add management command to get data from repositories

eodb/events/management/commands/feed_repositories.py

@@ -0,0 +1,53 @@
+import os
+import subprocess
+import time
+
+from django.core.management.base import BaseCommand
+from django.utils.dateparse import parse_datetime
+
+from eodb.events.models import Commit
+
+author_map = {
+    'fpeters@0d.be': 'fpeters@entrouvert.com',
+    'jschneider@entrouvrt.org': 'jschneider@entrouvert.com',
+    'bdauvergne@ciboulette.entrouvert.og': 'bdauvergne@entrouvert.com',
+    'bdauvergne@entrouvert.org': 'bdauvergne@entrouvert.com',
+    'bdauvergne@moria.entrouvert.org': 'bdauvergne@entrouvert.com',
+    'bdauvergne@lupin.entrouvert.com': 'bdauvergne@entrouvert.com',
+    'benjamin.dauvergne@a716ebb1-153a-0410-b759-cfb97c6a1b53': 'bdauvergne@entrouvert.com',
+    'mikael.ates@gmail.com': 'mates@entrouvert.com',
+    'mikael@xyzstation.(none)': 'mates@entrouvert.com',
+    'serghei@hekla.home': 'smihai@entrouvert.com',
+    'serghei.mihai@devlibre.net': 'smihai@entrouvert.com',
+    'Damien Laniel@localhost': 'dlaniel@entrouvert.com',
+}
+
+class Command(BaseCommand):
+
+    def add_arguments(self, parser):
+        parser.add_argument('repository', nargs='+')
+
+    def handle(self, *args, **options):
+        t0 = time.time()
+        for repository in sorted(options.get('repository')):
+            os.chdir(repository)
+            module = os.path.basename(repository).replace('.git', '')
+            print module
+            try:
+                for line in subprocess.check_output(['git', 'log', '--format=format:%H|%an|%ae|%aI|%cI']).splitlines():
+                    sha, author_name, author_email, author_date, commit_date = line.strip().split('|')
+                    if not author_email:
+                        author_email = author_name + '@entrouvert.com'
+                    author_email = author_map.get(author_email, author_email)
+                    commit, created = Commit.objects.get_or_create(sha=sha, defaults={
+                        'module': module,
+                        'author_datetime': parse_datetime(author_date),
+                        'commit_datetime': parse_datetime(commit_date),
+                        'author_email': author_email,})
+                    if commit.author_email != author_email:
+                        commit.author_email = author_email
+                        commit.save()
+            except subprocess.CalledProcessError as e:
+                print '  ', e
+
+        print time.time() - t0