NPZSee
A fast native viewer for multidimensional NumPy arrays.
简体中文 · Downloads · Platform notes
Open .npy and .npz files, navigate dimensions, inspect exact values, and plot
curves and distributions. NPZSee is written in Rust with egui. Standalone
downloads need no Python, NumPy, Qt, or Matplotlib.
Install and run
Download the standalone archive matching your OS and CPU from GitHub Releases, then extract it:
- Windows: double-click
npzsee.exe. - macOS 13+: open
NPZSee.app. Choosearm64for Apple Silicon andx86_64for Intel. - Linux: run
./npzseein a graphical desktop session. A glibc 2.28+ system, X11 or Wayland, and working OpenGL drivers are required.
Pass a file path with npzsee "path/to/arrays.npz", or drop a file on the window.
Downloads are currently unsigned release candidates.
pip
Python 3.10+ can install the matching precompiled wheel. The package has not yet been published to PyPI. For a downloaded wheel:
python -m pip install ./npzsee-0.1.0-py3-none-<your-platform>.whl
npzsee
npzsee "path/to/arrays.npz"
After the first PyPI publication, installation will be python -m pip install npzsee. The wheel installs the same native executable as the standalone
download, with no Python runtime dependencies. Without a matching wheel,
building the source distribution requires Rust and desktop development
libraries. See publishing.
Features
- Array browser: list and filter NPZ entries immediately; load arrays only when selected. Loading and analysis run in a cancellable background worker.
- Multidimensional navigation: descend through subarrays to a matrix or vector and return with breadcrumbs. C and Fortran memory orders both work.
- Table first: inspect exact numeric values, Unicode strings, scalars and JSON trees. Click a cell or JSON value to open its docked detail pane, read the full content and copy it without truncation. Long content does not open hover popups.
- Large tables: virtual rows and row/column pages keep every value reachable.
- Plot: vector, row and two-row/two-column X/Y views; complex components; zoom, pan, log axes, original-sample readouts and SVG export.
- Histogram: filled distributions, configurable bins, non-finite counts and summary statistics. A labelled sampled preview avoids scanning everything; exact statistics scan the entire selected slice when requested.
- Six interface languages: English, 简体中文, Français, Español, Русский, العربية. Change language in the toolbar; the choice is remembered. Arabic uses bidirectional layout and native contextual glyph shaping. Fonts are bundled under the SIL Open Font License.
- Read only: source files are never modified. Object/pickle arrays are not executed; store dictionary metadata as Unicode JSON instead.
Performance
NPY arrays are mapped read only; selected compressed NPZ members are decoded to temporary mapped files. A bounded cache helps with repeated selections. Plot overviews use min/max envelopes sized to the viewport, preserving spikes while drawing thousands of points instead of millions. Zoomed index plots revisit original samples.
These design choices matter as well as native execution: Rust does not make ZIP decompression inherently faster than NumPy. Prior local measurements on a 12.5-million-element float64 vector were approximately 240 ms to decode/map, 130 ms to build the envelope and below 1 ms to produce overview geometry. These are warm-cache component timings, not startup or cross-platform promises.
Supported data
NPY format versions 1/2/3, common integers, float16/32/64, complex64/128, booleans, Unicode and byte strings, and either byte order are supported. Structured fields have text previews. Unsupported fixed-width types show raw bytes; object arrays require conversion. X/Y overview plots connect samples in array order; use Table to inspect individual exact values.
Build and validate
Rust 1.96+ is required. Windows needs MSVC build tools; Linux needs desktop development libraries; macOS needs Xcode Command Line Tools.
cargo build --release --locked
cargo test --locked --all-features
cargo clippy --locked --all-targets --all-features -- -D warnings
cargo fmt --check
The executable is in target/release/. For the independent NumPy compatibility
oracle, install NumPy, build with --features bench, and run
python native/validate_numpy.py. The release workflow builds wheels and
standalone archives for Windows, Linux and macOS, each on x86_64 and ARM64.
Actual checks and limitations are in release status.
Language strings live in locales/. Translations can be improved independently
of the parser. Settings use the normal per-user configuration directory;
NPZSEE_CONFIG_DIR overrides that directory. --language en overrides startup
language without changing the saved preference.
Credits and license
NPZSee is derived from and inspired by LiukDiihMieu/npyquick. Thank you to the original project for its NumPy viewing workflow. NPZSee rewrites data loading and the desktop interface in Rust and adds hierarchical navigation, background analysis, large-array rendering, localization and native packaging. It is published under its own name while retaining upstream attribution.
Copyright © 2026 LiukDiihMieu and NPZSee contributors. Licensed under GNU GPL version 3 or later; see LICENSE, NOTICE, and the retained third-party notices in LICENSES. Matching source archives including locked dependency sources accompany release binaries.
