seqpro 0.17.0__tar.gz → 0.18.0__tar.gz

{seqpro-0.17.0 → seqpro-0.18.0}/.github/workflows/bench.yaml

@@ -30,7 +30,7 @@ jobs:
           environments: bench
           locked: true
       - name: Build extension
-        run: pixi run -e bench maturin develop --uv
+        run: pixi run -e bench maturin develop
       - name: Run benchmarks
         uses: CodSpeedHQ/action@v4
         with:

{seqpro-0.17.0 → seqpro-0.18.0}/CHANGELOG.md

@@ -1,3 +1,15 @@
+## 0.18.0 (2026-06-23)
+
+### Feat
+
+- **rag**: reshape/squeeze/to_packed preserve Ragged subclass
+- **rag**: __getitem__ preserves subclass on positional indexing
+- **rag**: add _with_layout subclass-preserving constructor
+
+### Fix
+
+- **rag**: non-tuple record indexing matches numpy (A[x] == A[(x,)])
+
 ## 0.17.0 (2026-06-23)
 
 ### Feat

{seqpro-0.17.0 → seqpro-0.18.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: seqpro
-Version: 0.17.0
+Version: 0.18.0
 Classifier: Programming Language :: Rust
 Classifier: Programming Language :: Python :: Implementation :: CPython
 Classifier: Programming Language :: Python :: Implementation :: PyPy

{seqpro-0.17.0 → seqpro-0.18.0}/pixi.lock

@@ -132,6 +132,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/parso-0.8.7-pyhcf101f3_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/patsy-1.0.2-pyhcf101f3_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/pexpect-4.9.0-pyhd8ed1ab_1.conda
+      - conda: https://conda.anaconda.org/conda-forge/noarch/pip-26.1.2-pyh145f28c_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/platformdirs-4.9.6-pyhcf101f3_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/pluggy-1.6.0-pyhd8ed1ab_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/prompt-toolkit-3.0.52-pyha770c72_0.conda
@@ -262,6 +263,7 @@ environments:
       - conda: https://conda.anaconda.org/conda-forge/noarch/parso-0.8.7-pyhcf101f3_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/patsy-1.0.2-pyhcf101f3_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/pexpect-4.9.0-pyhd8ed1ab_1.conda
+      - conda: https://conda.anaconda.org/conda-forge/noarch/pip-26.1.2-pyh145f28c_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/platformdirs-4.9.6-pyhcf101f3_0.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/pluggy-1.6.0-pyhf9edf01_1.conda
       - conda: https://conda.anaconda.org/conda-forge/noarch/prompt-toolkit-3.0.52-pyha770c72_0.conda
@@ -6564,6 +6566,18 @@ packages:
   - pkg:pypi/pickleshare?source=hash-mapping
   size: 11748
   timestamp: 1733327448200
+- conda: https://conda.anaconda.org/conda-forge/noarch/pip-26.1.2-pyh145f28c_0.conda
+  sha256: 9e673d3c6003f416df11670a14b026a04e3a45ebec55357987e15b860f138f2a
+  md5: 733cc07ed34162ac50b936464b163366
+  depends:
+  - python >=3.13.0a0
+  license: MIT
+  license_family: MIT
+  purls:
+  - pkg:pypi/pip?source=compressed-mapping
+  run_exports: {}
+  size: 1202377
+  timestamp: 1780262809860
 - conda: https://conda.anaconda.org/conda-forge/noarch/platformdirs-4.3.8-pyhe01879c_0.conda
   sha256: 0f48999a28019c329cd3f6fd2f01f09fc32cc832f7d6bbe38087ddac858feaa3
   md5: 424844562f5d337077b445ec6b1398a7

{seqpro-0.17.0 → seqpro-0.18.0}/pixi.toml

@@ -95,6 +95,10 @@ bench = "pytest tests/test_bench_tokenize.py --codspeed"
 rag-gate = "python benchmarks/bench_ragged_backends.py"
 
 [feature.bench.dependencies]
+# pip is required so `maturin develop` can install the editable build into
+# the bench env. (`--uv` instead trips a uv installer bug on conda-managed
+# transitive deps: "Installed distribution could not be found: ncls".)
+pip = "*"
 marimo = ">=0.23.5,<0.24"
 seaborn = ">=0.13.2,<0.14"
 statsmodels = ">=0.14.6,<0.15"

{seqpro-0.17.0 → seqpro-0.18.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "seqpro"
-version = "0.17.0"
+version = "0.18.0"
 authors = [
     { name = "David Laub", email = "dlaub@ucsd.edu" },
     { name = "Adam Klie", email = "aklie@ucsd.edu" },

