astro-legacy-archive/swas-fits-spectra
SWAS FITS Spectra The SWAS Tenth Public Release pointing-based IRAF-FITS Level 1.5 product serves 6,962 containers containing four or six named spectral image members. Use from datasets import load_dataset ds = load_dataset("astro-legacy-archive/swas-fits-spectra", "14498-5856_0001__SWAS_IMEXT001_B1", split="train") row = ds[0] values = row[ds.column_names[0]] print(len(values), values[:3]) 725 [0.00926896184682846, 0.007523952983319759, 0.001602768199518323]… See the full description on the dataset page: https://huggingface.co/datasets/astro-legacy-archive/swas-fits-spectra.
SWAS FITS Spectra
The SWAS Tenth Public Release pointing-based IRAF-FITS Level 1.5 product serves 6,962 containers containing four or six named spectral image members.
Use
from datasets import load_dataset
ds = load_dataset("astro-legacy-archive/swas-fits-spectra", "14498-5856_0001__SWAS_IMEXT001_B1", split="train")
row = ds[0]
values = row[ds.column_names[0]]
print(len(values), values[:3])725 [0.00926896184682846, 0.007523952983319759, 0.001602768199518323]The first configuration is 14498-5856_0001__SWAS_IMEXT001_B1. FITS configuration names combine the source filename stem with source EXTNAME and, for repeated EXTNAME values, source LINE. The complete configuration inventory is in the standalone .huggingface.yaml metadata.
import json, pyarrow.parquet as pq
from huggingface_hub import hf_hub_download
path = hf_hub_download("astro-legacy-archive/swas-fits-spectra", "d/00/000.parquet", repo_type="dataset")
metadata = json.loads(pq.read_metadata(path).metadata[b"spectrum"])
print(metadata["schema"], metadata["schema_version"], metadata["representation"])astro-legacy-archive/spectrum 1.1.0 fits-imageInstall the baseline readers with python -m venv .venv && .venv/bin/pip install datasets pyarrow huggingface_hub.
Records and fields
Each FITS Parquet table has one row and one source-named fixed-size vector of 725 float32 values in K. The FITS EXTNAME is the field name. LINE distinguishes repeated EXTNAMEs and retains CI, O2, H2O, 13CO, OTHER, or H2(18)O. The linear frequency-axis declarations remain in the member header in GHz.
There are 27,890 configurations and 27,890 Parquet rows. Source row order is retained. Units, dtypes, source declarations and logical FITS headers, source URL, served byte size and SHA-256 are attached to each artifact. Axes declared only in source metadata remain there.
The index names 6,963 FITS containers. It serves 6,962; NGC6334D_0008.fit returns HTTP 404 and is recorded, not reconstructed or replaced. The data-product overview identifies this pointing-based FITS archive and the CLASS archive as the two final community-release products, and warns that continuum offsets do not necessarily reflect the true continuum flux. The tenth-release table omits FITS-format grand co-adds for six comets and four planets because SWAS says a grand co-add for moving or time-variable targets would be misleading; their CLASS collections remain in astro-legacy-archive/swas-class-spectra. The mission page identifies the online data as the complete tenth release and the fifth through ninth releases as superseded, request-only subsets.
Source, rights, and rebuilding
Canonical FITS data index: https://lambda.gsfc.nasa.gov/product/swas/fits_data.html
No explicit data licence is published on the LAMBDA SWAS product pages.
The Smithsonian Astrophysical Observatory created the two final SWAS community-release products; LAMBDA serves this Submillimeter Wave Astronomy Satellite Tenth Public Release. No product-specific acknowledgement wording is published on the pinned SWAS product pages.
Rebuild with python tools/make_swas_fits_spectra.py --out data --source <source-directory> --provenance-commit <commit>.
Errata and requests
Please report corrections through the Hugging Face Community tab. A free Hugging Face account is enough to participate.
