NumPy Array (.npy / .npz)
.npy.npzNumPy (np.save / np.savez) — the standard array-serialization format across the Python scientific stack
What it is
NumPy's own on-disk array format. A .npy file holds exactly one array; a .npz file is a ZIP of several named .npy arrays (what np.savez writes). AltaiPlot reads both through a bundled Python/NumPy bridge.
Who produces it
Anyone working with matplotlib, pandas, or raw NumPy in Python who saved intermediate results with np.save/np.savez instead of exporting to CSV.
What we read / what we don't read
What we read
- 1D numeric arrays (float, int, uint, bool) → a single column
- 2D numeric arrays → one column per array column
- .npz archives with multiple named arrays — pick which one(s) to import
What we don't read
- Pickled/object-dtype arrays — refused outright (allow_pickle=False is never overridden, since loading a pickled array can execute arbitrary code embedded in the file)
- Arrays with more than 2 dimensions
How to import it into AltaiPlot
- Open AltaiPlot and drag your .npy/.npz file onto the window, or use File → Import.
- AltaiPlot detects the NumPy magic bytes and hands it to the bundled Python reader automatically.
- For a .npz archive, every array name/shape/dtype is listed — pick the one you need (a plain .npy always has exactly one).
- The array lands as plotting-ready column(s); pickled or higher-dimensional arrays are rejected with an explicit message, never silently skipped or misread.
Auditability note: this page's capability claims are sourced fromresources/python/app_sidecar.py:1032 — see the full format list for every extension AltaiPlot recognizes.