PyBMS Tools for VS Code

A practical VS Code extension for Pybms, a beginner-friendly Python data science and machine learning framework (currently alpha, v0.2.0).

Features

• 100-function sidebar grouped into Datasets (20), Cleaning (20), Arrays (15), Plots (15), Training (15), Optimization (10) and Syntax & help (5).
• Autocomplete and hover documentation for imports such as import pybms as pb, followed by pb.. PyBMS already ships .pyi stubs for Pylance; this extension adds a framework explorer, shortcuts and recipes.
• Starter projects and six recipes for Iris AutoML, CSV training, cleaning, NumPy dense neural networks, gradient descent and syntax repair. Use PyBMS: Create Starter Project or PyBMS: Create Example File.
• Python snippets: type pbimport, pbauto, pbdata, pbclean, pbnet, pbplot or pbsave in a Python file.
• Environment check, install/upgrade, and run current Python file using VS Code tasks. Respects pybms.pythonPath and auto-detects .venv.
• Syntax checking using Python's ast.parse, without executing source code. Automatic checks happen only when a saved Python file imports PyBMS.
• Reviewable syntax repair using pybms.fix_syntax(): opens a VS Code diff and asks for explicit confirmation before editing the original. Unsupported errors are not silently altered.
• No API key, telemetry or paid service required. Framework installation is separate and requires permission.

Quick start

1. Install Python 3.10+. Install this VS Code extension.
2. Open a folder in VS Code. Click the PyBMS icon in the Activity Bar, or run PyBMS: Get Started from the Command Palette (Ctrl+Shift+P).
3. Choose Create Starter Project. Pick the Iris AutoML recipe.
4. Run PyBMS: Install or Update Framework to install the current GitHub source into your configured Python environment.
5. Open your generated .py file and run PyBMS: Run Current Python File.

For example:

import pybms as pb

data = pb.load_data('iris')
result = pb.auto_train(data, target='target')
print(result.metrics)

Settings

On Windows, a manually selected interpreter can be C:\\Python313\\python.exe. Set this under Settings → Extensions → PyBMS. This extension does not automatically activate the Python extension interpreter; configure the same environment in both when necessary.

Safety and limitations

• The Install/Upgrade action explicitly asks before running pip install --upgrade from the official Pybms GitHub archive. Network access is needed for installation.
• Run Current Python File runs your own file on request; it does not execute files in the background.
• Syntax diagnostics use the Python parser. The syntax repair feature imports an installed PyBMS and calls fix_syntax; the source is sent through stdin and is not run. The diff and confirmation allow review of all changes.
• Some functions show a short fallback signature in the offline sidebar (name(...)). For the exact signature, use the complete PyBMS API reference or python -m pybms help <name>.
• PyBMS 0.2.0 is an alpha framework. The recipes are illustrative, not assurances of model quality. Avoid training/test leakage; use appropriate group/time-series validation where needed. Only load trusted .joblib artifacts.
• The publisher ID in package.json is aymanbms, matching the official publisher account used to upload this extension.

Build / packaging

The extension uses plain JavaScript with no runtime npm dependencies:

node --check src/extension.js
npm test
python scripts/package_vsix.py

The packaging script writes dist/pybms-tools-0.1.2.vsix. For the standard Microsoft packaging path, you can instead run npx @vscode/vsce package (requires network access). It includes an icon, marketplace manifest, README and license.

To install a VSIX locally, open VS Code Extensions, choose ... → Install from VSIX, and select the .vsix file. Upload the same VSIX from your Visual Studio Marketplace publisher portal.

Developer links

• Framework source
• 100-function reference
• Extension source/issues

MIT License — Ayman Boufounas.