CHANGELOG.md

Changelog

All notable changes to this project will be documented in this file. The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

2.14.3-241.1 - 2024-12-17

Added

• Support for copying user and response messages in their original format #791
• Right-click popup menu for response messages

Fixed

• Escaped code response characters #169, #437
• Model value change for empty conversations
• Order of Custom OpenAI applied settings #797

2.14.2-241.1 - 2024-12-12

Added

• Partial code completion acceptance

2.14.1-241.1 - 2024-12-06

Fixed

• Response streaming for o1 models
• IDE error caused by credential validation in EDT
• Submission handling when current response is ongoing

2.14.0-241.1 - 2024-12-04

Added

• Support for multi-line code completions
• Recent Qwen 2.5 Coder models (llama.cpp)
• Codestral as default code model (CodeGPT)

Fixed

• NPE caused by lookup element sorter

Removed

• StarCoder 7B model (CodeGPT)

2.13.1-241.1 - 2024-11-27

Fixed

• NPE when using IDE internal completions #777
• Inline completion triggering when editor has active lookup

2.13.0-241.1 - 2024-11-25

Added

• Single standalone view for prompt configuration
• Codestral model for code completion (CodeGPT)

Fixed

• Git diff logic for commit message generation

2.12.5-241.1 - 2024-11-18

Added

• Option to explain commits from VCS log tree #688

Fixed

• Chat response streaming for messages received faster than polling rate #757

Improved

• Default CodeGPT prompt instructions

2.12.4-241.1 - 2024-11-14

Added

• Qwen 2.5 Coder model for chat and code completion (CodeGPT)

Fixed

• Image popup notification condition fix #711

Improved

• Code completion user experience
  • Display and apply completions line by line
  • Make suggestion available as soon as new line is available
  • Add logic for whitespace adjustments
  • Post-insertion logic, i.e., move cursor to the beginning of next completion

2.12.3-241.1 - 2024-11-07

Fixed

• Files not persisting between messages in same session

2.12.2-241.1 - 2024-11-04

Fixed

• Ollama chat completion streaming #744
• Non-stream chat completions error handling (CodeGPT, OpenAI) #746
• Prompt text component height issues #747
• Endless loading when persona name or instructions are null #748

2.12.1-241.1 - 2024-11-01

Added

• Gemini 1.5 Pro model (CodeGPT)

Fixed

• Code completion state when session contains active completion

Removed

• StarCoder 16B model (CodeGPT)

2.12.0-241.1 - 2024-10-29

Added

• 'Auto Apply' feature for instant AI-suggested code updates #743 (CodeGPT)

Fixed

• Code completion enabled/disabled state
• Total token calculation for highlighted texts

2.11.7-241.1 - 2024-10-18

Fixed

• Edit Code request building #737 (OpenAI)
• Chat UI performance issues
• Minor issues related to git repository lookup

2.11.6-241.1 - 2024-10-15

Fixed

• High CPU usage #716
• NPE on suggestion actions when no files are opened in editor #725
• Inconsistent text rendering in the chat tool window

2.11.5-241.1 - 2024-10-10

Added

• Support high context limits (up to 200k tokens) via RAG when chatting with files (CodeGPT)
• Gemini 1.5 Pro model (Google)
• Legacy response support for text completions in code completions (Custom OpenAI)

Fixed

• Chat UI threading issues #730

2.11.4-241.1 - 2024-10-01

Added

• Qwen 2.5 (72B) and o1 models (CodeGPT, OpenAI)
• Free limited access to SOTA models (CodeGPT)

2.11.3-241.1 - 2024-09-25

Fixed

• Anthropic (Claude) chat requests #707

2.11.2-223 - 2024-09-24

Added

• Edit Code and Name Suggestions feature support for all providers
• Qwen 2.5 Coder models and prompt templates

Fixed

• Model popup submenu auto-hover #681
• Prompt text field newline creation #694
• Prompt text field theme issues #701

2.11.1-223 - 2024-09-12

Added

• Git history referencing in prompts
• Quick code snippet referencing in prompts #690
• Name suggestions for languages other than Java #615

