Skip to content

NOMAD integration

pynxtools-raman doesn't implement its own NOMAD parser or schema — that's handled generically by pynxtools for every NeXus application definition it knows about. What pynxtools-raman contributes on top is a search app tailored to Raman data. This page explains how the pieces fit together; for how to actually use the app, see How-to > Search Raman data in NOMAD.

How a converted file becomes searchable

Once NOMAD has pynxtools-raman installed as a plugin, any .nxs file uploaded to it is picked up by pynxtools's own NexusParser, which recognizes NeXus/HDF5 files generically — it doesn't need to know about NXraman specifically. The parser walks the HDF5 tree and populates NOMAD's metainfo structure from it, following the NXraman application definition to resolve group and field types.

Two things happen automatically during this:

  • Material normalization. NOMAD's own normalizer looks for chemical formula and atom-type information on any NXsample-typed group in the entry and populates results.material (elements, Hill/IUPAC/reduced formula, ...) from it. This is what drives the periodic-table filter in the Raman app and is independent of anything pynxtools-raman itself does — it's the same mechanism NOMAD uses for every NeXus-based plugin.
  • Entry metadata. The entry's NXentry/definition field (NXraman) is used to label the entry's type, and file-level references (such as identifier_experiment, mapped from a publication DOI for Raman Open Database records) become part of the searchable entry metadata.

The Raman app

pynxtools-raman registers one NOMAD app entry point, raman_app (src/pynxtools_raman/nomad/apps/__init__.py), pointing at the search quantities exposed by pynxtools's generic NeXus metainfo schema. The app declares which columns to show, which filters to expose in the side menu, and what the default dashboard looks like. See Reference > Raman NOMAD app for more information.

Because the app is built against pynxtools's shared NeXus schema rather than a Raman-specific one, its search-quantity paths follow the schema's own naming convention rather than something Raman-specific. If that schema changes upstream in pynxtools, the app's paths need to be updated to match — there's no independent Raman schema layer in between.