SciViewer

View scientific data files directly in VS Code.

Supported Formats

• HDF5 (.h5, .hdf5) - Hierarchical Data Format
• Pickle (.pkl, .pickle, .p) - Python serialized objects
• Parquet (.parquet, .pqt) - Columnar data format
• NumPy (.npy) - NumPy array files

Features

• 📁 Tree view of file structure with expand/collapse
• 📊 View dataset metadata (shape, dtype, attributes)
• 👁️ Preview dataset values
• 🎨 Integrates with VS Code light/dark themes
• ↻ Refresh button to reload changed files

Requirements

• Python 3.x with appropriate packages:

pip install h5py numpy pandas pyarrow

• h5py - required for HDF5 files
• numpy, pandas - recommended for full pickle support
• pyarrow - required for parquet files

Installation

1. Install from VS Code Marketplace (search "SciViewer")
2. Ensure Python is available in your PATH
3. (Optional) Configure sciViewer.pythonPath if Python is not auto-detected

Usage

Simply open any supported file in VS Code. The custom viewer will automatically display the file structure.

• Click on groups (📁) to expand/collapse
• Click on datasets (📊) to view details and data preview

Pickle Support

Pickle files display Python objects as a tree:

• dict → expandable group
• list, tuple, set → expandable group with indexed children
• numpy.ndarray → dataset with shape, dtype, preview
• pandas.DataFrame → dataset with shape, preview of first rows
• Scalars (int, float, str, etc.) → dataset with value preview
• Custom objects → dataset with str() representation

Parquet Support

Parquet files display a tabular view with:

• Summary - row count, column count, compression, file size
• Schema - column names and data types
• Preview - first 10 rows as a table

NumPy Array Support

NumPy .npy files display:

• Shape - array dimensions
• Dtype - data type
• Preview - first 20 elements (flattened)

Extension Settings

Troubleshooting

"No module named 'h5py'"

• Install h5py: pip install h5py
• Or set sciViewer.pythonPath to a Python environment that has h5py

"Failed to spawn Python"

• Ensure Python is installed and in your PATH
• Or set sciViewer.pythonPath to the full path of your Python executable

License

MIT