scripts/seed_commons_from_dms.py

#!/usr/bin/env python3
"""Seed the Field Atlas / commons from published DMS studies (ESM-free).

Turns a set of public DMS assays into de-identified substitution-effect rows
(dee.core.dms_seed → the same k-anonymized aggregation user data goes through)
so the commons has real, citable value on day one. No model needed — this only
pools measured values by substitution TYPE, keeping a substitution only when
≥ MIN_USERS independent studies measured it.

Input: a manifest JSON, a list of assays:
    [ {"name": "...", "csv": "path/to/dms.csv"}, ... ]
(ProteinGym-style CSVs: a 'mutant' column + a 'DMS_score' column.)

By default it WRITES the rows to a JSON file for review. Pass --push to upload
them to public.mutation_priors via dee.auth (requires SUPABASE creds in the
environment — run it where the service key lives, e.g. the deploy box).

Usage:
    python scripts/seed_commons_from_dms.py manifest.json [--out seed_rows.json] [--push]
"""
import argparse
import json
import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent.parent))

from dee.core.dms_seed import parse_proteingym_csv, seed_rows  # noqa: E402


def main():
    ap = argparse.ArgumentParser()
    ap.add_argument("manifest")
    ap.add_argument("--out", default="seed_rows.json")
    ap.add_argument("--push", action="store_true",
                    help="upload to public.mutation_priors (needs SUPABASE env)")
    args = ap.parse_args()

    assays = []
    for a in json.loads(Path(args.manifest).read_text(encoding="utf-8")):
        recs = parse_proteingym_csv(Path(a["csv"]).read_text(encoding="utf-8"))
        if recs:
            assays.append((a.get("name", a["csv"]), recs))
        print(f"  {a.get('name', a['csv']):28s} {len(recs):7d} records")

    rows = seed_rows(assays)   # enforces the effective-date gate + k-anonymity
    Path(args.out).write_text(json.dumps(rows, indent=2), encoding="utf-8")
    print(f"\n{len(rows)} de-identified substitution row(s) (>= MIN_USERS studies each) -> {args.out}")

    if args.push:
        from dee import auth
        result = auth.replace_mutation_priors(rows)
        print(f"push -> {result}")
    else:
        print("(dry run — pass --push to upload to the live commons)")


if __name__ == "__main__":
    main()