{seqpro-0.17.0 → seqpro-0.18.0}/python/seqpro/rag/_core.py

@@ -75,6 +75,15 @@ class Ragged(NDArrayOperatorsMixin, Generic[RDTYPE_co]):
             validate_layout(data)
         self._layout = data
 
+    def _with_layout(self, layout: Any) -> "Ragged[Any]":
+        """Reconstruct a same-kind container around ``layout``, preserving the
+        concrete subclass. Bypasses ``__init__`` (subclasses carry no state beyond
+        ``_layout``; see the subclassing contract). Used by structural transforms
+        so a ``Ragged`` subclass survives slicing/reshape/squeeze/to_packed."""
+        obj = object.__new__(type(self))
+        obj._layout = layout
+        return obj
+
     @property
     def _is_record(self) -> bool:
         return isinstance(self._layout, RecordLayout)
@@ -425,6 +434,19 @@ class Ragged(NDArrayOperatorsMixin, Generic[RDTYPE_co]):
     def __getitem__(
         self, where: Any
     ) -> "NDArray[Any] | bytes | dict[str, Any] | Ragged[Any]":
+        result = self._getitem(where)
+        # Preserve the concrete subclass for positional (structural) results.
+        # A string key is field extraction -> keep the bare field as base Ragged.
+        # Non-Ragged results (dict / bytes / ndarray / scalar) pass through.
+        if (
+            type(self) is not Ragged
+            and not isinstance(where, str)
+            and isinstance(result, Ragged)
+        ):
+            return self._with_layout(result._layout)
+        return result
+
+    def _getitem(self, where: Any) -> Any:
         # np.newaxis / None: insert a size-1 leading axis.
         # e.g. shape (batch, None)[None] -> (1, batch, None)
         if where is None:
@@ -813,7 +835,18 @@ class Ragged(NDArrayOperatorsMixin, Generic[RDTYPE_co]):
             except KeyError:
                 raise KeyError(where)
             return Ragged(field)  # field.offsets[0] is the shared object (zero-copy)
-        return self._getitem_record_rows(where)  # Task 8
+        # numpy contract: a non-tuple key is treated as a 1-tuple (A[x] == A[(x,)]).
+        # For a single-ragged-axis record with >1 leading fixed axis, route through
+        # the multidim peel path directly (under the SAME guard it requires, so we
+        # never re-dispatch through __getitem__ and risk recursion). Records that are
+        # not single-None fall through to the flat record-rows path unchanged.
+        if (
+            not isinstance(where, tuple)
+            and self.rag_dim > 1
+            and rec.shape.count(None) == 1
+        ):
+            return self._getitem_tuple_multidim((where,))
+        return self._getitem_record_rows(where)
 
     def _gather_indices(
         self, where: Any, starts: "NDArray[Any]", stops: "NDArray[Any]"
@@ -1161,6 +1194,14 @@ class Ragged(NDArrayOperatorsMixin, Generic[RDTYPE_co]):
 
     def squeeze(
         self, axis: int | tuple[int, ...] | None = None
+    ) -> "Ragged[Any] | NDArray[Any]":
+        result = self._squeeze_impl(axis)
+        if type(self) is not Ragged and isinstance(result, Ragged):
+            return self._with_layout(result._layout)
+        return result
+
+    def _squeeze_impl(
+        self, axis: int | tuple[int, ...] | None = None
     ) -> "Ragged[Any] | NDArray[Any]":
         if isinstance(self._layout, RecordLayout):
             return Ragged.from_fields(
@@ -1210,6 +1251,12 @@ class Ragged(NDArrayOperatorsMixin, Generic[RDTYPE_co]):
         )
 
     def reshape(self, *shape: int | None) -> "Ragged[Any]":
+        result = self._reshape_impl(*shape)
+        if type(self) is not Ragged and isinstance(result, Ragged):
+            return self._with_layout(result._layout)
+        return result
+
+    def _reshape_impl(self, *shape: int | None) -> "Ragged[Any]":
         if isinstance(self._layout, RecordLayout):
             return Ragged.from_fields(
                 {f: Ragged(fl).reshape(*shape) for f, fl in self._layout.fields.items()}
@@ -1270,6 +1317,12 @@ class Ragged(NDArrayOperatorsMixin, Generic[RDTYPE_co]):
         return raw.reshape(reshape_arg)
 
