statgen 0.2.6__tar.gz → 0.3.0__tar.gz

{statgen-0.2.6 → statgen-0.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: statgen
-Version: 0.2.6
+Version: 0.3.0
 Summary: Statistical genetics data objects and loaders for Python and MATLAB/Octave.
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown

{statgen-0.2.6 → statgen-0.3.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "statgen"
-version = "0.2.6"
+version = "0.3.0"
 description = "Statistical genetics data objects and loaders for Python and MATLAB/Octave."
 readme = "README.md"
 requires-python = ">=3.10"

{statgen-0.2.6 → statgen-0.3.0}/statgen/__init__.py

@@ -1,5 +1,5 @@
 from ._verbosity import get_verbosity, set_verbosity
 
-__version__ = "0.2.6"
+__version__ = "0.3.0"
 
 __all__ = ["__version__", "get_verbosity", "set_verbosity"]

{statgen-0.2.6 → statgen-0.3.0}/statgen/_bfile_utils.py

@@ -17,6 +17,7 @@ _FAM_NAMES = ["fid", "iid", "father_id", "mother_id", "sex", "pheno"]
 _PLOIDY_NAMES = ["ploidy_male", "ploidy_female"]
 _DNA_ALLELE_RE = re.compile(r"^[ACGT]+$")
 _BED_LOOKUP_INT8 = None
+_BED_MAX_READ_BYTES = 64 * 1024 * 1024
 
 
 def _read_csv_strict(path: Path, *, sep, names, description: str | None = None) -> pd.DataFrame:
@@ -256,20 +257,33 @@ def _bed_lookup_int8() -> np.ndarray:
 def _read_bed_rows_int8(path: Path, source_rows: np.ndarray, source_num_sample: int) -> np.ndarray:
     path = Path(path)
     source_rows = np.asarray(source_rows, dtype=np.int64).reshape(-1)
-    bytes_per_snp = (int(source_num_sample) + 3) // 4
-    decoded = np.empty((int(source_num_sample), source_rows.size), dtype=np.int8)
+    n_samples = int(source_num_sample)
+    bytes_per_snp = (n_samples + 3) // 4
+    if source_rows.size == 0:
+        return np.empty((n_samples, 0), dtype=np.int8)
+    unique_rows, inverse = np.unique(source_rows, return_inverse=True)
+    unique_decoded = np.empty((n_samples, unique_rows.size), dtype=np.int8)
+    rows_per_read = unique_rows.size if bytes_per_snp == 0 else max(1, _BED_MAX_READ_BYTES // bytes_per_snp)
     lookup = _bed_lookup_int8()
     with open(path, "rb") as f:
-        # PERF: loop over requested SNPs retained; each SNP is a separate BED row
-        #       requiring an individual seek. Inner decode is vectorized across subjects.
-        for j, source_row0 in enumerate(source_rows):
-            f.seek(3 + int(source_row0) * bytes_per_snp)
-            packed = f.read(bytes_per_snp)
-            if len(packed) != bytes_per_snp:
-                raise ValueError(
-                    f"{path}: short BED read for source SNP row {int(source_row0)}: "
-                    f"expected {bytes_per_snp} bytes, got {len(packed)}"
-                )
-            byte_values = np.frombuffer(packed, dtype=np.uint8)
-            decoded[:, j] = lookup[byte_values].reshape(-1)[:source_num_sample]
-    return decoded
+        group_starts = np.r_[0, np.flatnonzero(np.diff(unique_rows) != 1) + 1]
+        group_stops = np.r_[group_starts[1:], unique_rows.size]
+        # PERF: one seek/read per contiguous BED row block; requested order and
+        #       repeats are restored after block decoding. Blocks are capped to
+        #       bound transient packed/decode allocations for large requests.
+        for start, stop in zip(group_starts, group_stops, strict=True):
+            for chunk_start in range(start, stop, rows_per_read):
+                chunk_stop = min(chunk_start + rows_per_read, stop)
+                block_rows = unique_rows[chunk_start:chunk_stop]
+                expected_bytes = int(bytes_per_snp * block_rows.size)
+                f.seek(3 + int(block_rows[0]) * bytes_per_snp)
+                packed = f.read(expected_bytes)
+                if len(packed) != expected_bytes:
+                    raise ValueError(
+                        f"{path}: short BED read for source SNP rows {int(block_rows[0])}-{int(block_rows[-1])}: "
+                        f"expected {expected_bytes} bytes, got {len(packed)}"
+                    )
+                byte_values = np.frombuffer(packed, dtype=np.uint8)
+                block_decoded = lookup[byte_values].reshape(block_rows.size, bytes_per_snp * 4)
+                unique_decoded[:, chunk_start:chunk_stop] = block_decoded[:, :n_samples].T
+    return unique_decoded[:, inverse]

