bonesistools 1.2.5__tar.gz → 1.2.6__tar.gz

{bonesistools-1.2.5 → bonesistools-1.2.6}/.gitignore

@@ -62,6 +62,7 @@ Thumbs.db
 /tmp/
 /temp/
 *.tmp
+*tmp*
 *.bak
 *.bak.tsv
 

{bonesistools-1.2.5 → bonesistools-1.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: bonesistools
-Version: 1.2.5
+Version: 1.2.6
 Summary: Bioinformatics toolkit for upstream and downstream analyses around the BoNesis framework
 Project-URL: Repository, https://github.com/bnediction/bonesistools
 Author: Théo Roncalli
@@ -551,6 +551,7 @@ Requires-Python: >=3.7
 Requires-Dist: anndata>=0.8
 Requires-Dist: boolean-py>=4.0
 Requires-Dist: h5py>=3.7
+Requires-Dist: importlib-metadata; python_version < '3.8'
 Requires-Dist: importlib-resources; python_version < '3.9'
 Requires-Dist: matplotlib>=3.5
 Requires-Dist: networkx>=2.1

{bonesistools-1.2.5 → bonesistools-1.2.6}/pyproject.toml

@@ -5,7 +5,7 @@ build-backend = "hatchling.build"
 [project]
 name = "bonesistools"
 description = "Bioinformatics toolkit for upstream and downstream analyses around the BoNesis framework"
-version = "1.2.5"
+version = "1.2.6"
 license = {file = "LICENSE"}
 credits = ["BNeDiction", "PEPR Santé Numérique 2030"]
 authors = [
@@ -41,6 +41,7 @@ classifiers = [
 ]
 requires-python = ">=3.7"
 dependencies = [
+    "importlib-metadata; python_version < '3.8'",
     "importlib_resources; python_version < '3.9'",
     "typing-extensions>=4.0",
     "numpy>=1.21",

{bonesistools-1.2.5 → bonesistools-1.2.6}/src/bonesistools/__init__.py

@@ -29,6 +29,7 @@ from typing import List as _List
 from . import boolpy as bpy
 from . import databases as dbs
 from . import sctools as sct
+from ._metadata import package_version as _package_version
 from ._warnings import _warn_deprecated
 
 if _TYPE_CHECKING:
@@ -37,8 +38,10 @@ if _TYPE_CHECKING:
 del annotations
 
 __credits__ = "BNeDiction; PEPR Santé Numérique 2030"
+__version__ = _package_version()
 
 __all__ = [
+    "__version__",
     "sct",
     "bpy",
     "dbs",

bonesistools-1.2.6/src/bonesistools/_metadata.py

@@ -0,0 +1,70 @@
+#!/usr/bin/env python
+
+from __future__ import annotations
+
+import importlib
+from pathlib import Path
+from typing import Any, Optional, cast
+
+
+def package_version(distribution: str = "bonesistools") -> str:
+    """
+    Return the installed package version.
+
+    When the package is imported directly from a source checkout, distribution
+    metadata can be unavailable. In that case, read the version from
+    `pyproject.toml`.
+    """
+
+    source_version = _pyproject_version()
+    if source_version is not None:
+        return source_version
+
+    installed_version = _installed_version(distribution)
+    return installed_version if installed_version is not None else "0+unknown"
+
+
+def _installed_version(distribution: str) -> Optional[str]:
+
+    try:
+        metadata = importlib.import_module("importlib.metadata")
+    except ImportError:  # pragma: no cover - Python 3.7 compatibility
+        metadata = importlib.import_module("importlib_metadata")
+
+    try:
+        return cast(str, cast(Any, metadata).version(distribution))
+    except Exception as error:
+        if error.__class__.__name__ == "PackageNotFoundError":
+            return None
+        raise
+
+
+def _pyproject_version() -> Optional[str]:
+
+    pyproject = _find_pyproject()
+    if pyproject is None:
+        return None
+
+    in_project = False
+    for line in pyproject.read_text(encoding="utf-8").splitlines():
+        stripped = line.strip()
+        if stripped == "[project]":
+            in_project = True
+            continue
+        if in_project and stripped.startswith("["):
+            return None
+        if in_project and stripped.startswith("version"):
+            _, value = stripped.split("=", 1)
+            return value.strip().strip("\"'")
+
+    return None
+
+
+def _find_pyproject() -> Optional[Path]:
+
+    for parent in Path(__file__).resolve().parents:
+        pyproject = parent / "pyproject.toml"
+        if pyproject.exists():
+            return pyproject
+
+    return None

