mdPCA: Convert indID from labels DataFrame to string

[miriambt]

Concept

A bug was identified in mdPCA when processing the labels file, leading to the error below:

ValueError: Found array with 0 sample(s) (shape=(0, 0)) while a minimum of 1 is required by TruncatedSVD.

Issue Description

The error occurs because the indID column in the labels file might contain integer values (e.g., 403). However, the identifiers in laiobj are stored as strings. When filtering with:

labels = np.array(labels_df['label'][labels_df['indID'].isin(temp_ind_IDs)])

the mismatch between integer and string types means that no samples match, resulting in an empty array.

Resolution

To fix this, we ensure that the indID values in the labels file are treated as strings. We convert the indID column immediately after loading the file:

labels_df['indID'] = labels_df['indID'].astype(str)