Add genotyping-array manifest generator script

[joshuamschmidt]

Purpose

Add scripts/generate_genotyping_manifest.py. It builds the genotyping-array manifest CSV for a cohort, uploads it to gs://cpg-<project>/gtc_genotyping_array/manifests/, and registers it as a manifest analysis in Metamist. In production the transfer pipeline creates this manifest; this script does the same for hand-made cohorts, so any test cohort can be run through phase 1.

Proposed Changes

• Query the cohort's sequencing groups and build one row per SG from its meta (sentrix_barcode_a, sentrix_position_a, sample_plate). Same column layout as the production manifests.
• Check every derived GTC path exists in GCS before writing anything.
• Fail on a missing cohort, a non-genotypingarray SG, missing meta, a duplicate sentrix well, or a missing GTC.
• Refuse to overwrite an existing manifest without --force. --dry-run prints the CSV without uploading or registering.

Testing

• Dry-run against COH13479 (ourdna-test) and compared with the production manifest: identical values for every field the pipeline parser reads. Fields production left blank are filled from SG meta.
• Dry-run against COH13461 (ourdna-test): 95 rows, all GTC paths verified.

[Alexander-Stuckey]

Looks good to me