P4 (WIP): GenCall-equivalent calling, BAF/LRR, sex, GTC v5, simurg genotype

- genotype/clusters.py: EGT aligned to manifest order (float32 mean R)
- genotype/gencall.py: flair, six-case likelihood, priors/posteriors,
  fuzzy sub-scores and score map, vectorised with the reference's
  float/double evaluation order and libm transcendentals
- genotype/baf_lrr.py, sex.py (rules v1/v2), sample.py (IDAT -> Gtc)
- pipeline/inputs.py shared by ingest and genotype
- pipeline/genotype.py + CLI: GTC per sample, sample_summary.tsv/json
  (OUT-060..063), run manifest
- GTCs from IDATs are byte-identical to the oracle for 36 GSA samples x
  2 normalization versions (NaN sign canonicalised to the x86 pattern)

src/simurg/pipeline/ingest.py

@@ -36,9 +36,9 @@ from simurg.core.schemas import (
     write_parquet,
 )
 from simurg.formats.bpm import Manifest, read_bpm
-from simurg.formats.discovery import ResolvedSample, discover_idats, resolve_input_folder, resolve_samples
+from simurg.formats.discovery import ResolvedSample
 from simurg.formats.idat import check_pair, parse_idat_filename, read_idat
-from simurg.formats.samplesheet import SampleSheet, output_stem, parse_samplesheet
+from simurg.pipeline.inputs import check_paths, resolve_sample_set, sample_dir
 from simurg.signal.raw_xy import LocusMapCache, assemble_raw_xy, check_coverage
 
 SUMMARY_COLUMNS = [
@@ -93,7 +93,7 @@ def _init_worker(manifest: Manifest, manifest_sha256: str) -> None:
 
 
 def _sample_paths(out: Path, stem: str, layout: str) -> tuple[Path, Path]:
-    d = out / "samples" / stem if layout == "nested" else out
+    d = sample_dir(out, stem, layout)
     return d / f"{stem}.intensity.parquet", d / f"{stem}.controls.parquet"
 
 
@@ -202,61 +202,18 @@ def run_ingest(opts: IngestOptions, rep: Reporter) -> int:
         v = getattr(opts, k)
         rm.set_option(k, str(v) if v is not None else None, "cli")
 
-    for p, what in (
-        (opts.bpm, "BPM"),
-        (opts.sheet, "sample sheet"),
-        (opts.csv, "CSV manifest"),
-        (opts.idats, "IDAT folder"),
-    ):
-        if p is not None and not Path(p).exists():
-            raise SimurgError("E_FILE_NOT_FOUND", f"{what} not found: {p}", path=str(p))
-
-    # --- inputs (IN-001 … IN-022) --------------------------------------------------------------------
-    sheet: SampleSheet | None = parse_samplesheet(opts.sheet) if opts.sheet else None
-    if sheet:
-        rm.sheet_header = sheet.header
-        for code, msg in sheet.warnings:
-            rep.warn(code, msg)
-    folder, notes = resolve_input_folder(opts.idats, sheet)
-    for code, msg in notes:
-        rep.warn(code, msg) if code.startswith("W_") else rep.info(msg, code)
-    disc = discover_idats(folder, on_duplicate=opts.on_duplicate_idat)
-    for code, msg in disc.warnings:
-        rep.warn(code, msg)
-    samples, notes = resolve_samples(disc, sheet)
-    for code, msg in notes:
-        rep.info(msg, code)
-    if not samples:
-        raise SimurgError("E_NO_INPUT", f"no IDAT pairs found under {folder}")
-
-    # --- names and output paths (IN-010, IN-012) -----------------------------------------------------
-    use_idat_names = sheet is None or not sheet.has_sample_name
-    stems: dict[str, str] = {}
+    check_paths(
+        [(opts.bpm, "BPM"), (opts.sheet, "sample sheet"), (opts.csv, "CSV manifest"), (opts.idats, "IDAT folder")]
+    )
+    sset = resolve_sample_set(opts.idats, opts.sheet, opts.on_duplicate_idat, rep, rm)
+    samples = sset.samples
     jobs: list[SampleJob] = []
     rows: list[dict[str, Any]] = []
     for s in samples:
-        if use_idat_names and s.pair is not None and s.pair.grn is not None:
-            # sheet Sample_Name > IDAT sample name (field 406) > sample_id
-            try:
-                idat_name = read_idat(s.pair.grn, load_arrays=False).sample_id.strip()
-            except SimurgError:
-                idat_name = ""
-            if idat_name:
-                s.sample_name = idat_name
-        stem = output_stem(s.sample_name)
-        if stem in stems and stems[stem] != s.sample_id:
-            if not use_idat_names:  # sheet names that sanitise to the same stem
-                raise SimurgError(
-                    "E_SHEET_NAME_COLLISION",
-                    f"samples {stems[stem]} and {s.sample_id} both map to output name {stem!r}",
-                )
-            rep.info(f"{s.sample_id}: IDAT sample name {s.sample_name!r} is not unique; using the sample ID")
-            s.sample_name = s.sample_id
-            stem = output_stem(s.sample_id)
-        stems[stem] = s.sample_id
         if s.status != "READY" or s.pair is None or not s.pair.complete:
             rows.append(_missing_row(s))
             continue
+        stem = sset.stems[s.sample_id]
         ip, cp = _sample_paths(opts.out, stem, opts.layout)
         jobs.append(
             SampleJob(s.sample_id, s.sample_name, stem, str(s.pair.grn), str(s.pair.red), str(ip), str(cp), opts.resume)