Added soundex to utils (along with a test) and to FunCommands.

plugins/FunCommands.py

@@ -433,6 +433,20 @@ class FunCommands(callbacks.Privmsg):
         (s1, s2) = privmsgs.getArgs(args, needed=2)
         irc.reply(msg, str(utils.distance(s1, s2)))
 
+    def soundex(self, irc, msg, args):
+        """<string> [<length>]
+
+        Returns the Soundex hash to a given length.  The length defaults to
+        4, since that's the standard length for a soundex hash.  For unlimited
+        length, use 0.
+        """
+        (s, length) = privmsgs.getArgs(args, optional=1)
+        if length:
+            length = int(length)
+        else:
+            length = 4
+        irc.reply(msg, utils.soundex(s, length))
+
     modulechars = '%s%s%s' % (string.ascii_letters, string.digits, '_.')
     def pydoc(self, irc, msg, args):
         """<python function>

src/utils.py

@@ -37,6 +37,7 @@ from __future__ import generators
 
 from fix import *
 
+import string
 import sgmllib
 import htmlentitydefs
 
@@ -173,6 +174,23 @@ def distance(s, t):
                 cost = 1
             d[i][j] = min(d[i-1][j]+1, d[i][j-1]+1, d[i-1][j-1]+cost)
     return d[n][m]
-    
+
+_soundextrans = string.maketrans(string.ascii_uppercase,
+                                 '01230120022455012623010202')
+_notUpper = string.ascii.translate(string.ascii, string.ascii_uppercase)
+def soundex(s, length=4):
+    assert s
+    s = s.upper() # Make everything uppercase.
+    firstChar = s[0] # Save the first character.
+    s = s.translate(string.ascii, _notUpper) # Delete non-letters.
+    s = s.translate(_soundextrans) # Convert to soundex numbers.
+    s = s.lstrip(s[0]) # Remove all repeated first characters.
+    L = [firstChar]
+    for c in s:
+        if c != L[-1]:
+            L.append(c)
+    L = [c for c in L if c != '0'] + ['0', '0', '0']
+    s = ''.join(L)
+    return length and s[:length] or s
 
 # vim:set shiftwidth=4 tabstop=8 expandtab textwidth=78:

test/utils_test.py

@@ -77,3 +77,21 @@ class UtilsTest(unittest.TestCase):
         self.assertEqual(d['ab'], 'abc')
         self.assertEqual(d['fool'], 'fool')
         self.assertEqual(d['foo'], 'foo')
+
+    def testSoundex(self):
+        L = [('Euler', 'E460'),
+             ('Ellery', 'E460'),
+             ('Gauss', 'G200'),
+             ('Ghosh', 'G200'),
+             ('Hilbert', 'H416'),
+             ('Heilbronn', 'H416'),
+             ('Knuth', 'K530'),
+             ('Kant', 'K530'),
+             ('Lloyd', 'L300'),
+             ('Ladd', 'L300'),
+             ('Lukasiewicz', 'L222'),
+             ('Lissajous', 'L222')]
+        for (name, key) in L:
+            soundex = utils.soundex(name)
+            self.assertEqual(soundex, key,
+                             '%s was %s, not %s' % (name, soundex, key))