[Feature Request] - Add option to build a Leviathan index from a Sylph database

[jolespin]

Here's what a Sylph database looks like:

2026-07-09T20:31:17.369Z INFO  [sylph::inspect] Database file genomes.syldb processed with 2029 genomes
- database_file: genomes.syldb
  c: 200
  k: 31
  min_spacing_parameter: 30
  genome_files:
  - file_name: ../Datasets/CAMI-I/CAMI_high_toy/H_S003/genomes/H_S003__939857.1.fa.gz
    genome_kmers_num: 4181
    first_contig_name: H|S3|C877963
    genome_size: 979114
  - file_name: ../Datasets/CAMI-I/CAMI_high_toy/H_S004/genomes/H_S004__1262820.1.fa.gz

Here's what a Leviathan database looks like:

data = pd.read_pickle("genome_to_data.pkl.gz")
data["H_S001__1077974.1"]
{'id_genome_cluster': 'PSLC-1b2cdae03ed629b53d94da147b3ac9ab',
 'filepath': '../Datasets/CAMI-I/CAMI_high_toy/H_S001/genomes/H_S001__1077974.1.fa.gz'}

Idea is that a user can do the following:

• Download a big Sylph database (e.g., GlobDB)
• Use the cds.fasta.gz with only relevant hits
• --genomes will be mutually exclusive with --precomputed_sylph_database
• leviathan-index.py will need to make sure that all the genomes in the --feature_mapping are present in the --precomputed_sylph_database and vice versa (or perhaps a subset in case there are genomes that don't have features annotated but the user still wants to perform taxonomic profiling)