More flexibility with wildcards in selection (#2551)

* Fixes #2436 
* Selection strings changed to use fnmatch. This now allows for more flexible wildcard usage as well as for using multiple wildcards at once.
* Added two new tests to match the new functionality
* New doc section on wildcards
* Remove test for multiple wildcards
* Update in AUTHORS
* Update CHANGELOG

package/AUTHORS

@@ -128,6 +128,7 @@ Chronological list of authors
 2020
   - Charlie Cook
   - Yuanyu Chang
+  - Ivan Hristov
   - Michael Quevillon
   - Hao Tian
 

package/CHANGELOG

@@ -15,7 +15,7 @@ The rules for this file:
 ------------------------------------------------------------------------------
 mm/dd/yy richardjgowers, kain88-de, lilyminium, p-j-smith, bdice, joaomcteixeira,
          PicoCentauri, davidercruz, jbarnoud, RMeli, IAlibay, mtiberti, CCook96,
-         Yuan-Yu, xiki-tempula, HTian1997
+         Yuan-Yu, xiki-tempula, HTian1997, Iv-Hristov
 
   * 0.21.0
 
@@ -57,6 +57,7 @@ Fixes
   * Added parmed to setup.py
 
 Enhancements
+  * Changed selection wildcards to support multiple wildcards (#2436)
   * Added coordinate reader and writer for NAMD binary coordinate format (PR #2485)
   * Improved ClusterCollection and Cluster string representations (Issue #2464)
   * XYZ parser store elements attribute (#2420) and XYZ write uses the elements

package/MDAnalysis/core/selection.py

@@ -44,6 +44,7 @@
 
 import collections
 import re
+import fnmatch
 import functools
 import warnings
 
@@ -499,8 +500,8 @@ def apply(self, group):
 class StringSelection(Selection):
     """Selections based on text attributes
 
-    Supports the use of one wildcard at the start, 
-    end, and middle of strings
+    .. versionchanged:: 1.0.0
+        Supports multiple wildcards, based on fnmatch
     """
     def __init__(self, parser, tokens):
         vals = grab_not_keywords(tokens)
@@ -512,16 +513,8 @@ def __init__(self, parser, tokens):
     def apply(self, group):
         mask = np.zeros(len(group), dtype=np.bool)
         for val in self.values:
-            if val.count('*') > 1:
-                raise SelectionError('Can only use one wildcard in a string')
-            wc_pos = val.find('*')
-            if wc_pos == -1:  # No wildcard found
-                mask |= getattr(group, self.field) == val
-            else:
-                values = getattr(group, self.field).astype(np.str_)
-                mask |= np.char.startswith(values, val[:wc_pos])
-                mask &= np.char.endswith(values, val[wc_pos+1:])
-
+            values = getattr(group, self.field)
+            mask |= [fnmatch.fnmatch(x, val) for x in values]
         return group[mask].unique
 
 

package/doc/sphinx/source/documentation_pages/selections.rst

@@ -46,11 +46,6 @@ selection parser. The following applies to all selections:
   necessary).
 * Selections are parsed left to right and parentheses can be used for
   grouping.
-* Currently, a single wildcard ``*`` character can be at the start, middle, or 
-  end of a string for pattern matching. For example, ``GL*`` selects
-  all strings that start with "GL" such as "GLU", "GLY", "GLX29", "GLN". 
-  ``resname *N`` selects all residue names that end in "N", such as "ASN" and 
-  "GLN".
 
 
 Simple selections
@@ -104,6 +99,27 @@ moltype *molecule-type*
     select by molecule type, e.g. ``moltype Protein_A``. At the moment, only
     the TPR format defines the molecule type.
 
+Pattern matching
+----------------
+
+The pattern matching notation described below is used to specify 
+patterns for matching strings (based on :mod:`fnmatch`):
+
+``?`` 
+    Is a pattern that will match any single character. For example,
+    ``resname T?R`` selects residues named "TYR" and "THR".
+``*`` 
+    Is a pattern that will match multiple characters.  For example,
+    ``GL*`` selects all strings that start with "GL" such as "GLU",
+    "GLY", "GLX29", "GLN".
+``[seq]``
+    Would match any character in seq. For example, "resname GL[NY]" 
+    selects all residues named "GLN" or "GLY" but would not select
+    "GLU".
+``[!seq]``
+    Would match any character not in seq. For example, "resname GL[!NY]"
+    would match residues named "GLU" but would not match "GLN" or "GLY".
+
 Boolean
 -------
 

testsuite/MDAnalysisTests/core/test_atomselections.py

@@ -364,22 +364,22 @@ def test_global(self, universe):
         ag2 = ag.select_atoms("around 4 global backbone")
         assert_equal(ag2.indices, ag1.indices)
 
-    def test_wildcard_middle_selection(self, universe):
-        ag = universe.select_atoms('resname TYR or resname THR')
-        ag_wild = universe.select_atoms('resname T*R')
-        assert ag == ag_wild
-
-    def test_wildcard_start_selection(self, universe):
-        ag = universe.select_atoms('resname ASN GLN')
-        ag_wild = universe.select_atoms('resname *N')
-        assert ag == ag_wild
-
-    def test_wildcard_terminal_selection(self, universe):
-        ag = universe.select_atoms('resname ASN ASP')
-        ag_wild = universe.select_atoms('resname AS*')
+    @pytest.mark.parametrize('selstring, wildstring', [
+        ('resname TYR THR', 'resname T*R'),
+        ('resname ASN GLN', 'resname *N'),
+        ('resname ASN ASP', 'resname AS*'),
+        ('resname TYR THR', 'resname T?R'),
+        ('resname ASN ASP HSD', 'resname *S?'),
+        ('resname LEU LYS', 'resname L**'),
+        ('resname MET', 'resname *M*'),
+        ('resname GLN GLY', 'resname GL[NY]'),
+        ('resname GLU', 'resname GL[!NY]'),
+    ])
+    def test_wildcard_selection(self, universe, selstring, wildstring):
+        ag = universe.select_atoms(selstring)
+        ag_wild = universe.select_atoms(wildstring)
         assert ag == ag_wild
 
-
 class TestSelectionsAMBER(object):
     @pytest.fixture()
     def universe(self):
@@ -927,7 +927,6 @@ def universe():
         'index or protein',
         'prop mass < 4.0 hello',  # unused token
         'prop mass > 10. and group this',  # missing group
-        'resname E*Y*Z',  # >1 wildcards
     ])
     def test_selection_fail(self, selstr, universe):
         with pytest.raises(SelectionError):