Add metabolomics test data for metaproviz modules

[giuliaroeth]

Summary

Adds test data for the upcoming metaproviz/* modules, under a new top-level metabolomics/folder.

Why a new metabolomics/ folder

Discussed in nf-core Slack #modules, since no metabolomics test data existed yet. Following the same reasoning genomics uses (organize by data type, not by tool/software, for discoverability and reusability), we settled on metabolomics/raw_annotation_data/ rather than a tool-specific delete_me/metaproviz/ folder, since this data format (a feature-by-sample data matrix, sample metadata, and feature metadata) is a standard shape in metabolomics generally, not something specific to our modules.

Structure: metabolomics/raw_annotation_data/<cellline|patient>/<dataset>/, each with a data_matrix.csv, feature_matrix.csv, and sample_matrix.csv (the standard TSV/CSV shape), plus an SE.RData file (see below for why).

Future additions can follow the same data-type-based convention, e.g. metabolomics/raw_spectral_data/ or metabolomics/biocrates_data/ for other metabolomics data shapes.

Why both SE.RData and CSV files for the same data

Our modules accept two mutually exclusive input shapes: a Bioconductor SummarizedExperiment (.rds), or the three flat files directly (data_matrix/feature_matrix/sample_matrix). Both are real, separately tested code paths, so we need real test data in both shapes to cover them. We considered only bundling the SummarizedExperiment object, but that's
R-specific, so we kept the plain CSV/TSV files as the primary, standard, language-independent format, and added the .rds alongside it specifically to also exercise our modules' SummarizedExperiment input path in testing.

PR checklist

• We read the test-data specifications
• We tried testing our module with available data and there is no compatible data in test-datasets
• If we added data: we described the new files on the root README.md
• If we modified data: we checked whether the root README.md is still up-to-date (n/a, we only added new data, didn't modify any existing files)
• We made sure to submit the smallest dataset possible