Fixed

• Prompt text field persona suggestions #685
• Text field autofocus when opening the chat toolwindow #691
• Git commit dialog selected changes not being included in diff
• Ollama server host configuration overriding #684

Improved

• Model combobox menu UI/UX by extracting the providers into standalone subgroups #681
• Prompt text field inlay rendering

2.11.0-223 - 2024-09-05

Added

• Multiple tool window editor actions for generated code:
  • Direct apply: Directly replaces the highlighted code in the main editor with the generated code
  • Compare with Original: Opens a diff view to compare the generated code with the highlighted code
  • Insert at Caret: Inserts the generated code at the exact location of the caret in the main editor
• Vision support for Azure models
• General improvements to code completions, including:
  • Proper streaming support
  • Git context integration

Fixed

• Prompt text field background transparency issues
• OpenRouter default template for custom OpenAI provider
• Ollama streaming requests

2.10.2-223 - 2024-08-23

Added

• Progress display for web docs

Fixed

• Code completion prompt creation from placeholders
• Prompt text field inlay offset and up/down key events

2.10.1-223 - 2024-08-20

Fixed

• Codestral Infill templates for code completion in llama.cpp #667

2.10.0-223 - 2024-08-19

Added

• Web documentation support #650
• Improved user prompt text field component #665
  • Spellchecking
  • Auto-enclosing brackets, quotes, etc.
  • Removable suggestion action inlays
  • Context menu for copy-pasting, and more
• DeepSeek Coder V2 model for CodeGPT users
• Image support for Custom OpenAI-compatible providers

Fixed

• IDE freezes caused by long running file search queries #652

2.9.0-223 - 2024-08-03

Added

• Web search support - available only through the CodeGPT provider #641
• Customizable Personas - a quick way to define and switch between different system prompts. Initial release comes with over 600 pre-defined prompts. #638

Other Improvements

• The textbox suggestions popover now offers better usability, making it easier and quicker to add additional context to your session for more accurate responses.
• Support for attaching folders, allowing all files within the folder to be added to your current chat session.
• Optimized the chat user interface for better performance. We noticed that for faster models, the inference speed didn't match what was being rendered on the screen.

2.8.5-223 - 2024-07-24

Added

• Llama 3.1 405B model for CodeGPT Individual plan users

Fixed

• Wrong line separators exception during code completion (#637)

2.8.4-223 - 2024-07-19

Added

• GPT-4o mini model (free with our hosted service)
• New UI/UX design for the toolwindow's user input component
• Support for searching for and attaching files in the textarea popover when the '@' key is used

Fixed

• Numeric values in Custom providers' headers section (#619)
• Ollama settings sync (#616)

2.8.3-223 - 2024-07-15

Added

• Enter key shortcut for Fast Edit submission (#602)

Fixed

• Ollama form combobox state (#606)
• Replace code behaviour (#614)
• Wrong line separators on Windows (#605)

2.8.2-233 - 2024-07-01

Added

• Fast Code Edits #601

Fixed

• Chat toolwindow autoscroller issues
• Wrong line separators on Windows machine

Removed

• You.com provider support

2.8.1-223 - 2024-06-21

Added

• Claude 3.5 Sonnet model (CodeGPT)
• Additional maxTokens, temperature and apiKey fields (Ollama)

Fixed

• Model re-selection after refresh (Ollama)
• Commit message placeholders

2.8.0-223 - 2024-06-06

Added

• OpenAI and Anthropic models for CodeGPT subscribed users
• Support Stable Code Instruct 3B (#552)
• Support Phi-3 Medium 128K (#577)
• Support Codestral 22B 32K (#587)
• Add field for environment variables for Llama server (#550)
• Add OpenRouter service template (#581)

Fixed

• Remove trailing slashes from URL text fields (#579)
• Disable code completions by default

2.7.1-223 - 2024-05-15

Fixed

• Backward compatibility issues on plugin updates (#551)

Added

• Display a model name when starting the llama.cpp server (#546)

2.7.0-223 - 2024-05-14

Added

• Three new providers - Ollama, Google and CodeGPT 🚀🚀🚀
• OpenAI GPT-4o model (#547)
• New provider configuration user interface (#538)
• Support Phi-3 Mini, CodeGemma and CodeQwen1.5 models (#516, #524, #525, #527)
• Llama 3 download sizes (#498)
• Mistral AI service template (#532)
• Support for starting/stopping llama.cpp server from statusbar (#544)

