- Added a check to prevent sending a new command if one is already in progress
- Utilized UiUtilWrapper to display an error message when a second command is attempted
- Ensured that the check is reset after the command execution completes
- Ensured selection range now includes the last line
- Fixed an issue causing line duplication during refactoring
- Improves accuracy for function extraction in docstrings
- Comment out refDefsContext and defRefsContext usage in loadContexts
- Replace accessKey logic with defaultModel in several files
- Remove unused functions related to API key handling
Closes#286
- Removed redundant getFileContent function from diffHandler
- Centralized getFileContent into commonUtil for shared usage
- Adjusted imports in devchat to reflect the new location
- Updated package.json to include support for claude-3 models
- Removed references to claude-2 models in extension and ApiKeyManager
- Added API functionality and configured default settings for new models
- Wrap configuration update calls in try-catch to catch and log errors
- Ensure plugin initialization continues despite non-existent model data
- Log errors using `logger.channel()` for easier debugging and monitoring
- Added detailed comments to the isProviderHasSetted function to enhance understandability.
- Ensured comments include a brief description and the return type expectation.
- Aimed to address feedback on code documentation for better maintenance.
Closes#275
- Trim leading/trailing whitespace from OPENAI_API_KEY
- Prevent misleading 'APIConnectionError' due to key format
- Ensure API keys are in a valid format before network requests
Closes#265
- Added null checks to configuration retrieval to prevent failures
- Ensured default configurations are provided for missing settings
- Modified activation sequence to prevent errors with uninitiated UI utils
- Added a deactivate function to handle IDE shutdown events
- Utilize stopDevChatBase to terminate devchat subprocesses
- Included necessary import for the constants module
- Combine log insertion and retrieval steps into a single action
- Modify logInsert method to return an array of LogEntry objects
- Ensure backward compatibility by adapting existing code to new returns
- Convert file1 and file2 to lowercase before adding to filePairs
- Look up file pairs using lowercase file paths
- This addresses case sensitivity issues in file path on Windows
- Added a new endpoint /ide_name to return the IDE name
- Implemented handler to return hardcoded "vscode" as the IDE name
- Moved "/diff_apply" endpoint definition within functionRegistry
- Define model specific max_input_tokens to prevent overflow errors
- Ensure default token limits are set for each model configuration
- Integrate the default parameter setup within the activation process
- Deleted old topicView.ts as part of phasing out the view
- Removed the associated topicManager.ts which exclusively served topicView
- This commit will clean up the code after full transition to webview
- Deleted obsolete TopicManager & TopicTreeDataProvider classes
- Cleaned up Topic related logic across several modules
- Replaced TopicManager usage with new methods after webview migration
- Added "max_input_tokens" property to configuration settings
- Adjusted default values for "max_tokens" to ensure proper limit
- Renamed "gpt-4-1106-preview" to "gpt-4-turbo-preview" for consistency
- Simplified deletion logic in topicManager
- Now correctly records topic IDs in .deletedTopics
- Ensures topics are removed as intended from the application state
- Implement new topicHandler with getTopics and deleteTopic functions
- Refactor historyMessagesHandler to accept topicId for message retrieval
- Adjust historyMessagesBase function signature to take topicId as an argument
- Remove topic-related views, commands, and handlers from the codebase
- Refactor code to align with the new unified topic interaction method
- Streamline UI by eliminating unnecessary elements and consolidating views
- Add Model.CodeLlama-70b config settings support
- Implement Model.Mixtral-8x7B and Model.Minimax-abab6 support
- Modify UI and API integration to facilitate new model selections
- Include 'path' field in Command interface for directory awareness
- Add 'path' to command object creation for devchat run listing
- Update CommandEntry interface in devchat tool wrapper to hold 'path'
- Remove deprecated 1106 and 16k GPT-3.5 model configs
- Update GPT-3.5 config to latest stable version
- Update Claude, Xinghuo, GLM, ERNIEbot to latest versions
- Adjust default model to Claude 2.1
- Validate and test latest model integrations
- Remove unnecessary endpoint functions related to editor actions
- Simplify the JSON serialization process for VSCode data objects
- Add `runCode` function to facilitate executing code snippets
Closes#234
- Change condition for DocumentSymbol substructure processing
- Use symbol.children instead of instance type check
- Ensures complete information is returned by interface
