Published June 4, 2026 | Version 1.1.0

EPS Astro Extractor v1.1.0

  • 1. EPS Research

Description

 

Part of the EPS Research multi‑epoch kinematic platform: Full datasets, cross‑epoch corpora, executable examples, and the Astro‑Extractor are available at https://github.com/eps-research/rag-corpus-series

EPS Astro Extractor is a portable Streamlit application for extracting structured data from astronomical databases using a local large language model via LM Studio. No API key required. Runs entirely on local hardware.

The tool crawls any publicly accessible URL, feeds the content to a local LLM, and returns structured JSON with parameters, table data, and export options.

v1.1.0 adds:

  • Windows one-click launcher (run_extractor.bat) — double-click to install and launch, no command line required
  • EPS Corpora section in the sidebar with direct links to all five RAG corpora (v7/SPARC, Dwarf, GC, Z1, IntZ)
  • Auto-extract mode — leave the target blank to extract all structured data from the page
  • Suitable for high school students on Windows with no prior setup knowledge

Eleven quick targets are pre-configured across three categories: astronomical databases (SPARC, VizieR/CDS, NED, HyperLeda), HI survey catalogues (THINGS, LITTLE THINGS, WHISP, LVHIS), and EPS RAG Corpora (v7/SPARC, Dwarf, GC, Z1, IntZ).

Developed in support of the EPS Research RAG Corpus Series (github.com/eps-research/rag-corpus-series).

Full Platform: This record is part of the EPS Research multi‑epoch kinematic platform (SPARC + THINGS + LITTLE THINGS + WALLABY DR2 + Milky Way GC corpus + z≥6). All corpora, executable notebooks, the Astro‑Extractor, and the unified schema are available at: https://github.com/eps-research/rag-corpus-series

Files

eps_astro_extractor_banner.png

Files (74.5 kB)

Name Size Download all
md5:49a584def450e5de97e8f3f7f8ea9c47
63.3 kB Preview Download
md5:ff619cd5d841984275d60dc3f12eca3b
11.2 kB Preview Download

Additional details

Related works

Is part of
Dataset: 10.5281/zenodo.20398430 (DOI)
Is supplement to
Journal: 10.3389/fspas.2025.1680387 (DOI)
Preprint: 10.5281/zenodo.19563417 (DOI)