Fixed

• Commit message generation for Custom OpenAI services (#496)
• NPE when using unsupported model for code completions (#499)
• Editor Actions configuration UI issues (#518)
• IDE error caused by Git4Idea dependency (#526)
• Custom Service test connection (#531)
• UI issues around llama.cpp provider configuration (#543, #529)

Removed

• max_tokens configuration for code completions (will be handled internally)

2.6.3-223 - 2024-04-22

Added

• Support for Llama 3 model via llama.cpp port (#479)
• Code completion for "Custom OpenAI Service" (#476)
• Support for configuring llama.cpp server build parameters (#481)
• "Include file in context" to editor context menu (#475)
• Support for placeholders in the commit message system prompt (#458)

Fixed

• High CPU usage during new files check (#474)
• Persistence of credentials back into the PasswordSafe (#465)

2.6.2-222 - 2024-04-15

Fixed

• Text rendering anomalies upon streaming

2.6.1-222 - 2024-04-12

Fixed

• EncodingManager error handling for invalid inputs (#444)
• Initial default values for checkForPluginUpdatesCheckBox and checkForNewScreenshotsCheckBox (#446)
• Random IDE crashes caused by tree-sitter (#452, #446)
• Azure base url creation (#449)

2.6.0-222 - 2024-04-10

Added

• Vision support (image understanding) for OpenAI GPT-4 and Anthropic Claude models
• Total token panel for all providers
• Autofocus for UserTextArea when the tool window is visible

Fixed

• Git commit message generation
• Error when adding a single file to the context
• Several IntelliJ platform warnings

Removed

• Azure custom configuration (use OpenAI-compatible service to override the default configuration)

2.5.1 - 2024-03-14

Added

• Support for You Pro modes
• Basic post-processing for code completions
• Code completion feature toggle keyboard-shortcut
• Support for git commit message generation with Custom OpenAI and Anthropic services

Fixed

• Several IntelliJ platform warnings

Changed

• Supported minimum IDE build from 213 to 222

2.5.0 - 2024-03-06

Added

• Support for Anthropic Claude service

Fixed

• IndexOutOfBoundsException for Azure completions

2.4.0 - 2024-02-26

Added

• Support for custom OpenAI-compatible service (#383)
• Llama cpp for generation of git commit message (#380)
• Remote server settings for Windows + Mixtral Instruct template (#378)

Fixed

• Proxy settings for azure client (#382)

2.3.1 - 2024-02-19

Added

• Support for custom OpenAI model configuration (#250)

Fixed

• General Settings isModified state
• Azure service credential condition (#375)
• Caret offset location upon document changes (#367)

2.3.0 - 2024-02-14

Added

• Support for autocompletion (disabled by default)
• Support for auto resolving compilation errors

Replaced

• OpenAI chat models with the most recent ones
• CodeGPT header key

2.2.12 - 2023-12-21

Added

• Support for extended request parameters for llama.cpp

2.2.11 - 2023-12-14

Added

• Support for chatting with multiple files
• Support for generating commit messages and method names with Azure service

2.2.10 - 2023-12-12

Added

• Configurable commit-message system prompt

Fixed

• Redundant chat tab creation on provider/model change
• Azure bas host and path overriding

2.2.9 - 2023-12-04

Added

• IDE notification on new plugin versions

Fixed

• LLaMA settings state on server failure

Improved

• LLaMA server logging

2.2.8 - 2023-12-01

Fixed

• UI concurrency issues (run completion events on EDT)

2.2.7 - 2023-11-30

Fixed

• LLaMA server boot up when additional params are not set

2.2.6 - 2023-11-27

Added

• Support for automatic code formatting (#262)
• Support for additional command-line params for the llama server startup process
• Support for changing between different provides/models in the chat window (#227)
• Deepseek coder instruct models

Fixed

• Git commit message generation based on the actual user selected/checked files (#291)

