ensaioFetch versioned Fatiando geoscience datasets with Ensaio and preserve their hashes, citations and data conventions. Use for reproducible sample-data acquisition and replacing supported RockHound datasets; the fetchers return paths for format-specific readers.
Independent GitHub snapshot · Automated classification · Original documentation governs
Retrieve an explicit data version, then read it with a suitable format library. A package version and a dataset version identify different things.
Select a fetcher from the installed API, review the original data license and
citation, and estimate download/storage needs. ENSAIO_DATA_DIR selects a project
cache. Preserve original downloaded bytes; write derived products elsewhere.
The underlying Pooch registry verifies the expected content hash.
# example: alps-gps
from pathlib import Path
import hashlib
import ensaio
import pandas as pd
source_path = Path(ensaio.fetch_alps_gps(version=1))
source_sha256 = hashlib.sha256(source_path.read_bytes()).hexdigest()
gps = pd.read_csv(source_path)
This returns a compressed CSV path, not a DataFrame. A first fetch requires network access unless the verified bytes are already cached. A network failure is not a validated empty dataset, and disabling hash verification is not a fix.
Inspect column units, coordinate/velocity reference frames, epochs and uncertainty definitions from the original source. For this dataset, the published solution uses IGb08/ITRF2008 and the GRS80 ellipsoid, with horizontal velocities relative to Eurasia. Ensaio's short fetcher description says WGS84; do not use that shorthand to override the detailed source provenance or perform an unrecorded conversion.
Check station identifiers and missing values; keep velocity components and their standard deviations paired. Position uncertainty and velocity uncertainty are not interchangeable. Resampling or gridding must carry the reference frame and does not improve the observation accuracy automatically.
Record dataset version, source hash, package version, retrieval date and citation. See the fetcher documentation, original data and redistributed v1 release.