scoring-matrices - npm Package Compare versions

Comparing version 0.1.1 to 0.2.0

PKG-INFO

 Metadata-Version: 2.1
 Name: scoring-matrices
-Version: 0.1.1
+Version: 0.2.0
 Summary: Dependency free, Cython-compatible scoring matrices to use with biological sequences.
@@ -8,3 +8,3 @@ Home-page: https://github.com/althonos/score-matrices
 Author-email: [e-mail]
-License: GPLv3+
+License: MIT
 Project-URL: Bug Tracker, https://github.com/althonos/scoring-matrices/issues
@@ -52,3 +52,2 @@ Project-URL: Changelog, https://github.com/althonos/scoring-matrices/blob/main/CHANGELOG.md
 [![Source](https://img.shields.io/badge/source-GitHub-303030.svg?maxAge=2678400&style=flat-square)](https://github.com/althonos/scoring-matrices/)
-[![Mirror](https://img.shields.io/badge/mirror-EMBL-009f4d?style=flat-square&maxAge=2678400)](https://git.embl.de/larralde/scoring-matrices/)
 [![Issues](https://img.shields.io/github/issues/althonos/scoring-matrices.svg?style=flat-square&maxAge=600)](https://github.com/althonos/scoring-matrices/issues)
@@ -66,3 +65,3 @@ [![Docs](https://img.shields.io/readthedocs/scoring-matrices/latest?style=flat-square&maxAge=600)](https://scoring-matrices.readthedocs.io)
 
-The `scoring-matrices` packages is a dependency-free, batteries included library
+The `scoring-matrices` package is a dependency-free, batteries included library
 to handle and distribute common substitution matrices:
@@ -69,0 +68,0 @@

README.md

@@ -15,3 +15,2 @@ # 🧬🔠 `scoring-matrices` [![Stars](https://img.shields.io/github/stars/althonos/scoring-matrices.svg?style=social&maxAge=3600&label=Star)](https://github.com/althonos/scoring-matrices/stargazers)
 [![Source](https://img.shields.io/badge/source-GitHub-303030.svg?maxAge=2678400&style=flat-square)](https://github.com/althonos/scoring-matrices/)
-[![Mirror](https://img.shields.io/badge/mirror-EMBL-009f4d?style=flat-square&maxAge=2678400)](https://git.embl.de/larralde/scoring-matrices/)
 [![Issues](https://img.shields.io/github/issues/althonos/scoring-matrices.svg?style=flat-square&maxAge=600)](https://github.com/althonos/scoring-matrices/issues)
@@ -29,3 +28,3 @@ [![Docs](https://img.shields.io/readthedocs/scoring-matrices/latest?style=flat-square&maxAge=600)](https://scoring-matrices.readthedocs.io)
 
-The `scoring-matrices` packages is a dependency-free, batteries included library
+The `scoring-matrices` package is a dependency-free, batteries included library
 to handle and distribute common substitution matrices:
@@ -32,0 +31,0 @@

scoring_matrices.egg-info/PKG-INFO

 Metadata-Version: 2.1
 Name: scoring-matrices
-Version: 0.1.1
+Version: 0.2.0
 Summary: Dependency free, Cython-compatible scoring matrices to use with biological sequences.
@@ -8,3 +8,3 @@ Home-page: https://github.com/althonos/score-matrices
 Author-email: [e-mail]
-License: GPLv3+
+License: MIT
 Project-URL: Bug Tracker, https://github.com/althonos/scoring-matrices/issues
@@ -52,3 +52,2 @@ Project-URL: Changelog, https://github.com/althonos/scoring-matrices/blob/main/CHANGELOG.md
 [![Source](https://img.shields.io/badge/source-GitHub-303030.svg?maxAge=2678400&style=flat-square)](https://github.com/althonos/scoring-matrices/)
-[![Mirror](https://img.shields.io/badge/mirror-EMBL-009f4d?style=flat-square&maxAge=2678400)](https://git.embl.de/larralde/scoring-matrices/)
 [![Issues](https://img.shields.io/github/issues/althonos/scoring-matrices.svg?style=flat-square&maxAge=600)](https://github.com/althonos/scoring-matrices/issues)
@@ -66,3 +65,3 @@ [![Docs](https://img.shields.io/readthedocs/scoring-matrices/latest?style=flat-square&maxAge=600)](https://scoring-matrices.readthedocs.io)
 
