Add a rematch_packager management command

This allows quick resolution of all unmatched packages, especially after
tweaking the way find_user works.

Signed-off-by: Dan McGee <dan@archlinux.org>

3 files changed, 80 insertions, 4 deletions

diff --git a/devel/management/commands/rematch_packager.py b/devel/management/commands/rematch_packager.py
new file mode 100644
index 00000000..ba6e6a54
--- /dev/null
+++ b/devel/management/commands/rematch_packager.py
@@ -0,0 +1,64 @@
+# -*- coding: utf-8 -*-
+"""
+rematch_packager command
+
+Match all packages with a packager_str but NULL packager_id to a packager if we
+can find one.
+
+Usage: ./manage.py rematch_packager
+"""
+
+from django.core.management.base import NoArgsCommand
+
+import sys
+import logging
+
+from devel.utils import UserFinder
+from main.models import Package
+
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s -> %(levelname)s: %(message)s',
+    datefmt='%Y-%m-%d %H:%M:%S',
+    stream=sys.stderr)
+logger = logging.getLogger()
+
+class Command(NoArgsCommand):
+    help = "Runs a check on all active mirror URLs to determine if they are reachable via IPv4 and/or v6."
+
+    def handle_noargs(self, **options):
+        v = int(options.get('verbosity', None))
+        if v == 0:
+            logger.level = logging.ERROR
+        elif v == 1:
+            logger.level = logging.INFO
+        elif v == 2:
+            logger.level = logging.DEBUG
+
+        return match_packager()
+
+def match_packager():
+    finder = UserFinder()
+    logger.info("getting all unmatched packages")
+    package_count = matched_count = 0
+    unknown = set()
+
+    for package in Package.objects.filter(packager__isnull=True):
+        logger.debug("package %s, packager string %s",
+                package.pkgname, package.packager_str)
+        package_count += 1
+        user = finder.find(package.packager_str)
+        if user:
+            package.packager = user
+            logger.debug("  found user %s" % user.username)
+            package.save()
+            matched_count += 1
+        else:
+            unknown.add(package.packager_str)
+
+    logger.info("%d packages checked, %d newly matched",
+            package_count, matched_count)
+    logger.debug("unknown packagers:\n%s",
+            "\n".join(unknown))
+
+# vim: set ts=4 sw=4 et:
diff --git a/devel/tests.py b/devel/tests.py
index c982e502..36691179 100644
--- a/devel/tests.py
+++ b/devel/tests.py
@@ -1,7 +1,7 @@
 from django.test import TestCase
 
 from django.contrib.auth.models import User
-from devel.utils import UserFinder 
+from devel.utils import UserFinder
 from main.models import UserProfile
 
 class DevelTest(TestCase):
@@ -87,6 +87,16 @@ class FindUserTest(TestCase):
         self.assertEqual(self.user3,
                 self.finder.find("Bob Jones <bjones AT Arch Linux DOT org>"))
 
+    def test_by_invalid(self):
+        self.assertEqual(self.user1,
+                self.finder.find("Joe User <user1@example.com"))
+        self.assertEqual(self.user1,
+                self.finder.find("Joe 'nickname' User <user1@example.com"))
+        self.assertEqual(self.user1,
+                self.finder.find("Joe \"nickname\" User <user1@example.com"))
+        self.assertEqual(self.user1,
+                self.finder.find("Joe User <joe@differentdomain.com"))
+
     def test_cache(self):
         # simply look two of them up, but then do it repeatedly
         for i in range(50):
diff --git a/devel/utils.py b/devel/utils.py
index 3a6ad699..6bc52c89 100644
--- a/devel/utils.py
+++ b/devel/utils.py
@@ -88,10 +88,12 @@ class UserFinder(object):
             return None
         if userstring in self.cache:
             return self.cache[userstring]
-        matches = re.match(r'^([^<]+)? ?<([^>]*)>', userstring)
+
+        name = email = None
+
+        matches = re.match(r'^([^<]+)? ?<([^>]*)>?', userstring)
         if not matches:
-            name = userstring
-            email = None
+            name = userstring.strip()
         else:
             name = matches.group(1)
             email = matches.group(2)