Chem Db Spectra

chem-db-spectra

Goal

To query experimental molecular spectra—such as Infrared (IR), Mass Spectrometry, and UV-Vis—from standard experimental databases like the NIST Chemistry WebBook for baseline comparison against properties computed via MLIPs or DFT.

Instructions

Step 1. Query Spectra Data

Use the query_spectra.py script to fetch available spectrographic data for a given chemical formula. The script automatically searches the NIST Chemistry WebBook, resolves the compound, and downloads all available JCAMP-DX (.jdx) formatted spectra into the specified directory.

# Env: base-agent
python .agents/skills/chem-db-spectra/scripts/query_spectra.py   [--type {IR,Mass,UVVis,All}]

Parameters:

• formula: The chemical formula to query (e.g., CH4, C2H6O).
• output_dir: The directory path to save the downloaded JCAMP-DX (.jdx) spectrum files.
• --type: Filter the spectrum type to download (options: IR, Mass, UVVis, All). Default is All.

Constraints

• Environments: The scripts require the base-agent Conda environment.
• Spectrum Format: Output data is universally saved in JCAMP-DX (.jdx) standard format. You must use appropriate spectroscopic python package (e.g., jcamp) if you need to parse it back into numpy arrays.
• Scraping Dependability: This script utilizes HTML scraping (beautifulsoup4). If the NIST WebBook layout changes in the future, the selectors may need to be updated.

References

• Linstrom, P.J. and Mallard, W.G., Eds., NIST Chemistry WebBook, NIST Standard Reference Database Number 69, National Institute of Standards and Technology, Gaithersburg MD. URL
• JCAMP-DX for Infrared, Raman, and Related Data. URL

---

Author: Bowen Deng Contact: GitHub @learningmatter-mit

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: learningmatter-mit
• Source: learningmatter-mit/AtomisticSkills
• License: MIT
• Homepage: https://arxiv.org/abs/2605.24002

Install and usage instructions live in the source repository linked above.