     def to_packed(self, *, copy: bool = True) -> "Ragged[Any]":
+        result = self._to_packed_impl(copy=copy)
+        if type(self) is not Ragged and isinstance(result, Ragged):
+            return self._with_layout(result._layout)
+        return result
+
+    def _to_packed_impl(self, *, copy: bool = True) -> "Ragged[Any]":
         from ._ops import _pack_parts
 
         if isinstance(self._layout, RecordLayout):

seqpro-0.18.0/tests/test_ragged_record_indexing.py

@@ -0,0 +1,47 @@
+import numpy as np
+import pytest
+
+from seqpro.rag import Ragged
+
+
+def _multidim_record():
+    """A (3, 2, None) numeric record: 6 ragged segments over a (batch=3, ploidy=2)
+    leading grid."""
+    lengths = np.array([[1, 2], [3, 1], [2, 2]], np.uint32)  # (3, 2)
+    data = np.arange(int(lengths.sum()), dtype=np.int32)
+    start = Ragged.from_lengths(data, lengths.reshape(-1)).reshape(3, 2, None)
+    return Ragged.from_fields({"start": start})
+
+
+@pytest.mark.parametrize(
+    "key",
+    [
+        0,
+        slice(1, 3),
+        slice(None),
+        np.array([True, False, True]),
+        np.array([0, 2]),
+    ],
+    ids=["int", "slice", "full_slice", "mask", "int_array"],
+)
+def test_nontuple_equals_one_tuple_multidim_record(key):
+    """numpy contract: rec[k] must equal rec[(k,)] for a multi-leading-axis record."""
+    rec = _multidim_record()
+    got = rec[key]
+    want = rec[(key,)]
+    assert type(got) is type(want), (type(got), type(want))
+    if isinstance(want, Ragged):
+        assert got.shape == want.shape, (got.shape, want.shape)
+        np.testing.assert_array_equal(
+            np.asarray(got["start"].data), np.asarray(want["start"].data)
+        )
+        np.testing.assert_array_equal(
+            np.asarray(got["start"].offsets[0]), np.asarray(want["start"].offsets[0])
+        )
+
+
+def test_slice_preserves_ploidy_axis():
+    """Regression for the specific bug: rec[1:3] kept the ploidy axis."""
+    rec = _multidim_record()
+    assert rec[1:3].shape == (2, 2, None)
+    assert rec[0].shape == (2, None)

seqpro-0.18.0/tests/test_ragged_subclass.py

@@ -0,0 +1,60 @@
+import numpy as np
+
+from seqpro.rag import Ragged
+
+
+class _Sub(Ragged):
+    __slots__ = ()
+
+
+def _ragged():
+    return Ragged.from_lengths(
+        np.arange(6, dtype=np.int32), np.array([2, 1, 3], np.uint32)
+    )
+
+
+def test_with_layout_preserves_subclass():
+    sub = _Sub(_ragged()._layout)
+    out = sub._with_layout(_ragged()._layout)
+    assert type(out) is _Sub
+
+
+def test_with_layout_base_returns_base():
+    base = _ragged()
+    out = base._with_layout(_ragged()._layout)
+    assert type(out) is Ragged
+
+
+def _record():
+    a = Ragged.from_lengths(
+        np.arange(6, dtype=np.int32), np.array([2, 1, 3], np.uint32)
+    )
+    b = Ragged.from_lengths(
+        np.arange(6, dtype=np.int32) * 10, np.array([2, 1, 3], np.uint32)
+    )
+    return Ragged.from_fields({"a": a, "b": b})
+
+
+def test_getitem_positional_preserves_subclass():
+    sub = _Sub(_record()._layout)
+    assert type(sub[0:2]) is _Sub  # positional row slice -> subclass
+    assert type(sub[np.array([0, 2])]) is _Sub
+
+
+def test_getitem_field_extraction_stays_base():
+    sub = _Sub(_record()._layout)
+    assert type(sub["a"]) is Ragged  # string key -> bare field, base Ragged
+
+
+def test_reshape_squeeze_to_packed_preserve_subclass():
+    sub = _Sub(_record()._layout)  # shape (3, None) record
+    assert type(sub.reshape(1, 3, None)) is _Sub
+    assert type(sub.reshape(1, 3, None).squeeze(0)) is _Sub
+    assert type(sub.to_packed()) is _Sub
+
+
+def test_reshape_squeeze_to_packed_base_unchanged():
+    base = _record()
+    assert type(base.reshape(1, 3, None)) is Ragged
+    assert type(base.reshape(1, 3, None).squeeze(0)) is Ragged
+    assert type(base.to_packed()) is Ragged