app’s settings but also lays down a structured approach for potential future additions to user preferences and settings-related functionalities.
This commit refines the `populateSelectedNodeHierarchy` method in the `SkillTreeViewModel` to accurately represent nodes that possess multiple parent nodes, ensuring a comprehensive and non-redundant representation of the entire hierarchy leading back to the root nodes.
Modifications and Features:
- The method now employs recursion to traverse all possible paths back to the root nodes from a selected node, capturing every unique node in the hierarchies.
- A `Set` is utilized to monitor and ensure that each node is only added once to the `_selectedNodeHierarchy` list, eliminating the possibility of duplicates.
- The finalized `_selectedNodeHierarchy` list is constructed such that the root of the tree is the last item in the list, providing a more logical representation of the hierarchy.
These enhancements ensure a more accurate and efficient representation of the skill tree structure, particularly in scenarios where nodes have multiple parents, facilitating better navigation and interaction within the skill tree.
This commit introduces the `SettingsViewModel` class, part of the MVVM architecture, responsible for managing the state and business logic for the settings feature of the app. This ViewModel is associated with the `SettingsView`.
Features Managed by this ViewModel:
- Dark Mode: Toggle the state of Dark Mode and notify listeners.
- Developer Mode: Toggle the state of Developer Mode and notify listeners.
- Base URL: Update the state of Base URL and notify listeners.
- Continuous Mode Steps: Increment and decrement the number of Continuous Mode Steps and notify listeners.
Each change in the state is followed by a notification to the listeners to rebuild the UI components that depend on this state, ensuring a reactive UI. To-do comments have been added where the state needs to be persisted or synchronized with a server.
This commit integrates the `LeaderboardService` into `SkillTreeViewModel` to enable benchmark report submissions to the leaderboard. A `BenchmarkRun` object is created from the evaluation response and submitted using the `submitReport` method from `LeaderboardService`.
* New benchmark data models
* Update _benchmarkBaseUrl
* Remove ReportRequestBody
* Update benchmark service methods for proxy approach
* Add eval id to SkillNodeData
* Refactor runBenchmark Method for proxy approach
This commit updates the runBenchmark method in the SkillTreeViewModel class to align with the new report generation flow. The updated method does the following:
1. Checks if a benchmark is already running to prevent overlapping runs.
2. Sets a flag to indicate that the benchmark is running and notifies the UI.
3. Reverses the selected node hierarchy for report generation.
4. Loops through each node in the reversed hierarchy to:
- Generate a unique UUID for each test run.
- Create a ReportRequestBody object.
- Call the generateSingleReport method in the BenchmarkService.
- Update the UI after each single report is generated.
5. After all single reports are generated, it calls the generateCombinedReport method in the BenchmarkService, passing in all the generated UUIDs.
6. Finally, it sets the benchmark running flag to false and notifies the UI.
This change improves the report generation flow and allows for both individual and combined reports.
This commit significantly expands the functionalities of TaskViewModel to manage both tasks and test suites in a unified manner. The view model now serves as the primary business logic class that interacts with the UI for task and test suite management.
Key Enhancements:
- Introduced `_testSuites` list to store TestSuite objects.
- Added `combinedDataSource` to hold both tasks and test suites.
- Introduced `selectTestSuite` and `deselectTestSuite` methods for TestSuite selection management.
- Added methods for TestSuite CRUD operations (`addTestSuite`, `fetchTestSuites`, `_saveTestSuitesToPrefs`).
- Created `fetchAndCombineData` method to fetch and combine tasks and test suites into a single list, `combinedDataSource`.
This update provides a more robust and unified approach for managing tasks and test suites, thereby improving the application's modularity and scalability.
Added a state variable isBenchmarkRunning in SkillTreeViewModel to track the status of benchmark execution. This state variable is used to conditionally disable specific UI components:
- The "Initiate test suite" button in TaskQueueView is disabled during the benchmark.
- All IconButtons in SideBarView are disabled during the benchmark.
- Node selection in SkillTreeView is disabled during the benchmark.
This ensures that the user cannot interact with these components while a benchmark test is running, thereby improving UX and preventing potential issues.
This commit extends the `SkillTreeViewModel` to include `BenchmarkService` as a dependency. This integration allows for leveraging benchmark-related API calls within the skill tree logic.
Two new methods have been added to `SkillTreeViewModel`:
1. `callGenerateReport`: This method attempts to call the `generateReport` function from the `BenchmarkService`. Currently, it only prints the API response and is incomplete in terms of full functionality.
2. `callPollUpdates`: Similar to `callGenerateReport`, this method aims to call `pollUpdates` from `BenchmarkService` and prints the API response. This is also incomplete and will require further development.
Both methods are preliminary and will require additional features to become fully functional.
