VSCode N-Gram Code Suggester
⚠️ Experimental Project – Not for Production Use
This repository demonstrates a research prototype that implements a simple n‑gram language model for code completion in VS Code. It is not guaranteed to be stable, fast, or secure enough for production workloads. Use at your own risk.
Overview
VSCode N‑Gram Code Suggester is a proof‑of‑concept that combines:
| Component |
Description |
| VS Code extension |
Hooks into the editor’s Inline Suggest API and uses the trained n‑gram model to surface context‑aware completions. |
| Model file |
A lightweight model that stores trigram/tag frequencies. |
The idea is to show that even a single‑sentence context can yield useful suggestions, without the heavy machinery of large neural models.
Current functional
- Autocompletion generation based on a pre-trained model
- Project context awareness (open documents are indexed and included in autocompletion generation)
- Support for model training and autocompletion generation for: C#, JavaScript, TypeScript, Python
⚠️ Important about configuration
If you use a large model (more than 2 million patterns), you may need to disable "Fuzzy search" and "Use Smoothing" for better performance. If suggestions are still slow, enable "Use Trigger Characters" option. Pre-trained model in builded extension is so large.
For own smaller models, keep these settings enabled for better suggestion quality.
Extension Settings
Below is a quick reference to all user‑configurable options for the extension.
Add any of these to your workspace or user settings.json to tweak the behaviour.
| Setting | Type | Default | Constraints | Description |
| -------------------------------------- | ------- | ------------------------ | ----------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| codeSuggester.modelPath | string | ./models/model.json.gz | – | Path to the trained model file. Supports plain .json or gzipped .json.gz. |
| codeSuggester.maxSuggestions | number | 5 | 1 – 10 | Maximum number of suggestions displayed in the IntelliSense list. |
| codeSuggester.maxFuzzyChecks | number | 2000 | ≥ 1000 | Maximum number of fuzzy‑search checks performed. Higher values give better matches but can be slow on large models. |
| codeSuggester.minConfidence | number | 0.2 | 0.0 – 1.0 | Minimum confidence threshold for a suggestion to be shown. |
| codeSuggester.enableFuzzyMatching | boolean | false | – | Turns on fuzzy matching for similar code patterns. ⚠️Use only on small models⚠️ |
| codeSuggester.useSmoothing | boolean | false | – | Enables smoothing algorithms to better handle rare n‑grams. ⚠️Use only on small models⚠️ |
| codeSuggester.useTriggerCharacters | boolean | false | – | When enabled, suggestions are only triggered when the cursor is placed on a trigger character (. , ( ) [ { : ; =). Useful if auto‑suggestions feel sluggish. |
| codeSuggester.useProjectContext | boolean | true | - | Use project context from open files for suggestions |
| codeSuggester.updateOnFileChange | boolean | false | - | Update project model when files are modified (may impact performance) |
Model Training
For information on training your own models, please refer to the full documentation on GitHub.
Build Extension from sources
For information how to build extension from sources, please refer to the full documentation on GitHub.
License
Distributed under the MIT License. See the LICENSE file for details.
