.travis.yml

@@ -11,8 +11,8 @@ python:
     - 'pypy'
     - 'pypy3'
 install:
-    - pip wheel -f wheelhouse coverage biopython cython pysam pyvcf || true
-    - pip install -f wheelhouse biopython cython pysam pyfasta coverage pyvcf || true
+    - pip wheel -f wheelhouse coverage biopython cython pysam pyvcf numpy || true
+    - pip install -f wheelhouse biopython cython pysam pyfasta coverage pyvcf numpy || true
     - python setup.py install
     - if [ ! -f samtools-1.2 ]; then curl -sL https://github.com/samtools/samtools/releases/download/1.2/samtools-1.2.tar.bz2 | tar -xjv; fi
     - cd samtools-1.2

README.rst

@@ -261,6 +261,17 @@ Sequence names are truncated on any whitespace. This is a limitation of the inde
     gi|557361099|gb|KF435150.1| Homo sapiens MDM4 protein variant Y (MDM4) mRNA, complete cds, alternatively spliced
     gi|557361097|gb|KF435149.1| Homo sapiens MDM4 protein variant G (MDM4) mRNA, complete cds
 
+Records can be accessed efficiently as numpy arrays:
+
+.. code:: python
+
+    # new in v0.5.4
+    >>> from pyfaidx import Fasta
+    >>> import numpy as np
+    >>> genes = Fasta('tests/data/genes.fasta')
+    >>> np.asarray(genes['NM_001282543.1'])
+    array(['C', 'C', 'C', ..., 'A', 'A', 'A'], dtype='|S1')
+
 Sequence can be buffered in memory using a read-ahead buffer
 for fast sequential access:
 

debian/changelog

+python-pyfaidx (0.5.4-1) unstable; urgency=medium
+
+  * New upstream version
+
+ -- Andreas Tille <tille@debian.org>  Tue, 29 May 2018 08:57:08 +0200
+
 python-pyfaidx (0.5.3.1-1) unstable; urgency=medium
 
   * New upstream version

pyfaidx/__init__.py

@@ -5,6 +5,7 @@ Fasta file -> Faidx -> Fasta -> FastaRecord -> Sequence
 
 from __future__ import division
 import os
+import sys
 from os.path import getmtime
 from six import PY2, PY3, string_types, integer_types
 from six.moves import zip_longest
@@ -19,9 +20,12 @@ import warnings
 from math import ceil
 from threading import Lock
 
+if sys.version_info > (3, ):
+    buffer = memoryview
+
 dna_bases = re.compile(r'([ACTGNactgnYRWSKMDVHBXyrwskmdvhbx]+)')
 
-__version__ = '0.5.3.1'
+__version__ = '0.5.4'
 
 
 class KeyFunctionError(ValueError):
@@ -762,6 +766,7 @@ class Faidx(object):
 
 class FastaRecord(object):
     __slots__ = ['name', '_fa']
+
     def __init__(self, name, fa):
         self.name = name
         self._fa = fa
@@ -887,6 +892,16 @@ class FastaRecord(object):
         """ Read the actual defline from self._fa.faidx mdshw5/pyfaidx#54 """
         return self._fa.faidx.get_long_name(self.name)
 
+    @property
+    def __array_interface__(self):
+        """ Implement numpy array interface for issue #139"""
+        return {
+            'shape': (len(self), ),
+            'typestr': '|S1',
+            'version': 3,
+            'data': buffer(str(self).encode('ascii'))
+        }
+
 
 class MutableFastaRecord(FastaRecord):
     def __init__(self, name, fa):

tests/test_FastaRecord.py

@@ -9,6 +9,7 @@ from difflib import Differ
 path = os.path.dirname(__file__)
 os.chdir(path)
 
+
 class TestFastaRecord(TestCase):
     def setUp(self):
         pass
@@ -28,8 +29,9 @@ class TestFastaRecord(TestCase):
         reference_upper = Fasta(filename, sequence_always_upper=True)
         reference_normal = Fasta(filename)
         os.remove('data/genes.fasta.lower.fai')
-        assert reference_upper['gi|557361099|gb|KF435150.1|'][1:100].seq == reference_normal['gi|557361099|gb|KF435150.1|'][1:100].seq.upper()
-
+        assert reference_upper['gi|557361099|gb|KF435150.1|'][
+            1:100].seq == reference_normal['gi|557361099|gb|KF435150.1|'][
+                1:100].seq.upper()
 
     def test_long_names(self):
         """ Test that deflines extracted using FastaRecord.long_name are
@@ -60,7 +62,8 @@ class TestFastaRecord(TestCase):
                         line_len = len(line)
                         fasta_uniform_len.write(line)
                     elif line_len > len(line):
-                        fasta_uniform_len.write(line.rstrip() + b'N' * (line_len - len(line)) + b'\n')
+                        fasta_uniform_len.write(line.rstrip() + b'N' *
+                                                (line_len - len(line)) + b'\n')
                     else:
                         fasta_uniform_len.write(line)
         fasta = Fasta('data/issue_62.fa', as_raw=True)
@@ -98,6 +101,14 @@ class TestFastaRecord(TestCase):
         os.remove('data/padded.fasta')
         os.remove('data/padded.fasta.fai')
 
+    def test_numpy_array(self):
+        """ Test the __array_interface__ """
+        import numpy
+        filename = "data/genes.fasta.lower"
+        reference = Fasta(filename)
+        np_array = numpy.asarray(reference[0])
+        assert isinstance(np_array, numpy.ndarray)
+
 
 class TestMutableFastaRecord(TestCase):
     def setUp(self):
@@ -124,7 +135,8 @@ class TestMutableFastaRecord(TestCase):
         fasta = Fasta('data/genes.fasta', mutable=False)
         chunk = fasta['gi|557361099|gb|KF435150.1|'][0:100]
         mutable['gi|557361099|gb|KF435150.1|'][0:100] = chunk.seq
-        assert str(fasta['gi|557361099|gb|KF435150.1|']) == str(mutable['gi|557361099|gb|KF435150.1|'])
+        assert str(fasta['gi|557361099|gb|KF435150.1|']) == str(
+            mutable['gi|557361099|gb|KF435150.1|'])
 
     def test_mutate_fasta_to_N(self):
         mutable = Fasta('data/genes_mutable.fasta', mutable=True)