{statgen-0.2.6 → statgen-0.3.0}/statgen/annotations.py

@@ -12,13 +12,13 @@ _CACHE_SCHEMA = "annotations_cache/0.1"
 
 
 
-def _as_sparse_binary(mat) -> sparse.csr_matrix:
+def _as_sparse_binary(mat, name: str = "annomat") -> sparse.csr_matrix:
     if sparse.issparse(mat):
         csr = mat.tocsr(copy=True)
     else:
         arr = np.asarray(mat)
         if arr.ndim != 2:
-            raise ValueError("annomat must be a 2D matrix")
+            raise ValueError(f"{name} must be a 2D matrix")
         csr = sparse.csr_matrix(arr)
 
     # Guard against CSR matrices carrying explicit stored zeros in .data.
@@ -30,7 +30,7 @@ def _as_sparse_binary(mat) -> sparse.csr_matrix:
         bad = data != 1
         if bad.any():
             idx = int(np.flatnonzero(bad)[0])
-            raise ValueError(f"annomat contains non-binary value: {data[idx]!r}")
+            raise ValueError(f"{name} contains non-binary value: {data[idx]!r}")
 
     csr = csr.astype(np.uint8)
     csr.data[:] = 1
@@ -336,17 +336,17 @@ class AnnotationPanel:
 
         return AnnotationPanel(out_shards, lhs_names + rhs_names)
 
-    def save_cache(self, path, format=None) -> None:
-        save_annotations_cache(self, path, format=format)
+    def save_cache(self, path) -> None:
+        save_annotations_cache(self, path)
 
 
-def create_annotations(reference, annomat, annonames) -> AnnotationPanel:
-    names = _coerce_annonames(annonames)
+def create_annotations(reference, annotation_matrix, annotation_names) -> AnnotationPanel:
+    names = _coerce_annonames(annotation_names)
     n = int(reference.num_snp)
-    mat = _as_sparse_binary(annomat)
+    mat = _as_sparse_binary(annotation_matrix, "annotation_matrix")
     if mat.shape != (n, len(names)):
         raise ValueError(
-            f"annomat shape mismatch: expected ({n}, {len(names)}), got {mat.shape}"
+            f"annotation_matrix shape mismatch: expected ({n}, {len(names)}), got {mat.shape}"
         )
 
     out_shards = []
@@ -369,7 +369,7 @@ def create_annotation(reference, annovec, annoname) -> AnnotationPanel:
         raise ValueError("annoname must be non-empty")
     vec = _coerce_annovec(annovec, int(reference.num_snp))
     mat = sparse.csr_matrix(vec.reshape(-1, 1))
-    return create_annotations(reference, mat, [name])
+    return create_annotations(reference, annotation_matrix=mat, annotation_names=[name])
 
 
 def load_annotations(bed_paths, reference) -> AnnotationPanel:
@@ -384,10 +384,10 @@ def load_annotations(bed_paths, reference) -> AnnotationPanel:
             raise FileNotFoundError(f"BED file not found: {p}")
 
     annomat, annonames = _paint_annotations(paths, reference)
-    return create_annotations(reference, annomat, annonames)
+    return create_annotations(reference, annotation_matrix=annomat, annotation_names=annonames)
 
 
-def save_annotations_cache(panel: AnnotationPanel, path, format=None) -> None:
+def save_annotations_cache(panel: AnnotationPanel, path) -> None:
     meta = {
         "schema": _CACHE_SCHEMA,
         "shard_labels": [s.label for s in panel.shards],

{statgen-0.2.6 → statgen-0.3.0}/statgen/genotype.py

@@ -519,8 +519,8 @@ class GenotypePanel:
                 same_ploidy = ploidy_male.reshape(1, -1)
                 geno[np.ix_(is_unknown, cols)] *= same_ploidy / 2.0
 
-    def save_cache(self, path, format=None) -> None:
-        save_genotype_cache(self, path, format=format)
+    def save_cache(self, path) -> None:
+        save_genotype_cache(self, path)
 
 
 def _fam_equal(lhs: pd.DataFrame, rhs: pd.DataFrame) -> bool:
@@ -642,7 +642,11 @@ def load_genotype(bfile_prefix, reference) -> GenotypePanel:
     else:
         source_layout = "non_sharded"
         shared = _load_source_record(prefix, None)
-        if "X" in ref_labels and Path(prefix + ".ploidy").is_file() is False:
+        if (
+            "X" in ref_labels
+            and (shared.bim["chr"] == "X").any()
+            and Path(prefix + ".ploidy").is_file() is False
+        ):
             warnings.warn(
                 f"{prefix}: chrX genotype source has no .ploidy sidecar; "
                 "defaulting chrX rows to male/female ploidy (1, 2)",
@@ -692,7 +696,7 @@ def load_genotype(bfile_prefix, reference) -> GenotypePanel:
     )
 
