Skip to content

Native Thermo OMNIC reader

SpectraSherpa's read-only Thermo Nicolet OMNIC parser implements a bounded, fail-closed grammar for independently qualified SPA single spectra, SPG compatible spectrum groups, and legacy SRS time series. Thermo's formats are proprietary; the public block-table and series-marker interpretation was cross-checked against:

  • spectrochempy-omnic 0.2.1, commit 2eb6b7d3964451d35eeb0c185cb99a0cc147c7cd, Copyright © 2025 LCS — Laboratoire Catalyse et Spectrochimie, Caen, France, CeCILL-B.

SpectraSherpa does not import or execute that project and does not bundle its source or test data. The native parser is an independent implementation over SpectraSherpa's immutable snapshot, resource-limit, typed-asset, provenance, and fail-closed format registry authorities. Exact-hash files from an external SpectroChemPy data installation may be used by the test suite as local conformance references; those proprietary instrument files are not copied into SpectraSherpa's source distribution or wheel.

The parser preserves the feature-axis order declared in the file. For SPG, sample rows are stably ordered by their explicit acquisition timestamp and the original directory index is retained as sample metadata. SRS data-point axes remain data-point axes; SpectraSherpa does not invent an optical-path conversion or silently reverse a spectrum.

One redistributable OMNIC file is bundled in-tree and always exercised: tests/fixtures/omnic/openspecy-polyethylene-reflectance.spa, published by the Open Specy project under CC-BY-4.0 (see open-specy.md). It covers the SPA single-spectrum variant with reflectance in percent, and SpectroChemPy 0.8.1 independently decodes identical shape, axis, and values from it.

SPG group and SRS series conformance still runs on Thermo files SpectraSherpa is not licensed to redistribute, so those cases execute only where that corpus is installed. Closing that gap requires instrument files SpectraSherpa owns; it is tracked rather than papered over, because a group or series reader qualified only on synthesized bytes is qualified against its own assumptions.

The qualified variants and exact expected-science checks are recorded in docs/evidence/native-omnic-reader-conformance.json.