Speed up parsing by about 2x on average, fix __getitem__(), and use setuptools if available.

PeterScott · PeterScott · commit 7d8aea828154 · 2012-01-12T17:58:17.000-08:00
diff --git a/nameparser/__init__.py b/nameparser/__init__.py
@@ -1,4 +1,4 @@
-VERSION = (0, 1, 3)
+VERSION = (0, 1, 4)
 __version__ = '.'.join(map(str, VERSION))
 __author__ = "Derek Gulbranson"
 __author_email__ = 'derek73@gmail.com'
diff --git a/nameparser/constants.py b/nameparser/constants.py
@@ -5,7 +5,7 @@
 re_mac = re.compile(r'^(ma?c)(\w)', re.I)
 re_initial = re.compile(r'^(\w\.|[A-Z])?$')
 
-TITLES = (
+TITLES = set([
     'dr','doctor','miss','misses','mr','mister','mrs','ms','sir',
     'rev','madam','madame','AB','2ndLt','Amn','1stLt','A1C','Capt','SrA','Maj',
     'SSgt','LtCol','TSgt','Col','BrigGen','1stSgt','MajGen','SMSgt','LtGen',
@@ -19,7 +19,7 @@
     'Gen','SgtMaj','SgtMajMC','WO-1','CWO-2','CWO-3','CWO-4','CWO-5','ENS','SA',
     'LTJG','SN','LT','PO3','LCDR','PO2','CDR','PO1','CAPT','CPO','RDML','SCPO',
     'RADM','MCPO','VADM','MCPON','ADM','FADM','WO1','CWO2','CWO3','CWO4','CWO5'
-)
+])
 
 # QUESTIONABLE_TITLES could be last names or they could be titles.
 # TODO: need to find best way to deal with these, not doing anything special yet.
@@ -28,20 +28,20 @@
 
 # PUNC_TITLES could be names or titles, but if they have period at the end they're a title
 PUNC_TITLES = ('hon.',)
-PREFICES = (
+PREFICES = set([
     'abu','bon','ben','bin','da','dal','de','del','der','de','di','e','ibn',
     'la','le','san','st','ste','van','vel','von'
-)
-SUFFICES = (
+])
+SUFFICES = set([
     'esq','esquire','jr','sr','2','i','ii','iii','iv','v','clu','chfc',
     'cfp','md','phd'
-)
+])
 CAPITALIZATION_EXCEPTIONS = {
     'ii': 'II',
     'iii': 'III',
     'iv': 'IV',
     'md': 'M.D.',
     'phd': 'Ph.D.'
 }
-CONJUNCTIONS = ('&', 'and', 'et', 'e', 'und', 'y')
+CONJUNCTIONS = set(['&', 'and', 'et', 'e', 'und', 'y'])
 
diff --git a/nameparser/parser.py b/nameparser/parser.py
@@ -113,7 +113,7 @@ def __ne__(self, other):
         return not unicode(self).lower() == unicode(other).lower()
     
     def __getitem__(self, key):
-        return [getattr(self, x) for x in self.members[key]]
+        return getattr(self, self.members[key])
     
     def next(self):
         if self.count >= len(self.members):
diff --git a/setup.py b/setup.py
@@ -1,5 +1,8 @@
 #!/usr/bin/env python
-from distutils.core import setup
+try:
+  from setuptools import setup
+except ImportError:
+  from distutils.core import setup
 import nameparser
 import os
 

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-VERSION = (0, 1, 3)`
	`1`	`+VERSION = (0, 1, 4)`
`2`	`2`	`__version__ = '.'.join(map(str, VERSION))`
`3`	`3`	`__author__ = "Derek Gulbranson"`
`4`	`4`	`__author_email__ = 'derek73@gmail.com'`