astro-legacy-archive/abs-field-a-maps

ABS Field A maps

Regenerate the preview from the published Parquet data with python tools/render_hub_preview.py abs-field-a-maps datasets/abs-field-a-maps/preview.png.

This dataset contains the Atacama B-mode Search Field A map product served by LAMBDA. It contains the coadded Stokes Q and U maps shown in Kusaka et al. (2018), together with the source QQ, QU, and UU covariance maps.

python -m venv .venv && .venv/bin/pip install datasets huggingface_hub pyarrow

from datasets import load_dataset
dataset = load_dataset("astro-legacy-archive/abs-field-a-maps", "ABS_data_fielda", split="train[:1]", columns=["PIXLIST", "Q", "U"])
row = dataset[0]
print(row["PIXLIST"], row["Q"], row["U"])

505364 -0.0008958747555850673 -0.002862118511720226

import json
from huggingface_hub import hf_hub_download
import pyarrow.parquet as pq
path = hf_hub_download("astro-legacy-archive/abs-field-a-maps", "ABS_data_fielda.parquet", repo_type="dataset")
print(json.loads(pq.read_schema(path).metadata[b"explicit_pixel_table"])["pixelisation"]["scheme"]["nside"]["value"])

256

Dataset structure

The configuration contains 80,735 source rows and six scalar fields in source order: int32 PIXLIST, followed by float64 Q, U, QQ, QU, and UU. The source FITS table does not contain TUNIT keywords. LAMBDA documents Q and U in thermodynamic microkelvin and QQ, QU, and UU in microkelvin squared; those units and their document basis travel in the field metadata without altering the source FITS definitions.

PIXLIST is the source's explicit HEALPix pixel identity. The table header states NSIDE=256, and the LAMBDA product description identifies the file as HEALPix. A complete NSIDE-256 sky contains 786,432 pixels; the released table contains only the Field A rows. Missing full-sky pixels are therefore visible from the documented total and measured row count. No row is inserted, filled, or classified by this mirror.

The source FITS does not state an ORDERING or INDXSCHM keyword, so the metadata does not invent either statement. Without externally establishing whether the indexing is RING or NESTED, the PIXLIST values cannot be mapped unambiguously to sky positions. The metadata records only the source-backed NSIDE, HEALPix product description, explicit PIXLIST field, row count, and documented full-sky total.

Relationship to other ABS products

LAMBDA publishes Field A as its own map product. The final BB, EB, and EE power-spectrum tables are mirrored separately in abs-bandpowers. The Field A mask, spatial window, transfer function, likelihood, and 100 noise simulations are separate releases and are not table columns or configurations here. The Tau A product is also separate and consists of three 201 by 201 ZEA image HDUs, rather than this explicit-pixel table shape.

The power spectra were formed after applying the released mask and correcting for transfer and window functions. Those later operations do not modify the Q/U and covariance values mirrored here. Analyses reproducing published spectra need the separately released mask and response products.

Provenance and reconstruction

The canonical Field A download page serves ABS_data_fielda.fits, delivered 2018-12-05. The current file is 3,559,680 bytes with SHA-256 5b98c7fd2c676d6a578b48e36941f418d0e3d1d9d58416786df5325d1f9aa79e.

The embedded document contains both FITS headers, source field definitions, documented units and pixelisation bases, source URL and pin, credit, and the fixed regeneration command. Rebuild with python tools/make_abs_field_a_maps.py data/abs-field-a-maps.

No explicit licence identifier is published. Credit, verbatim: “If you use these data products please acknowledge Kusaka et al.” The cited paper is Kusaka et al. (2018), JCAP 2018(09), 005, doi:10.1088/1475-7516/2018/09/005.

Licence

HEASARC states verbatim:

HEASARC materials are all available freely for your use.

This is a free-use statement rather than a named formal licence.

Errata and requests

To report an erratum, propose a correction, or ask a question about this dataset, open a Discussion or Pull Request in its Community tab; posting requires a free Hugging Face account. Dataset requests and archive-wide questions may be posted in the README space's Community tab or the Community tab of any dataset. Confirmed corrections are made in the source conversion tools and published by rebuilding the dataset.