-The `scoring-matrices` packages is a dependency-free, batteries included library
+The `scoring-matrices` package is a dependency-free, batteries included library
 to handle and distribute common substitution matrices:
@@ -69,0 +68,0 @@

scoring_matrices/__init__.py

@@ -1,2 +0,2 @@
-__version__ = "0.1.1"
+__version__ = "0.2.0"
 __author__ = "Martin Larralde <[e-mail]>"
@@ -3,0 +3,0 @@ __license__ = "MIT"

scoring_matrices/lib.pxd

@@ -14,9 +14,13 @@ # distutils: language = c
     cdef float** _matrix
+    cdef char*   _alphabet
 
     cdef int _allocate(self, size_t length) except 1 nogil
 
-    cdef const float* data(self) except NULL nogil
-    cdef const float** matrix(self) except NULL nogil    
+    cdef const float* data_ptr(self) except NULL nogil
+    cdef const float** matrix_ptr(self) except NULL nogil    
+    cdef const char* alphabet_ptr(self) except NULL nogil
+    cdef size_t size(self) noexcept nogil
 
     cpdef bint is_integer(self)
+    cpdef bint is_symmetric(self)
     cpdef float min(self)
@@ -23,0 +27,0 @@ cpdef float max(self)

scoring_matrices/lib.pyi

@@ -6,3 +6,2 @@ import typing
 
-
 class ScoringMatrix:
@@ -16,2 +15,18 @@ DEFAULT_ALPHABET: ClassVar[str]
     def from_str(cls: Type[S], text: str, name: Optional[str] = None) -> S: ...
+    @classmethod
+    def from_diagonal(
+        cls: Type[S],
+        diagonal: Iterable[float],
+        mismatch_score: float = 0.0,
+        alphabet: str = DEFAULT_ALPHABET,
+        name: Optional[str] = None,
+    ) -> S: ...
+    @classmethod
+    def from_match_mismatch(
+        cls: Type[S],
+        match_score: float = 1.0,
+        mismatch_score: float = -0.0,
+        alphabet: str = DEFAULT_ALPHABET,
+        name: Optional[str] = None,
+    ) -> S: ...
     def __init__(
@@ -38,4 +53,5 @@ self,
     def is_integer(self) -> bool: ...
+    def is_symmetric(self) -> bool: ...
     def min(self) -> float: ...
     def max(self) -> float: ...
-    def shuffle(self: S, alphabet: str) -> S: ...
+    def shuffle(self: S, alphabet: str) -> S: ...

scoring_matrices/lib.pyx

 # distutils: language = c
 # cython: language_level=3, linetrace=True, binding=True
+"""Dependency free, Cython-compatible scoring matrices for bioinformatics.
+"""
 
@@ -10,3 +12,3 @@ cimport cython
 from libc.stdlib cimport realloc, free
-from libc.string cimport memcpy
+from libc.string cimport memcpy, memset
 
@@ -103,2 +105,65 @@ from .matrices cimport _NAMES, _ALPHABETS, _SIZES, _MATRICES
 
