datasite: me@madhavajay.com
http_relay_servers:
  - syftbox.net
public_url: "syft://me@madhavajay.com/public/biovault/participants.yaml"
private_url: "syft://me@madhavajay.com/private/biovault/participants.yaml"

# Mock data anchors for testing
mock_data_grch38: &mock_data_grch38
  ref_version: GRCh38
  ref: https://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/reference/GRCh38_reference_genome/GRCh38_full_analysis_set_plus_decoy_hla.fa
  ref_index: https://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/reference/GRCh38_reference_genome/GRCh38_full_analysis_set_plus_decoy_hla.fa.fai
  aligned: https://ftp-trace.ncbi.nih.gov/1000genomes/ftp/1000G_2504_high_coverage/data/ERR3239276/NA06985.final.cram
  aligned_index: https://ftp-trace.ncbi.nih.gov/1000genomes/ftp/1000G_2504_high_coverage/data/ERR3239276/NA06985.final.cram.crai
  # BLAKE3 checksums for verification (much faster than SHA-256)
  ref_b3sum: "49cbaceaf79ebc1da6581b2f7599cb03e6552ccce87584d1a0eaec59c3629368"
  ref_index_b3sum: "002cf8e0066a2226616b5d9cc09994ac06831cd907e13e521bef6dc69403d147"
  aligned_b3sum: "4556b84f32e58e1a5c4d7238352e9fc0bcaabd2478250733252f2b76047ba3ca"
  aligned_index_b3sum: "6914d3c6842670bdde272b8cc4dfaf858a84f379e9e79d8b24c1a89d577262e2"

mock_data_snp: &mock_data_snp
  snp: https://github.com/OpenMined/biovault-data/raw/main/snp/23andme_genome_v4_Full.zip
  snp_b3sum: "2f4aeca4ff424abcc187fcecc0f3776376a43ea674a9498c9b24b4ad4964ccea"
  snp_post_process:
    uncompress: true
    file: genome_Zeeshan_Usamani_v4_Full.txt

participants:
  MADHAVA_23andme:
    id: MADHAVA_23andme
    url: "{root.private_url}#participants.MADHAVA_23andme"
    snp: "{url}.snp"
    mock: *mock_data_snp
  MADHAVA_WGS:
    id: MADHAVA_WGS
    url: "{root.private_url}#participants.MADHAVA_WGS"
    ref_version: GRCh38
    ref: "{url}.ref"
    ref_index: "{url}.ref_index"
    aligned: "{url}.aligned"
    aligned_index: "{url}.aligned_index"
    mock: *mock_data_grch38