simplify how samples are specified

Come up with an intuitive data structure for internal use, and then parse into it from the following options:
1. Specify as a samples.tsv file
2. Specify as a custom user-provided py script
3. Specify as yaml containing strings as paths with wildcards in them

You could create a parser script for this.

Start by comparing the three input forms described in the issue: samples.tsv, a custom Python script, and YAML paths with wildcards. Define the internal sample structure and the parser entry point, then confirm that each format can be converted into it and that the resulting sample data is usable by the pipeline.

Written by the indexing model from the issue text.