PGx: ClinPGx variant-based genes (PGX-008) and --guideline dpwg (PGX-025)

- fetch_sources.sh downloads ClinPGx clinicalVariants (CC BY-SA 4.0); pgx db build adds genotype-only
  variant-based genes (levels 1A-2B, level 3 for the parity list), rsIDs resolved from the manifest's probes:
  70 genes in the DB.
- pgx phenotype --guideline dpwg: CPIC mappings for CPIC genes (as ClinPGx curates DPWG), no phenotype for
  DPWG-only genes (flags dpwg_via_cpic_mapping / dpwg_mapping_not_available). K-029 updated.

tests/unit/test_pgx_db.py

@@ -156,3 +156,46 @@ def test_pharmvar_core_alleles() -> None:  # PGX-001 PharmVar source
     core = pharmvar_core_alleles("CYP2A6", {"CYP2A6-6.2/GRCh38/CYP2A6_9.vcf": vcf,
                                             "CYP2A6-6.2/GRCh38/CYP2A6_9.001.vcf": vcf})  # fmt: skip
     assert core == {"*9": [("chr19", 100, "A", "C", "rs1")]}  # sub-alleles (*9.001) are not core definitions
+
+
+def test_clinpgx_variant_genes() -> None:  # PGX-008 variant-based genes from ClinPGx
+    from simurg.pgx.db import clinpgx_variant_genes
+
+    tsv = ("variant\tgene\ttype\tlevel of evidence\tchemicals\tphenotypes\n"
+           "rs1\tOPRM1\tEfficacy\t3\tx\t\nrs2\tFOO\tEfficacy\t3\tx\t\nrs3\tFOO\tEfficacy\t2A\tx\t\n"
+           "CYP2C9*3\tCYP2C9\tPK\t1A\tx\t\nrs4\tCYP2C9\tPK\t1A\tx\t\nrs5\tA,B\tPK\t1A\tx\t\n")  # fmt: skip
+    assert clinpgx_variant_genes(tsv, {"CYP2C9"}) == {"OPRM1": {"rs1"}, "FOO": {"rs3"}}
+
+
+def test_dpwg_mode(tmp_path: Path) -> None:  # PGX-025
+    from simurg.core.log import Reporter
+    from simurg.pipeline.pgx_pheno import PhenoOptions, run_phenotype
+
+    src = tmp_path / "sources" / "cpic-20260101"
+    src.mkdir(parents=True)
+    t = _tables()
+    t["gene"][0]["symbol"] = "CYP2D6"
+    for tbl in ("sequence_location", "allele_definition", "allele", "gene_result"):
+        for r in t[tbl]:
+            r["genesymbol"] = "CYP2D6"
+    for k, v in t.items():
+        (src / f"{k}.json").write_text(json.dumps(v))
+    (src / "source.json").write_text(json.dumps({"name": "CPIC", "version": "2026-01-01", "retrieved": "20260101"}))
+    fa = tmp_path / "g.fa"
+    fa.write_text(">chr1\n" + SEQ + "\n")
+    r = CliRunner().invoke(app, ["pgx", "db", "build", "--sources", str(tmp_path / "sources"), "--fasta", str(fa),
+                                 "--out", str(tmp_path / "db"), "--quiet"])  # fmt: skip
+    assert r.exit_code == 0, r.output
+    db = next((tmp_path / "db").glob("pgx_db-1-*.sqlite"))
+    calls = tmp_path / "calls"
+    calls.mkdir()
+    (calls / "pgx_calls.json").write_text(json.dumps({"samples": [{"sample_id": "s1", "genes": [
+        {"gene": "CYP2D6", "genotype": "*2/*2", "feasible_solutions": ["*2/*2"], "candidates": [
+            {"genotype": "*2/*2", "score": 1.0, "raw_score": 1.0}]},
+        {"gene": "CYP3A4", "genotype": "*1/*22", "feasible_solutions": ["*1/*22"], "candidates": []}]}]}))  # fmt: skip
+    rep = Reporter(command="t", quiet=True)
+    assert run_phenotype(PhenoOptions(calls, db, guideline="dpwg"), rep) == 0
+    doc = json.loads((calls / "s1.pgx.json").read_text())
+    by = {g["gene"]: g for g in doc["genes"]}
+    assert by["CYP2D6"]["phenotype"] == "Poor Metabolizer" and "dpwg_via_cpic_mapping" in by["CYP2D6"]["flags"]
+    assert by["CYP3A4"]["phenotype"] is None and "dpwg_mapping_not_available" in by["CYP3A4"]["flags"]