+    @classmethod
+    def from_diagonal(
+        cls, 
+        object diagonal, 
+        float mismatch_score=0.0, 
+        str alphabet not None = DEFAULT_ALPHABET, 
+        str name = None
+    ):
+        """Create a scoring matrix from a diagonal vector.
+
+        Arguments:
+            diagonal (sequence of `float`): The diagonal of the scoring
+                matrix, used to score character matches.
+            mismatch_score (`float`): The mismatch score to use for 
+                every mismatches.
+            alphabet (`str`): The alphabet to use with the scoring matrix.
+            name (`str` or `None`): A name for the scoring matrix, if any.
+
+        Example:
+            >>> matrix = ScoringMatrix.from_diagonal(
+            ...     diagonal=[2, 2, 3, 3],
+            ...     mismatch_score=-3.0,
+            ...     alphabet="ATGC",
+            ... )
+            >>> for row in matrix:
+            ...     print(row)
+            [2.0, -3.0, -3.0, -3.0]
+            [-3.0, 2.0, -3.0, -3.0]
+            [-3.0, -3.0, 3.0, -3.0]
+            [-3.0, -3.0, -3.0, 3.0]
+
+        .. versionadded:: 0.2.0
+
+        """
+        cdef list   matrix = []
+        cdef size_t length = len(alphabet)
+
+        for i, x in enumerate(diagonal):
+            row = [ x if j == i else mismatch_score for j in range(length) ]
+            matrix.append(row)
+        return cls(matrix, alphabet=alphabet, name=name)
+
+    @classmethod
+    def from_match_mismatch(
+        cls,
+        float match_score = 1.0, 
+        float mismatch_score = -1.0,
+        str alphabet not None = DEFAULT_ALPHABET,
+        str name = None,
+    ):
+        """Create a scoring matrix from two match/mismatch scores.
+
+        .. versionadded:: 0.2.0
+
+        """
+        cdef list   matrix = []
+        cdef size_t length = len(alphabet)
+
+        for i in range(length):
+            row = [ match_score if j == i else mismatch_score for j in range(length) ]
+            matrix.append(row)
+        return cls(matrix, alphabet=alphabet, name=name)
+
     # --- Magic methods --------------------------------------------------------
@@ -119,2 +184,27 @@
         """Create a new scoring matrix.
+
+        Arguments:
+            matrix (array-like of `float`): A square matrix with dimensions
+                equal to the ``alphabet`` length, storing the scores for each
+                pair of characters.
+            alphabet (`str`): The alphabet used to index the rows and columns
+                of the scoring matrix.
+            name (`str` or `None`): The name of the scoring matrix, if any.
+
+        Example:
+            >>> matrix = ScoringMatrix(
+            ...     [[91, -114,  -31, -123],
+            ...      [-114, 100, -125, -31],
+            ...      [-31, -125,  100, -114],
+            ...      [-123, -31, -114,  91]],
+            ...     alphabet="ACGT",
+            ...     name="HOXD70",
+            ... )
+
+        Raises:
+            `ValueError`: When the matrix is not a valid matrix or does not
+                match the given alphabet.
+            `MemoryError`: When memory for storing the scores could not be
+                allocated successfully.
+
         """
@@ -137,4 +227,4 @@ cdef ssize_t i
 
-        assert self._data != NULL
-        assert self._matrix != NULL
+        for i, c in enumerate(self.alphabet):
+            self._alphabet[i] = ord(c)
         for i, row in enumerate(matrix):
@@ -255,7 +345,5 @@ if len(row) != size:
         self._data = <float*> realloc(self._data, sizeof(float) * size * size)
-        if self._data is NULL:
-            raise MemoryError("Failed to allocate matrix")
-
         self._matrix = <float**> realloc(self._matrix, sizeof(float*) * size)
-        if self._matrix is NULL:
+        self._alphabet = <char*> realloc(self._alphabet, sizeof(char) * (size + 1))
+        if self._data is NULL or self._matrix is NULL or self._alphabet is NULL:
             raise MemoryError("Failed to allocate matrix")
@@ -267,2 +355,3 @@
             self._matrix[i] = &self._data[i * self._size]
+        memset(self._alphabet, 0, sizeof(char) * (size + 1))
 
@@ -273,3 +362,16 @@ return 0
 
-    cdef const float* data(self) except NULL nogil:
+    cdef size_t size(self) noexcept nogil:
+        """Get the size of the scoring matrix.
+        """
+        return self._size
+
+    cdef const char* alphabet_ptr(self) except NULL nogil:
+        """Get the alphabet of the scoring matrix as a C-string.
+        """
+        if self._alphabet == NULL:
+            with gil:
+                raise RuntimeError("uninitialized scoring matrix")
+        return <const char*> self._alphabet
+
+    cdef const float* data_ptr(self) except NULL nogil:
         """Get the matrix scores as a dense array.
@@ -282,3 +384,3 @@ """
 
-    cdef const float** matrix(self) except NULL nogil:
+    cdef const float** matrix_ptr(self) except NULL nogil:
         """Get the matrix scores as an array of pointers.
@@ -299,2 +401,5 @@ """
 
+        Returns:
+            `bool`: `True` if the matrix only contains integer scores.
+
         Example:
@@ -309,11 +414,10 @@ >>> blosum62 = ScoringMatrix.from_name("BLOSUM62")
         """
-        assert self._data != NULL
+        cdef size_t       i
+        cdef float        x
+        cdef bint         integer = True
+        cdef const float* _data   = self.data_ptr()
 
-        cdef size_t i
-        cdef float  x
-        cdef bint   integer = True
-
         with nogil:
             for i in range(self._nitems):
-                x = self._data[i]
+                x = _data[i]
                 if lrintf(x) != x:
@@ -324,2 +428,24 @@ integer = False
 
