🔬 SciViewer — Seamless Python Plot Visualization in VS Code

Capture matplotlib plots automatically and view them inside VS Code — no X11 forwarding, no savefig boilerplate, no broken GUI backends over SSH.

SciViewer intercepts your plt.show() calls, saves each figure, and displays it in a live panel right inside the editor. It's built for researchers and data scientists working in remote Python environments where opening a plot window is painful or impossible.

✨ Features

• Zero-setup auto-injection — open a Python file, run it, and plots are captured automatically. No code changes.
• Live plot panel — figures appear in a VS Code panel as soon as plt.show() runs.
• Thumbnail gallery — browse recent plots; zoom and pan for detailed inspection.
• Rich metadata — each plot records its source script, function, line number, figure size, DPI, and axes info.
• Jupyter-aware — detects and patches notebook kernels alongside scripts.
• Automatic cleanup — keeps the most recent 100 plots, removes older ones.
• Status bar integration — shows whether injection is active and how many sessions are patched.
• Cross-platform — Windows, macOS, Linux.

🚀 Quick Start

1. Install SciViewer from the VS Code Marketplace (or search "SciViewer" in the Extensions view).
2. Open any Python file that uses matplotlib.
3. Run it. Your plots appear in the SciViewer panel automatically.

import matplotlib.pyplot as plt
import numpy as np

x = np.linspace(0, 10, 100)
plt.plot(x, np.sin(x))
plt.title("Auto-captured by SciViewer")
plt.show()   # → appears in the SciViewer panel

No %matplotlib magic, no backend configuration, no manual savefig.

🔧 How It Works

SciViewer injects a small Python patch (monkey_patch.py) into your active Python session. The patch replaces matplotlib.pyplot.show() and Figure.show() so that, instead of opening a GUI window, each figure is:

1. Saved as a high-quality PNG (150 DPI) to a temp folder ($TMPDIR/sciviewer).
2. Recorded with metadata (source location, figure properties, timestamp).

The extension watches that folder and renders new plots in its panel. Because the capture happens at the matplotlib level, it works the same locally, over SSH Remote, in containers, and in Codespaces — anywhere your Python runs.

Note: the patch overrides plt.show() for the patched session, so plots are routed to SciViewer rather than a native window. Originals are stored so behavior can be restored.

⚙️ Configuration

Open settings (Ctrl+,) and search for "SciViewer":

🎮 Commands

Run via the Command Palette (Ctrl+Shift+P):

📋 Requirements

• VS Code 1.93.0 or newer
• Python with matplotlib installed (pip install matplotlib)

🐛 Troubleshooting

Plots not appearing?

1. Check the status bar for injection status.
2. Confirm matplotlib is installed: pip install matplotlib.
3. Trigger manual injection: Ctrl+Shift+P → SciViewer: Inject Into Current Python Session.
4. Review settings: Ctrl+, → search "sciviewer".

Using remote development? SciViewer works with VS Code Remote SSH, GitHub Codespaces, Docker containers, and WSL — the capture runs wherever Python runs.

🗺️ Roadmap

• Multi-library support (Seaborn, Plotly, Bokeh)
• Group plots by experiment/run
• Export helpers (figure numbering, LaTeX)
• Restore-to-native-window toggle

🤝 Contributing

Contributions welcome.

git clone https://github.com/ofirbartal100/sciviewer
cd sciviewer
npm install
npm run compile
# Press F5 in VS Code to launch a development instance

1. Fork the repo
2. Create a feature branch
3. Make your changes
4. Open a pull request

📄 License

MIT — see LICENSE.

Built for researchers and data scientists who live in remote Python environments.