evocaition-vscext v0.1.1

This is the source code for the Visual Studio Code extension named evocaition, which gives the editor the ability to invoke the evocaition command-line tool to generate text using the content from the current editor window.

Features

• Seamless AI Integration: Connect to Cloud APIs (e.g., OpenRouter) or utilize Local AI solutions (e.g., LM Studio) for text generation.
• Customizable Experience: Tailor your text prediction with Configurable Sampler Settings.

The extension supports the same features that the evocaition command-line application supports, such as:

• Cloud API Integration: Use any OpenAI compatible endpoint, such as , to generate text with AI.
• Local AI Integration: Instead of a cloud provider, use an application like to run a local server process.
• Configurable Sampler Settings: Set values such as tempature, topK, topP, minP in the VS Code configuration JSON file to control prediction.

Requirements

An installed binary for evocaition must already be installed somewhere accessible by your PATH environment.

Usage

Make sure to set the API key first, if needed. This can be done through the Evocation: Set API Key command.

Running Text Prediction

The default keybinding for text prediction is ctrl+k, p or cmd+k, p on MacOS. This will run the evocaition app in the background using the sampler settings pulled from the configuration file to control the LLM text prediction. Once the task is complete, the text will be inserted at the cursor's location in the active window.

Another mode offered is bound to ctrl+k, ctrl+p or cmd+k, cmd+p on MacOS which will attempt to predict only one sentence. Even if more tokens are predicted, only the completion of the sentence will get inserted.

Setting Configuration Values

To set configuration values, you can use the commands available in the Command Palette:

1. Open the Command Palette (Ctrl+Shift+P or Cmd+Shift+P on macOS).
2. Type the command you want to run (e.g., Evocaition: Set API Key) and press Enter.
3. Follow the prompt to enter the desired value.

Alternatively, you can directly edit the User Settings (JSON) file:

1. Run the command Preferences: Open User Settings (JSON) (Cmd+, on macOS).
2. Add the appropriate key-value pairs for extension as documented below.

Configuration

API Settings

• evocaition.apiKey: (string) If using a cloud API like openrouter, set your API key here.
  • Example: "evocaition.apiKey": "<LONG API KEY TEXT>"
• evocaition.apiEndpoint: (string) Defaults to openrouter. Set a different API endpoint if needed (e.g., for local AI integration).
  • Example: "evocaition.apiEndpoint": "http://127.0.0.1:1234"

Model and Generation Settings

• evocaition.modelId: (string) The ID of the model to use for text generation.
  • Example: "evocaition.modelId": "meta-llama/llama-3.3-70b-instruct"
  • Note: The example above uses OpenRouter syntax. LM Studio would expect something like "llama-3.3-70b-instruct" for the same model or "llama-3.3-70b-instruct@q4_k_s" for specifying a quantization size.
• evocaition.maxTokens: (integer) The maximum number of tokens to generate.
  • Default: 0 (no limit)
  • Example: "evocaition.maxTokens": 512
• evocaition.sentenceMaxTokens: (integer) The maximum number of tokens to generate in 'sentence' mode.
  • Default: 100
  • Example: "evocaition.sentenceMaxTokens": 50
• evocaition.temperature: (number) The temperature setting for the model (controls randomness).
  • Default: 1.0
  • Example: "evocaition.temperature": 0.7

Sampler Settings

• evocaition.topP: (number) The top P value for nucleus sampling.
  • Default: 0.9
  • Example: "evocaition.topP": 0.8
• evocaition.minP: (number) The minimum P value for nucleus sampling.
  • Default: 0.0
  • Example: "evocaition.minP": 0.05
• evocaition.topK: (integer) The top K value for sampling.
  • Default: 0
  • Example: "evocaition.topK": 50
• evocaition.repPen: (number) The repetition penalty for the model.
  • Default: 0
  • Example: "evocaition.repPen": 1.04

Additional Settings

These settings are only able to be set manually in the user settings JSON file.

• evocaition.seed: (integer) An optional seed value for reproducibility.
  • Example: "evocaition.seed": 42
  • Note: Not currently configurable via a specific command; set directly in User Settings JSON.
• evocaition.documentContextCharacterLength (integer) An optional number of characters to pull from the active document to send to the AI as part of the prompt.
  • Default: 3000
  • Example: "evocaition.documentContextCharacterLength": 10240
• evocaition.promptTemplate (string) An optional string template to pass as the prompt for text generation. It uses nunjucks for template syntax.
  • Default: "You are a creative writing specialist AI. Continue the following text:\n\n{{ documentBody }}"
  • Example: "evocaition.documentContextCharacterLength": 10240

Installation

You can install the evocaition-vscext extension via the Visual Studio Code Marketplace:

1. Open Visual Studio Code.
2. Go to the Extensions view by clicking on the square icon on the Sidebar or pressing Ctrl+Shift+X.
3. In the search bar, type evocaition.
4. Click on the extension and then click Install.

Development Installs

Alternatively, after cloning from github you can install this locally yourself if you have all of the development tools installed via Node already:

npm install -g @vscode/vsce

The extension can be built into a package file and then installed:

vsce package
code --install-extension evocaition-vscext-<VERSION>.vsix

License

This project is licensed under the MIT License.