Getting started

Install the CLI

macOS — Homebrew (recommended)

brew tap Latka-Industries/tetration https://github.com/Latka-Industries/tetration
brew install tetration
tet --help

cargo install

cargo install tetration

Default features need system HDF5 and NetCDF dev libraries for tet convert. Without them:

cargo install tetration --no-default-features

tet info, tet query, and Zarr import still work.

From source

git clone https://github.com/Latka-Industries/tetration.git
cd tetration
cargo build --release
export PATH="$PWD/target/release:$PATH"

Convert a file

tet convert volume.h5 volume.tet     # HDF5
tet convert data.nc data.tet         # NetCDF
tet convert zarr_store/ output.tet   # Zarr v3 directory

Inspect a file

tet info volume.tet
tet info volume.tet --metadata       # axis names and coord labels
tet verify volume.tet                # health check

Run a query

Query documents are flat JSON or TOML — same wire shape, either format:

mean.json

{ "dataset": "temperature", "mean": [] }

mean.toml

dataset = "temperature"
mean = []

# Inline JSON
tet query '{"dataset":"temperature","mean":[]}' -t volume.tet -x -q

# TOML file
tet query mean.toml -t volume.tet -x -q

# Table preview
tet query mean.toml -t volume.tet -x --format table --preview 6

Rust library

[dependencies]
tetration = "0.1.9"

use tetration::prelude::*;
// TetWriterSession, TetFile, execute_query_json, …

See Rust quick start.

Python (tet-py)

pip install tet-py

import tet

with tet.open("volume.tet") as f:
    print(tet.__version__, tet.core_version())
    print(f.mean("temperature"))

    # Optional capped preview while querying (CLI: --preview N)
    r = f.mean("temperature", preview=32)
    r.scalar, r.preview.shape

    # Dense export — three sinks (same as the query engine)
    arr = f.read_numpy("temperature")                              # ram
    z = f.transform.to_numpy.zscore("temperature")                 # transform → ram
    spill = f.transform.to_spill.zscore("temperature", path="z.bin")
    spill.to_numpy()
    side = f.transform.to_sidecar.zscore("temperature", path="z.tet")  # sidecar → .tet
    side.to_numpy(f)

See Python overview, quick start, and NumPy interchange (ram / spill / sidecar).

Next steps

• Format spec — .tet on-disk layout
• CLI reference — all tet commands
• Query engine — what queries can do
• Query cookbook — copy-paste examples
• Python (tet-py) — PyPI install, query API, NumPy
• Guides — format comparisons, mmap patterns, catalog integration