Installation¶

The CLI and the Python library ship in one PyPI package, datamanifestpy; Julia support is the separate package DataManifest.jl.

CLIPythonJulia

pipx install datamanifestpy

pip install datamanifestpy

using Pkg
Pkg.add("DataManifest")

The sections below cover each client in full.

The command-line client¶

pipx install datamanifestpy   # the `datamanifest` command, in its own environment
# or
pip install datamanifestpy

pipx keeps the command isolated from your project environments; pip installs the CLI (and the Python library) into the current environment.

Requirements: Python ≥ 3.10. The core dependencies (httpx, tqdm, tomli_w, platformdirs) install automatically. Nothing else is needed to fetch, verify, extract and track datasets from the shell — the optional extras below only matter when you load data from Python.

Python library¶

pip install datamanifestpy

This installs the library and also puts the datamanifest command on your path. Requirements: Python ≥ 3.10.

The core package fetches, verifies, extracts and tracks your datasets and caches your computed results. The data loaders (turning a file into a pandas/xarray/… object) are optional, so you only install the backends you actually use:

pip install "datamanifestpy[csv]"       # pandas CSV
pip install "datamanifestpy[parquet]"   # pandas + pyarrow
pip install "datamanifestpy[nc]"        # xarray + netcdf4
pip install "datamanifestpy[yaml]"      # pyyaml (intake / DVC import)
pip install "datamanifestpy[fsspec]"    # object-store access (s3://, gs://, …)
pip install "datamanifestpy[all]"       # all of the above

Without an extra, datamanifest still downloads, verifies and tracks data and hands you the on-disk path — load_dataset only needs the matching backend for the format you ask it to load.

Julia package¶

using Pkg
Pkg.add("DataManifest")

Requirements: Julia ≥ 1.10.

In Julia, data loaders are ordinary packages (e.g. CSV.jl) that you Pkg.add yourself and reference from the manifest — see language bindings. The Julia API reference is mirrored on this site; the Julia walkthrough lives in the DataManifest.jl repository.

Once installed, head to the quickstart.