{bonesistools-1.2.5 → bonesistools-1.2.6}/src/bonesistools/databases/hcop/_orthologs.py

@@ -14,6 +14,7 @@ from typing import (
     List,
     Optional,
     Sequence,
+    Set,
     Tuple,
     Union,
     cast,
@@ -131,11 +132,84 @@ class Orthologs:
             table=table,
         )
 
+    @overload
+    def __call__(
+        self,
+        data: str,
+        *args: Any,
+        **kwargs: Any,
+    ) -> List[str]: ...
+
+    @overload
+    def __call__(
+        self,
+        data: List[str],
+        *args: Any,
+        **kwargs: Any,
+    ) -> List[str]: ...
+
+    @overload
+    def __call__(
+        self,
+        data: Tuple[str, ...],
+        *args: Any,
+        **kwargs: Any,
+    ) -> Tuple[str, ...]: ...
+
+    @overload
+    def __call__(
+        self,
+        data: Set[str],
+        *args: Any,
+        **kwargs: Any,
+    ) -> Set[str]: ...
+
+    @overload
+    def __call__(
+        self,
+        data: Sequence[str],
+        *args: Any,
+        **kwargs: Any,
+    ) -> Sequence[str]: ...
+
+    @overload
+    def __call__(
+        self,
+        data: InteractionList,
+        *args: Any,
+        **kwargs: Any,
+    ) -> InteractionList: ...
+
+    @overload
+    def __call__(
+        self,
+        data: pd.DataFrame,
+        *args: Any,
+        **kwargs: Any,
+    ) -> pd.DataFrame: ...
+
+    @overload
+    def __call__(
+        self,
+        data: Graph[Any],
+        *args: Any,
+        **kwargs: Any,
+    ) -> Graph[Any]: ...
+
+    @overload
+    def __call__(
+        self,
+        data: "BooleanNetworkLike",
+        *args: Any,
+        **kwargs: Any,
+    ) -> "BooleanNetworkLike": ...
+
     def __call__(
         self,
         data: Union[
             str,
             Sequence[str],
+            Set[str],
             InteractionList,
             pd.DataFrame,
             Graph[Any],
@@ -143,7 +217,7 @@ class Orthologs:
         ],
         *args: Any,
         **kwargs: Any,
-    ):
+    ) -> Any:
         """
         Translate genes in a supported object.
 
@@ -153,7 +227,8 @@ class Orthologs:
 
         Parameters
         ----------
-        data: str, sequence, InteractionList, DataFrame, Graph or BooleanNetworkLike
+        data: str, sequence, set, InteractionList, DataFrame, Graph or
+            BooleanNetworkLike
             Object containing human gene symbols to translate.
         *args: Any
             Positional arguments forwarded to the selected translation method.
@@ -181,7 +256,11 @@ class Orthologs:
         if (
             isinstance(data, SequenceInstance) and not isinstance(data, str)
         ) or isinstance(data, set):
-            return self.translate_sequence(cast(Sequence[str], data), *args, **kwargs)
+            return self.translate_sequence(
+                cast(Union[Sequence[str], Set[str]], data),
+                *args,
+                **kwargs,
+            )
         if isinstance(data, pd.DataFrame):
             return self.translate_df(data, *args, **kwargs)
         if isinstance(data, Graph):
@@ -190,7 +269,7 @@ class Orthologs:
             return self.translate_bn(data, *args, **kwargs)
         raise TypeError(
             f"unsupported argument type for 'data': "
-            f"expected str, sequence, interaction list, {pd.DataFrame}, {Graph} "
+            f"expected str, sequence, set, interaction list, {pd.DataFrame}, {Graph} "
             f"or Boolean network-like object but received {type(data)}"
         )
 