 
-def save_genotype_cache(panel: GenotypePanel, path, format=None) -> None:
+def save_genotype_cache(panel: GenotypePanel, path) -> None:
     meta = {
         "schema": _CACHE_SCHEMA,
         "n_shards": len(panel.shards),

{statgen-0.2.6 → statgen-0.3.0}/statgen/reference.py

@@ -341,8 +341,8 @@ class ReferencePanel:
                 ok = False
         return ok
 
-    def save_cache(self, path, format=None) -> None:
-        save_reference_cache(self, path, format=format)
+    def save_cache(self, path) -> None:
+        save_reference_cache(self, path)
 
 
 def load_reference(path, shards=None) -> ReferencePanel:
@@ -403,7 +403,7 @@ def load_reference(path, shards=None) -> ReferencePanel:
     return ReferencePanel(out_shards)
 
 
-def save_reference_cache(panel: ReferencePanel, path, format=None) -> None:
+def save_reference_cache(panel: ReferencePanel, path) -> None:
     # Metadata (schema, labels, checksums) as a compact JSON blob stored in the npz.
     # SNP-axis numeric vectors (bp) and string arrays are stored as native
     # binary numpy arrays — not JSON — per the performance contract.

{statgen-0.2.6 → statgen-0.3.0}/statgen/sumstats.py

@@ -252,8 +252,8 @@ class Sumstats:
         by_label = {s.label: s for s in self._shards}
         return Sumstats([by_label[label] for label in selected])
 
-    def save_cache(self, path, format=None) -> None:
-        save_sumstats_cache(self, path, format=format)
+    def save_cache(self, path) -> None:
+        save_sumstats_cache(self, path)
 
 
 def _derive_logp(p_vals: np.ndarray) -> np.ndarray:
@@ -393,26 +393,26 @@ def load_sumstats(path, reference) -> Sumstats:
 
 def create_sumstats(
     reference,
-    pvec,
-    zvec=None,
-    nvec=None,
-    beta_vec=None,
-    se_vec=None,
-    eaf_vec=None,
-    info_vec=None,
+    p,
+    z=None,
+    n=None,
+    beta=None,
+    se=None,
+    eaf=None,
+    info=None,
 ) -> Sumstats:
-    n = int(reference.num_snp)
-    aligned_p = _coerce_aligned_vec("pvec", pvec, n=n, allow_inf=True)
-    _validate_pvec(aligned_p, "pvec")
+    num_snp_value = int(reference.num_snp)
+    aligned_p = _coerce_aligned_vec("p", p, n=num_snp_value, allow_inf=True)
+    _validate_pvec(aligned_p, "p")
     aligned_logp = _derive_logp(aligned_p)
 
     aligned_optional = {
-        "z": None if zvec is None else _coerce_aligned_vec("zvec", zvec, n=n, allow_inf=False),
-        "n": None if nvec is None else _coerce_aligned_vec("nvec", nvec, n=n, allow_inf=False),
-        "beta": None if beta_vec is None else _coerce_aligned_vec("beta_vec", beta_vec, n=n, allow_inf=False),
-        "se": None if se_vec is None else _coerce_aligned_vec("se_vec", se_vec, n=n, allow_inf=False),
-        "eaf": None if eaf_vec is None else _coerce_aligned_vec("eaf_vec", eaf_vec, n=n, allow_inf=False),
-        "info": None if info_vec is None else _coerce_aligned_vec("info_vec", info_vec, n=n, allow_inf=False),
+        "z": None if z is None else _coerce_aligned_vec("z", z, n=num_snp_value, allow_inf=False),
+        "n": None if n is None else _coerce_aligned_vec("n", n, n=num_snp_value, allow_inf=False),
+        "beta": None if beta is None else _coerce_aligned_vec("beta", beta, n=num_snp_value, allow_inf=False),
+        "se": None if se is None else _coerce_aligned_vec("se", se, n=num_snp_value, allow_inf=False),
+        "eaf": None if eaf is None else _coerce_aligned_vec("eaf", eaf, n=num_snp_value, allow_inf=False),
+        "info": None if info is None else _coerce_aligned_vec("info", info, n=num_snp_value, allow_inf=False),
     }
 
     return _build_sumstats_from_aligned(
@@ -422,7 +422,7 @@ def create_sumstats(
     )
 
 
-def save_sumstats_cache(sumstats: Sumstats, path, format=None) -> None:
+def save_sumstats_cache(sumstats: Sumstats, path) -> None:
     meta = {
         "schema": _CACHE_SCHEMA,
         "shard_labels": [s.label for s in sumstats.shards],

{statgen-0.2.6 → statgen-0.3.0}/statgen.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: statgen
-Version: 0.2.6
+Version: 0.3.0
 Summary: Statistical genetics data objects and loaders for Python and MATLAB/Octave.
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown