kallithea Changeset - 90e06f53af8c

Changeset - 90e06f53af8c

Parent rev.

Child rev.

[Not reviewed]

beta

0 2 0

Marcin Kuzminski - 13 years ago 2012-05-28 16:26:47
marcin@python-works.com

Implemented cache-map on main page to save executing select
statements that checks if cache should be invalidated.
It reduces number of executed queries from N which is number of
repos to 1 which is needed to fetch all keys from database.
On pages with large number of repos this could reduce load time by half

2 files changed with 56 insertions and 7 deletions:

rhodecode/model/db.py

rhodecode/model/scm.py

0 comments (0 inline, 0 general)

rhodecode/model/db.py

➞

Show inline comments

@@ @@ -648,108 +648,117 @@ class Repository(Base, BaseModel): @@
     def get_changeset(self, rev=None):
         return get_changeset_safe(self.scm_instance, rev)
     @property
     def tip(self):
         return self.get_changeset('tip')
     @property
     def author(self):
         return self.tip.author
     @property
     def last_change(self):
         return self.scm_instance.last_change
     def comments(self, revisions=None):
         """
         Returns comments for this repository grouped by revisions
         :param revisions: filter query by revisions only
         """
         cmts = ChangesetComment.query()\
             .filter(ChangesetComment.repo == self)
         if revisions:
             cmts = cmts.filter(ChangesetComment.revision.in_(revisions))
         grouped = defaultdict(list)
         for cmt in cmts.all():
             grouped[cmt.revision].append(cmt)
         return grouped
     #==========================================================================
     # SCM CACHE INSTANCE
     #==========================================================================
     @property
     def invalidate(self):
         return CacheInvalidation.invalidate(self.repo_name)
     def set_invalidate(self):
         """
         set a cache for invalidation for this instance
         """
         CacheInvalidation.set_invalidate(self.repo_name)
     @LazyProperty
     def scm_instance(self):
         return self.__get_instance()
     @property
     def scm_instance_cached(self):
     def scm_instance_cached(self, cache_map=None):
         @cache_region('long_term')
         def _c(repo_name):
             return self.__get_instance()
         rn = self.repo_name
         log.debug('Getting cached instance of repo')
         inv = self.invalidate
         if inv is not None:
         if cache_map:
             # get using prefilled cache_map
             invalidate_repo = cache_map[self.repo_name]
             if invalidate_repo:
                 invalidate_repo = (None if invalidate_repo.cache_active
                                    else invalidate_repo)
         else:
             # get from invalidate
             invalidate_repo = self.invalidate
         if invalidate_repo is not None:
             region_invalidate(_c, None, rn)
             # update our cache
-            CacheInvalidation.set_valid(inv.cache_key)
+            CacheInvalidation.set_valid(invalidate_repo.cache_key)
         return _c(rn)
     def __get_instance(self):
         repo_full_path = self.repo_full_path
         try:
             alias = get_scm(repo_full_path)[0]
             log.debug('Creating instance of %s repository' % alias)
             backend = get_backend(alias)
         except VCSError:
             log.error(traceback.format_exc())
             log.error('Perhaps this repository is in db and not in '
                       'filesystem run rescan repositories with '
                       '"destroy old data " option from admin panel')
             return
         if alias == 'hg':
             repo = backend(safe_str(repo_full_path), create=False,
                            baseui=self._ui)
             # skip hidden web repository
             if repo._get_hidden():
                 return
         else:
             repo = backend(repo_full_path, create=False)
         return repo
 class RepoGroup(Base, BaseModel):
     __tablename__ = 'groups'
     __table_args__ = (
         UniqueConstraint('group_name', 'group_parent_id'),
         CheckConstraint('group_id != group_parent_id'),
         {'extend_existing': True, 'mysql_engine':'InnoDB',
          'mysql_charset': 'utf8'},
+    )
     __mapper_args__ = {'order_by': 'group_name'}
     group_id = Column("group_id", Integer(), nullable=False, unique=True, default=None, primary_key=True)
     group_name = Column("group_name", String(length=255, convert_unicode=False, assert_unicode=None), nullable=False, unique=True, default=None)
     group_parent_id = Column("group_parent_id", Integer(), ForeignKey('groups.group_id'), nullable=True, unique=None, default=None)
     group_description = Column("group_description", String(length=10000, convert_unicode=False, assert_unicode=None), nullable=True, unique=None, default=None)
     repo_group_to_perm = relationship('UserRepoGroupToPerm', cascade='all', order_by='UserRepoGroupToPerm.group_to_perm_id')