@@ -257,13 +336,41 @@ class Orthologs:
             return list(self._mapping[gene])
         return [gene] if keep_if_missing else []
 
+    @overload
+    def translate_sequence(
+        self,
+        genes: List[str],
+        keep_if_missing: bool = True,
+    ) -> List[str]: ...
+
+    @overload
+    def translate_sequence(
+        self,
+        genes: Tuple[str, ...],
+        keep_if_missing: bool = True,
+    ) -> Tuple[str, ...]: ...
+
+    @overload
+    def translate_sequence(
+        self,
+        genes: Set[str],
+        keep_if_missing: bool = True,
+    ) -> Set[str]: ...
+
+    @overload
     def translate_sequence(
         self,
         genes: Sequence[str],
         keep_if_missing: bool = True,
-    ) -> Sequence[str]:
+    ) -> Sequence[str]: ...
+
+    def translate_sequence(
+        self,
+        genes: Union[Sequence[str], Set[str]],
+        keep_if_missing: bool = True,
+    ) -> Union[Sequence[str], Set[str]]:
         """
-        Translate a sequence of human gene symbols.
+        Translate a sequence or set of human gene symbols.
 
         Each gene is translated to one target symbol using the deterministic
         HCOP ranking. Missing genes are either kept or removed depending on
@@ -271,15 +378,15 @@ class Orthologs:
 
         Parameters
         ----------
-        genes: sequence of str
+        genes: sequence or set of str
             Human gene symbols to translate.
         keep_if_missing: bool (default: True)
             If True, keep original gene symbols with no ortholog.
 
         Returns
         -------
-        sequence of str
-            Translated sequence, preserving the input sequence type when
+        sequence or set of str
+            Translated values, preserving the input collection type when
             possible.
         """
 

{bonesistools-1.2.5 → bonesistools-1.2.6}/src/bonesistools/databases/ncbi/_genesyn.py

@@ -386,7 +386,7 @@ class GeneSynonyms:
         self.ncbi_file = self.__resolve_gene_info_file(self.version)
 
         self.__download_gene_info()
-        self.__initialize_mappings(show_warnings=show_warnings)
+        self._initialize_mappings(show_warnings=show_warnings)
 
     def __call__(
         self,
@@ -1038,7 +1038,7 @@ class GeneSynonyms:
         self.ncbi_file = self.__resolve_gene_info_file(self.version)
 
         self.__download_gene_info()
-        self.__initialize_mappings(show_warnings=show_warnings)
+        self._initialize_mappings(show_warnings=show_warnings)
 
     def get_mapping(self):
         """
@@ -1693,7 +1693,7 @@ class GeneSynonyms:
 
         return gene_aliases_mapping
 
-    def __initialize_mappings(self, show_warnings: bool) -> None:
+    def _initialize_mappings(self, show_warnings: bool) -> None:
 
         self.show_warnings = show_warnings
         try:
@@ -1724,6 +1724,10 @@ class GeneSynonyms:
             *self.databases,
         )
 
+    def __initialize_mappings(self, show_warnings: bool) -> None:
+
+        self._initialize_mappings(show_warnings=show_warnings)
+
     @support_legacy_gene_synonyms_args
     def __conversion_function(
         self,

