Adding meta data to the spectral container?

[SvenJachalke]

Hi @dgeorgiev21

One question (or maybe its already a feature request :D):

Is it planned to include also some meta data to your spectral containers? I don't know what the *.mat files provide, but from my side it would be very nice for example to know the measurement type (as you know the WITec tools provide things like area scan, single spectrum, line scan, etc.) in advance. Right now I'm getting the information from the shape, e.g. 1x1 = single spectrum, m x n = area scan ... a 1 x m can be also line scan but also a time series.

Also some "spatial data", e.g. from which part of the sample the data was recorded, would be great. In the WITec software one can set at which coordinates (+step size, etc.) a scan is performed. Having that in the container would be also valuable.

Are such informations stored in the *.mat files? Or is this unique to the *.wip files?
Maybe you can comment on that?

Thanks!
Sven

[dgeorgiev21]

Hi @SvenJachalke

Thank you for the question and your engagement with the development of RamanSPy – it’s really helpful!

Currently, the spectral container does not store explicit metadata about measurement type or spatial acquisition details. As you noted, this information can sometimes be inferred from the data shape, but it’s not formally preserved in the container structure.

Adding metadata is something we’ve been considering for a while. The main challenge is balancing usefulness with simplicity – we want to focus on fields that are truly important to users. Some of the fields we have been thinking about include:

Data format–related:

• Date created
• Author / operator
• Modality (e.g. single point, line scan, imaging, volumetric)
• Step sizes (for line, image or volumetric scans)

These could be implemented in the __init__ method of the SpectralObject class and populated via the load_* methods. Challenges include the lack of consistent example data from all instruments and evolving data formats.

Preprocessing history:

• A list of applied preprocessing steps, including:
  • Method name and parameters
  • Date/time applied

This would be easier to implement via the PreprocessingStep.apply() method.

We are curious – is there anything else you think would be particularly useful to include as metadata in your workflows?

---

If you find RamanSPy useful, please consider starring the project on GitHub! Your support means a lot!