Improved

• Toolwindow Chat UI

2.2.5 - 2023-11-23

Added

• Option to set the number of threads for local LLM models

Fixed

• Tool window chat editor enabled/disabled logic
• Tool window text pane caret visibility
• Editor actions for createNewChatOnEachAction state
• Plugin from crashing when the LLaMA server process is stuck

2.2.4 - 2023-11-20

Added

• Support for git commit message generation
• Support for method/function name lookup generation

Fixed

• Llama client read/connect timeout configuration

2.2.3 - 2023-11-15

Added

• Expand/Collapse logic for tool window editors
• Interactive total token count label

Removed

• Azure model selection (redundant field)

2.2.2 - 2023-11-07

Added

• OpenAI GPT-4 Turbo and the latest GPT-3.5 models

Fixed

• Credential validation when using llama service

2.2.1 - 2023-11-06

Fixed

• Right click context menu (#251)
• Azure path (#232)
• Llama telemetry service value
• Include all feature builds by removing the value (#242)

2.2.0 - 2023-11-03

Added

• Support for running local LLMs via LLaMA C/C++ port (#249)
• Support for fetching You.com subscription

Fixed

• Project build for Windows ARM64

Improved

• You.com coupon design
• Overall UI design

2.1.7 - 2023-10-26

Replaced

• You.com GPT-4 toggle component with checkbox
• Telemetry segment service with rudderstack

Fixed

• You.com web search not being displayed regardless of the flag

2.1.6 - 2023-10-12

Added

• Default destination value for toolwindow editor save file action
• Suggestion to switch to different LLM provider on quota exceeded error
• Support for You.com GPT-4 model

Fixed

• The style of wrapping used if the text area is wrapping lines
• Memory leaks (correct disposals for Configurable classes)

2.1.5 - 2023-10-02

Added

• Support for diffing generated code

2.1.4 - 2023-10-01

Added

• Support for modifying generated code within the chat toolwindow
• Support for creating new files directly from the chat toolwindow

Fixed

• Privacy policy url in the settings

2.1.3 - 2023-09-28

Added

• More telemetry actions

Fixed

• Error logging

2.1.2 - 2023-09-28

Fixed

• Incorrect window selection during chat or toolwindow actions

2.1.1 - 2023-09-27

Added

• Telemetry based on redhat impl

2.1.0 - 2023-09-21

Added

• You API integration
• Ability to override completion path

Updated

• Chat History UI

2.0.6 - 2023-08-29

Removed

• Functionality to fetch/use OpenAI account name

Fixed

• Rendering user input's newlines

Added

• Support for closing other tabs (#172)
• Support for configuring custom hosts for OpenAI and Azure services

2.0.5 - 2023-06-12

Fixed

• Tool window not opening on editor actions (#157)

Added

• Support for changing the editor action behaviour (#157)
• Support for overriding completions request parameters (#152)
• User text area autofocus on creating a new chat (#155)
• Keymap tool window actions (#148)

2.0.4 - 2023-05-27

Fixed

• TypeScript and C# code highlighting

• ToolWindow usability when virtual space option turned ON (#125)

Added

• ToolWindow code editor copy/replace header actions
• Custom prompt main editor action (#144)
• Support for 2023.2 EAP builds (#149)

2.0.3 - 2023-05-18

Fixed

• Empty editor context menu item text (#137)
• Temp file path resolving (#130)

Improved

• Response streaming

Added

• Reset chat window toolbar action (#138)

2.0.2 - 2023-05-16

Fixed

• Settings deserialization error

Improved

• Memory consumption by disposing unused editors which are no longer needed

Removed

• Main editor focus stealing on response streaming

2.0.1 - 2023-05-14

Added

• New GPT-3.5-16k model

Fixed

• NPE when displayName couldn't be fetched

Improved

• Proxy support by disabling the default trustmanager

Removed

• Off-screen rendering setting option
• Automatic textarea focus on stream completion (#126)

2.0.0 - 2023-05-03

Added

• Automatic retry logic on stream timeouts

Improved

• Input prompt text field UI/UX (height grows with the content)

Removed

• Custom prompt editor action (users can now provide custom actions within the chat window itself)

Replaced

• ToolWindow HTML content with native Swing components

Secured

• OPENAI_API_KEY persistence, key is saved in the OS password safe from now on