bonesistools-1.2.6/src/bonesistools/sctools/_stats.py

@@ -0,0 +1,90 @@
+#!/usr/bin/env python
+
+from __future__ import annotations
+
+import warnings
+from typing import Any, Optional, Tuple, cast
+
+import numpy as np
+from scipy import sparse
+
+
+def _column_mean_variance(
+    matrix: Any,
+    max_memory: Optional[int] = None,
+) -> Tuple[np.ndarray, np.ndarray]:
+
+    n_obs = int(matrix.shape[0])
+    n_vars = int(matrix.shape[1])
+    if sparse.issparse(matrix):
+        sparse_format = getattr(matrix, "format", None)
+        if sparse_format == "csr":
+            sums = np.bincount(
+                matrix.indices,
+                weights=matrix.data,
+                minlength=n_vars,
+            )
+            squared_sums = np.bincount(
+                matrix.indices,
+                weights=matrix.data * matrix.data,
+                minlength=n_vars,
+            )
+        elif sparse_format == "csc":
+            sums = np.zeros(n_vars, dtype=np.float64)
+            squared_sums = np.zeros(n_vars, dtype=np.float64)
+            nonempty_columns = np.diff(matrix.indptr) > 0
+            if np.any(nonempty_columns):
+                starts = matrix.indptr[:-1][nonempty_columns]
+                data = np.asarray(matrix.data, dtype=np.float64)
+                sums[nonempty_columns] = np.add.reduceat(data, starts)
+                squared_sums[nonempty_columns] = np.add.reduceat(data * data, starts)
+        else:
+            return _column_mean_variance(
+                matrix.tocsr(),
+                max_memory=max_memory,
+            )
+
+        means = sums / n_obs
+        mean_squares = squared_sums / n_obs
+    else:
+        dense_matrix = cast(np.ndarray, matrix)
+        means = np.empty(n_vars, dtype=np.float64)
+        mean_squares = np.empty(n_vars, dtype=np.float64)
+        chunk_size = _dense_column_chunk_size(dense_matrix, max_memory)
+        for start in range(0, n_vars, chunk_size):
+            end = min(start + chunk_size, n_vars)
+            dense_chunk = dense_matrix[:, start:end]
+            means[start:end] = np.asarray(
+                dense_chunk.mean(axis=0, dtype=np.float64)
+            ).ravel()
+            squared_chunk = np.multiply(dense_chunk, dense_chunk)
+            mean_squares[start:end] = np.asarray(
+                squared_chunk.mean(axis=0, dtype=np.float64)
+            ).ravel()
+
+    variances = np.maximum(mean_squares - means**2, 0)
+    if n_obs > 1:
+        variances *= n_obs / (n_obs - 1)
+    return means, variances
+
+
+def _dense_column_chunk_size(
+    matrix: np.ndarray,
+    max_memory: Optional[int],
+) -> int:
+
+    n_obs = int(matrix.shape[0])
+    n_vars = int(matrix.shape[1])
+    if max_memory is None:
+        return max(1, n_vars)
+
+    bytes_per_value = np.dtype(matrix.dtype).itemsize
+    bytes_per_column = n_obs * bytes_per_value
+    if max_memory < bytes_per_column:
+        warnings.warn(
+            "Requested memory budget is smaller than the memory required for a "
+            "single variable. Computation will proceed with chunk_size=1.",
+            RuntimeWarning,
+            stacklevel=2,
+        )
+    return max(1, int(max_memory // bytes_per_column))

{bonesistools-1.2.5 → bonesistools-1.2.6}/src/bonesistools/sctools/datasets/__init__.py

@@ -20,9 +20,12 @@ else:
 import anndata as _ad
 from anndata import AnnData as _AnnData
 
+from ._geo import from_geo
+
 del annotations
 
 __all__ = [
+    "from_geo",
     "nestorowa",
 ]
 
@@ -37,14 +40,11 @@ def nestorowa() -> _AnnData:
     AnnData
         Preprocessed AnnData object containing highly variable genes.
 
-    Notes
-    -----
-    Dataset derived from:
-
-        Nestorowa et al. (2016)
-        "A single-cell resolution map of mouse hematopoietic stem and "
-        "progenitor cell differentiation"
-        Blood 128(8): e20-e31
+    References
+    ----------
+    Nestorowa et al. (2016). A single-cell resolution map of mouse
+    hematopoietic stem and progenitor cell differentiation. Blood, 128(8),
+    e20-e31.
     """
 
     path = _files(__package__) / "nestorowa_hvg.h5ad"