NewickIO: parse branch lengths in scientific notation (Issue #220)

Included a unit test for it, too.

Thanks to mathli for reporting this bug and offering a fix.

Bio/Phylo/NewickIO.py

@@ -20,13 +20,13 @@
 class NewickError(Exception):
     """Exception raised when Newick object construction cannot continue."""
     pass
-    
-    
+
+
 tokens = [
     (r"\(",                         'open parens'),
     (r"\)",                         'close parens'),
     (r"[^\s\(\)\[\]\'\:\;\,]+",     'unquoted node label'),
-    (r"\:[0-9]*\.?[0-9]+",          'edge length'),
+    (r"\:[0-9]*\.?[0-9]+([eE][+-]?[0-9]+)?", 'edge length'),
     (r"\,",                         'comma'),
     (r"\[(\\.|[^\]])*\]",           'comment'),
     (r"\'(\\.|[^\'])*\'",           'quoted node label'),
@@ -70,11 +70,11 @@ def _parse_confidence(text):
         # assert 0 <= current_clade.confidence <= 1
     except ValueError:
         return None
-        
-        
+
+
 def _format_comment(text):
     return '[%s]' % (text.replace('[', '\\[').replace(']', '\\]'))
-    
+
 def _get_comment(clade):
     if hasattr(clade, 'comment') and clade.comment:
         return _format_comment(str(clade.comment))
@@ -207,12 +207,13 @@ def new_clade(self, parent=None):
     def process_clade(self, clade):
         """Final processing of a parsed clade. Removes the node's parent and
         returns it."""
-        if (clade.name and not (self.values_are_confidence or self.comments_are_confidence)
+        if (clade.name and not (self.values_are_confidence or
+                                self.comments_are_confidence)
             and clade.confidence is None):
             clade.confidence = _parse_confidence(clade.name)
             if not clade.confidence is None:
                 clade.name = None
-            
+
         if hasattr(clade, 'parent'):
             parent = clade.parent
             parent.clades.append(clade)

Tests/test_Phylo.py

@@ -67,6 +67,13 @@ def test_newick_write(self):
                 if c is not None)
         self.assertEqual(internal_names, set(('E', 'F')))
 
+    def test_newick_read_scinot(self):
+        """Parse Newick branch lengths in scientific notation."""
+        tree = Phylo.read(StringIO("(foo:1e-1,bar:0.1)"), 'newick')
+        clade_a = tree.clade[0]
+        self.assertEqual(clade_a.name, 'foo')
+        self.assertAlmostEqual(clade_a.branch_length, 0.1)
+
     def test_format_branch_length(self):
         """Custom format string for Newick branch length serialization."""
         tree = Phylo.read(StringIO('A:0.1;'), 'newick')