Metadata
Release files for npzsee 0.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| npzsee-0.1.0.tar.gz | 14.4 MB | With provenance | Details for npzsee-0.1.0.tar.gz |
Built distributions (wheels)
| File | Actions Reset | |||
|---|---|---|---|---|
| npzsee-0.1.0-py3-none-win_arm64.whl | Python 3 | none | Windows ARM64 | Details for npzsee-0.1.0-py3-none-win_arm64.whl |
| npzsee-0.1.0-py3-none-win_amd64.whl | Python 3 | none | Windows x86-64 | Details for npzsee-0.1.0-py3-none-win_amd64.whl |
| npzsee-0.1.0-py3-none-manylinux_2_28_x86_64.whl | Python 3 | none | Linux glibc 2.28+ x86-64 | Details for npzsee-0.1.0-py3-none-manylinux_2_28_x86_64.whl |
| npzsee-0.1.0-py3-none-manylinux_2_28_aarch64.whl | Python 3 | none | Linux glibc 2.28+ ARM64 | Details for npzsee-0.1.0-py3-none-manylinux_2_28_aarch64.whl |
| npzsee-0.1.0-py3-none-macosx_13_0_x86_64.whl | Python 3 | none | macOS 13.0+ x86-64 | Details for npzsee-0.1.0-py3-none-macosx_13_0_x86_64.whl |
| npzsee-0.1.0-py3-none-macosx_13_0_arm64.whl | Python 3 | none | macOS 13.0+ ARM64 | Details for npzsee-0.1.0-py3-none-macosx_13_0_arm64.whl |
Total release size: 124.5 MB
Release files / npzsee-0.1.0.tar.gz
| Download URL | npzsee-0.1.0.tar.gz |
|---|---|
| Size | 14.4 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
892c54b53973fe462c9eea42c9a2575f3668c520373c88c2b503afc44e2d549e
|
|
BLAKE2b-256 checksum How to use checksums |
8023e4b5e49a67db621e2545b43268105fa32c6e7b54d423082812dd8ca321fc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / npzsee-0.1.0-py3-none-win_arm64.whl
| Download URL | npzsee-0.1.0-py3-none-win_arm64.whl |
|---|---|
| Size | 17.8 MB |
| Tags | Python 3 Windows ARM64 |
|
SHA-256 checksum How to use checksums |
1814097f84522fb24764fd5d9e4f9d39c82ad5d72477e891f55ce49ecea49c75
|
|
BLAKE2b-256 checksum How to use checksums |
8f2d69e25e0fe7b3b5a160dedd9ef2dd458a1b56fc095279c06721f06d13207a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / npzsee-0.1.0-py3-none-win_amd64.whl
| Download URL | npzsee-0.1.0-py3-none-win_amd64.whl |
|---|---|
| Size | 18.1 MB |
| Tags | Python 3 Windows x86-64 |
|
SHA-256 checksum How to use checksums |
244c3844a1bc6394ab42d5e20d835ec00d1d360dd184962160d4e117a0e12543
|
|
BLAKE2b-256 checksum How to use checksums |
83469029f86c4ea7fcfb6a11aa88f6ef8e22198489306abe2371e0f55256ea1c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / npzsee-0.1.0-py3-none-manylinux_2_28_x86_64.whl
| Download URL | npzsee-0.1.0-py3-none-manylinux_2_28_x86_64.whl |
|---|---|
| Size | 19.1 MB |
| Tags | Linux glibc 2.28+ x86-64 Python 3 |
|
SHA-256 checksum How to use checksums |
85245e826b9c052f84b5bbf1a8544d53cd93406d52766394da18493ecaf55238
|
|
BLAKE2b-256 checksum How to use checksums |
2966a2d03f41a6d4deacef07c9c23a0f24222eedeb564448e8dd8053457973e2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / npzsee-0.1.0-py3-none-manylinux_2_28_aarch64.whl
| Download URL | npzsee-0.1.0-py3-none-manylinux_2_28_aarch64.whl |
|---|---|
| Size | 18.8 MB |
| Tags | Linux glibc 2.28+ ARM64 Python 3 |
|
SHA-256 checksum How to use checksums |
f53c27f02acb83ebef11c43977cc3418a306657118d3c22024b258768394fa0e
|
|
BLAKE2b-256 checksum How to use checksums |
af19c7d003f67751d941b659bb63240cfb83587d31ed8d16377647de63a1a731
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / npzsee-0.1.0-py3-none-macosx_13_0_x86_64.whl
| Download URL | npzsee-0.1.0-py3-none-macosx_13_0_x86_64.whl |
|---|---|
| Size | 18.2 MB |
| Tags | Python 3 macOS 13.0+ x86-64 |
|
SHA-256 checksum How to use checksums |
a3b5d825530920d4fb3ca28a3dd220ff59253449be44a08de50367be4f6a2946
|
|
BLAKE2b-256 checksum How to use checksums |
526a683d81174b6776332acd2fe1e638deb4b495fd446be091945bfc8ce0c710
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logRelease files / npzsee-0.1.0-py3-none-macosx_13_0_arm64.whl
| Download URL | npzsee-0.1.0-py3-none-macosx_13_0_arm64.whl |
|---|---|
| Size | 18.0 MB |
| Tags | Python 3 macOS 13.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
9ff55dfa6057cc3652e43527834439304fbe7cdf2f5ecc00cb5ec8613792eb74
|
|
BLAKE2b-256 checksum How to use checksums |
50b3c1553aa0f7d1d2d69dfa31d5bed76bed7bf354c4baefe38e05c210a6d8ce
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.16
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logSecurity for npzsee 0.1.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 7 files in this release have been uploaded with provenance metadata.
Attestation for npzsee-0.1.0.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 24, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.