Added performance evaluations.

 * Raw timings for the pairwise alignment of the first two sequences of
   each balibase file
 * Timings for alignment of mouse and human titin, with a timeout
   controlled by `UNIT_TIMEOUT`

.gitignore

@@ -2,3 +2,4 @@
 balibase/
 __pycache__/
 balibase.zip
+atlas

.gitlab-ci.yml

@@ -6,4 +6,8 @@ cache:
                 - balibase/
 test:
         script:
-                - python3 -m unittest alignementseq_tests.py
+                - python3 -m unittest alignementseq_tests
+
+test_performance:
+        script:
+                - python3 -m unittest alignementseq_tests.PerformanceTestCase

alignementseq_tests.py

@@ -4,14 +4,23 @@ import unittest
 import urllib.request
 import os
 import zipfile
+import warnings
+import timeit
+import multiprocessing
+import pandas as pd
 
 import Bio.SeqIO
 import Bio.pairwise2
+import Bio.Entrez
+
 from Bio.SeqRecord import SeqRecord
 from Bio.SubsMat.MatrixInfo import blosum62
 
 class AlignmentSeqTestCase(unittest.TestCase):
     def setUp(self):
+        
+        self.unit_timeout = int(os.environ.get("UNIT_TIMEOUT") or 60)
+        
         balibase_zippath = "balibase.zip"
         self.balibase_path = "balibase"
         testfile_path = os.path.join(
@@ -51,7 +60,9 @@ class AlignmentSeqTestCase(unittest.TestCase):
             )
 
     def test_simple_align(self):
-        """Test alignments with the simplest metric. As there can be a huge number of
+        """Test alignments with the simplest metric.
+        
+        As there can be a huge number of
         solutions, we check only that we got the right score, and one valid
         alignment.
         """
@@ -82,6 +93,9 @@ class AlignmentSeqTestCase(unittest.TestCase):
                 raise
 
     def test_blosum_align(self):
+        """Tests alignments with blosum but no gap extension.
+        """
+
         from alignementseq import align, vec_align
         
         for s1, s2, filename in self.get_dataset_heads():
@@ -107,6 +121,63 @@ class AlignmentSeqTestCase(unittest.TestCase):
                     print(ex_r2)
                     raise
 
+class PerformanceTestCase(AlignmentSeqTestCase):
+    """Performance tests, slow.
+
+    Excluded from default test suite, run it with
+    ``python -m unittest alignmentseq_tests.PerformanceTestCase``
+    """
+
+    def setUp(self):
+        """Loads titin sequences.
+        """
+        super().setUp()
+
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            self.titin_human, self.titin_mouse = Bio.SeqIO.parse(
+                    Bio.Entrez.efetch(db="protein", id=["CAA62188","EDL27217"],
+                        rettype="gp", retmode="text"),
+            "genbank")
+
+    def test_performance(self):
+        from alignementseq import align, vec_align
+
+        methods = (vec_align, align)
+        times = pd.DataFrame(columns=("length1", "length2", "method", "time"))
+        for s1, s2, filename in self.get_dataset_heads():
+            for method in methods:
+                def to_time():
+                    method(s1, s2)
+                t = timeit.Timer(to_time).timeit(1)
+                times.loc[len(times)] = (
+                        len(s1.seq), len(s2.seq), method.__name__, t
+                        )
+        print(times)
+
+    def test_titin(self):
+        from alignementseq import align, vec_align
+
+        for method in vec_align, align:
+            def to_time():
+                method(self.titin_human, self.titin_mouse)
+            def run():
+                t = timeit.Timer(to_time).timeit(1)
+                print("{}: {}".format(method.__name__, t))
+            p = multiprocessing.Process(target=run)
+            p.start()
+            p.join(self.unit_timeout)
+            # Note: if the task eats up your memory, it can take a while for
+            # it to terminate if it times out.
+            if p.is_alive():
+                p.terminate()
+                p.join()
+                print("{}: timeout".format(method.__name__))
+            
+
+def load_tests(loader, standard_tests, pattern):
+    return unittest.defaultTestLoader.loadTestsFromTestCase(AlignmentSeqTestCase)
+
 if __name__ == '__main__':
-    unittest.main()
+    unittest.main(verbosity=2)
 

requirements.txt

 biopython==1.73
+pandas==0.23.3