Skip to content

vscode.md

Latest commit

 

History

History
85 lines (59 loc) · 2.79 KB

vscode.md

File metadata and controls

85 lines (59 loc) · 2.79 KB
Raw

effGen VSCode Extension

Experimental. This extension is a developer-experience preview shipped from the repository (tools/vscode-effgen). It is not published to the VSCode Marketplace and is not covered by effGen's stability guarantees. Build it from source (below).

The effGen VSCode extension adds editor support for the effGen framework inside Visual Studio Code.

Features

Feature Description
Prompt-template completion Triggers on LibraryPrompt(, effgen.prompts., or %effgen_ — offers all built-in templates with snippet insertion.
Run code lens Shows a ▶ Run with effGen button above any LibraryPrompt(, effgen_chat(, %%effgen_agent, or %effgen_chat line. Clicking sends the prompt to the configured effGen server and prints the response to the effGen output channel.
Hover docs Hover over a quoted template name (e.g. "coding") to see the template's description, category, and usage snippet.
Prompt registry viewer Run effGen: Show Prompt Registry from the Command Palette (Ctrl+Shift+P) to open a webview listing all templates.

Requirements

  • VS Code ≥ 1.85
  • Node.js 18+ (for building the .vsix from source)

Installation

From source (development)

cd tools/vscode-effgen
npm install
npm run compile
# To package as .vsix:
# npm install -g @vscode/vsce
# vsce package

Then install the generated .vsix via Extensions → Install from VSIX…

Pre-built .vsix

Download vscode-effgen-0.3.2.vsix from the release page and install via Extensions → Install from VSIX…

Configuration

Open Settings (Ctrl+,) and search for effgen:

Setting Default Description
effgen.serverUrl http://localhost:8080 URL of your running effGen API server
effgen.defaultModel gpt-5-nano Default model id sent to the server for the Run code lens (use any id your server can serve)
effgen.enableCompletion true Toggle prompt-template auto-completion
effgen.enableHover true Toggle hover documentation

Usage

Prompt-template completion

In any Python file, start typing a LibraryPrompt( call and press Tab:

LibraryPrompt("cod|")  # ← Tab here completes to "coding" with a snippet

Run code lens

# A ▶ Run with effGen lens appears above this line ↓
result = effgen_chat("Explain gradient descent in one paragraph")

Clicking the lens sends the extracted prompt to effgen.serverUrl/v1/chat/completions and shows the response in the effGen output channel.

Hover docs

LibraryPrompt("reasoning", ...)
#              ↑ hover here for docs

Development

# Watch mode (incremental compile)
npm run watch