+    cpdef bint is_symmetric(self):
+        """Test whether the scoring matrix is symmetric.
+
+        Returns:
+            `bool`: `True` if the matrix is a symmetric matrix.
+
+        .. versionadded:: 0.2.0
+
+        """
+        cdef size_t        i
+        cdef size_t        j
+        cdef bint          symmetric = True
+        cdef const float** _matrix   = self.matrix_ptr()
+
+        with nogil:
+            for i in range(self._nitems):
+                for j in range(i + 1, self._nitems):
+                    if _matrix[i][j] != _matrix[j][i]:
+                        symmetric = False
+                        break
+        return symmetric
+
     cpdef float min(self):
@@ -403,3 +529,3 @@ """Get the minimum score of the scoring matrix.
         for letter in alphabet:
-            row = self[letter]
+            row = self[<str> letter]
             matrix.append([row[j] for j in indices])
@@ -406,0 +532,0 @@

scoring_matrices/tests/test_scoring_matrix.py

@@ -53,2 +53,29 @@ import pickle
 
+    def test_from_diagonal(self):
+        m = ScoringMatrix.from_diagonal([1, 2, 3, 4], 0.0, alphabet="ATGC")
+        self.assertEqual(m[0], [1.0, 0.0, 0.0, 0.0])
+        self.assertEqual(m[1], [0.0, 2.0, 0.0, 0.0])
+        self.assertEqual(m[2], [0.0, 0.0, 3.0, 0.0])
+        self.assertEqual(m[3], [0.0, 0.0, 0.0, 4.0])
+
+        m = ScoringMatrix.from_diagonal([1, 2, 3, 4], -1.0, alphabet="ATGC")
+        self.assertEqual(m[0], [ 1.0, -1.0, -1.0, -1.0])
+        self.assertEqual(m[1], [-1.0,  2.0, -1.0, -1.0])
+        self.assertEqual(m[2], [-1.0, -1.0,  3.0, -1.0])
+        self.assertEqual(m[3], [-1.0, -1.0, -1.0,  4.0])
+
+    def test_from_diagonal_invalid_length(self):
+        self.assertRaises(
+            ValueError,
+            ScoringMatrix.from_diagonal,
+            [ 3, 3, 3, 3, 3, 3 ],
+            alphabet="ATGC"
+        )
+        self.assertRaises(
+            ValueError,
+            ScoringMatrix.from_diagonal,
+            [ 3, 3, 3 ],
+            alphabet="ATGC"
+        )
+
     def test_list(self):
@@ -72,2 +99,7 @@ aa = ScoringMatrix.from_name("BLOSUM50")
 
+    def test_init_empty(self):
+        m = ScoringMatrix([], alphabet="")
+        self.assertEqual(len(m), 0)
+        self.assertFalse(bool(m))
+
     def test_init_invalid_length(self):
@@ -107,2 +139,12 @@ with self.assertRaises(ValueError):
         self.assertEqual(sm1.alphabet, sm2.alphabet)
-        self.assertEqual(list(sm1), list(sm2))
+        self.assertEqual(list(sm1), list(sm2))
+
+    def test_shuffle_invalid_alphabet(self):
+        matrix = ScoringMatrix.from_name("BLOSUM62")
+        self.assertRaises(KeyError, matrix.shuffle, "ARNJOU")
+
+    def test_shuffle_empty(self):
+        matrix = ScoringMatrix.from_name("BLOSUM62")
+        empty = matrix.shuffle("")
+        self.assertEqual(len(empty), 0)
+        self.assertFalse(bool(empty))

setup.cfg

@@ -10,3 +10,3 @@ [metadata]
 long_description_content_type = text/markdown
-license = GPLv3+
+license = MIT
 platform = any
@@ -13,0 +13,0 @@ keywords = bioinformatics, sequence, substitution, matrix, score

setup.py

@@ -138,3 +138,8 @@ import configparser
                 nitems = len(matrix) * len(matrix)
-                dst.write(f"float _MATRIX_{id_}[{nitems}] = {{ { ', '.join(map(repr, itertools.chain.from_iterable(matrix))) } }};\n")
+                dst.write(f"float _MATRIX_{id_}[{nitems}] = {{")
+                for i, item in enumerate(itertools.chain.from_iterable(matrix)):
+                    if i != 0:
+                        dst.write(", ")
+                    dst.write(f"{item!r}F")
+                dst.write("};\n")
 
@@ -141,0 +146,0 @@ dst.write(f"const float* _MATRICES[{len(names) + 1}] = {{")

Alert delta unavailable

Currently unable to show alert delta for PyPI packages.

Improved metrics

Total package byte prevSize

364888

15.62%

Lines of code

541

12.24%