     users_group_to_perm = relationship('UsersGroupRepoGroupToPerm', cascade='all')
     parent_group = relationship('RepoGroup', remote_side=group_id)
@@ @@ -1027,198 +1036,234 @@ class UsersGroupRepoGroupToPerm(Base, Ba @@
 class Statistics(Base, BaseModel):
     __tablename__ = 'statistics'
     __table_args__ = (
          UniqueConstraint('repository_id'),
          {'extend_existing': True, 'mysql_engine':'InnoDB',
           'mysql_charset': 'utf8'}
+    )
     stat_id = Column("stat_id", Integer(), nullable=False, unique=True, default=None, primary_key=True)
     repository_id = Column("repository_id", Integer(), ForeignKey('repositories.repo_id'), nullable=False, unique=True, default=None)
     stat_on_revision = Column("stat_on_revision", Integer(), nullable=False)
     commit_activity = Column("commit_activity", LargeBinary(1000000), nullable=False)#JSON data
     commit_activity_combined = Column("commit_activity_combined", LargeBinary(), nullable=False)#JSON data
     languages = Column("languages", LargeBinary(1000000), nullable=False)#JSON data
     repository = relationship('Repository', single_parent=True)
 class UserFollowing(Base, BaseModel):
     __tablename__ = 'user_followings'
     __table_args__ = (
         UniqueConstraint('user_id', 'follows_repository_id'),
         UniqueConstraint('user_id', 'follows_user_id'),
         {'extend_existing': True, 'mysql_engine':'InnoDB',
          'mysql_charset': 'utf8'}
+    )
     user_following_id = Column("user_following_id", Integer(), nullable=False, unique=True, default=None, primary_key=True)
     user_id = Column("user_id", Integer(), ForeignKey('users.user_id'), nullable=False, unique=None, default=None)
     follows_repo_id = Column("follows_repository_id", Integer(), ForeignKey('repositories.repo_id'), nullable=True, unique=None, default=None)
     follows_user_id = Column("follows_user_id", Integer(), ForeignKey('users.user_id'), nullable=True, unique=None, default=None)
     follows_from = Column('follows_from', DateTime(timezone=False), nullable=True, unique=None, default=datetime.datetime.now)
     user = relationship('User', primaryjoin='User.user_id==UserFollowing.user_id')
     follows_user = relationship('User', primaryjoin='User.user_id==UserFollowing.follows_user_id')
     follows_repository = relationship('Repository', order_by='Repository.repo_name')
     @classmethod
     def get_repo_followers(cls, repo_id):
         return cls.query().filter(cls.follows_repo_id == repo_id)
 class CacheInvalidation(Base, BaseModel):
     __tablename__ = 'cache_invalidation'
     __table_args__ = (
         UniqueConstraint('cache_key'),
         Index('key_idx', 'cache_key'),
         {'extend_existing': True, 'mysql_engine':'InnoDB',
          'mysql_charset': 'utf8'},
+    )
     cache_id = Column("cache_id", Integer(), nullable=False, unique=True, default=None, primary_key=True)
     cache_key = Column("cache_key", String(length=255, convert_unicode=False, assert_unicode=None), nullable=True, unique=None, default=None)
     cache_args = Column("cache_args", String(length=255, convert_unicode=False, assert_unicode=None), nullable=True, unique=None, default=None)
     cache_active = Column("cache_active", Boolean(), nullable=True, unique=None, default=False)
     def __init__(self, cache_key, cache_args=''):
         self.cache_key = cache_key
         self.cache_args = cache_args
         self.cache_active = False
     def __unicode__(self):
         return u"<%s('%s:%s')>" % (self.__class__.__name__,
                                   self.cache_id, self.cache_key)
     @classmethod
     def clear_cache(cls):
         cls.query().delete()
     @classmethod
     def _get_key(cls, key):
         """
         Wrapper for generating a key, together with a prefix
         :param key:
         """
         import rhodecode
         prefix = ''
         iid = rhodecode.CONFIG.get('instance_id')
         if iid:
             prefix = iid
         return "%s%s" % (prefix, key), prefix, key.rstrip('_README')
     @classmethod
     def get_by_key(cls, key):
         return cls.query().filter(cls.cache_key == key).scalar()
     @classmethod
     def _get_or_create_key(cls, key, prefix, org_key):
         inv_obj = Session.query(cls).filter(cls.cache_key == key).scalar()
         if not inv_obj:
             try:
                 inv_obj = CacheInvalidation(key, org_key)
                 Session.add(inv_obj)
                 Session.commit()
+                #Session.commit()
             except Exception:
                 log.error(traceback.format_exc())
                 Session.rollback()
         return inv_obj
     @classmethod
     def invalidate(cls, key):
         """
         Returns Invalidation object if this given key should be invalidated
         None otherwise. `cache_active = False` means that this cache
         state is not valid and needs to be invalidated
         :param key:
         """
         key, _prefix, _org_key = cls._get_key(key)
         inv = cls._get_or_create_key(key, _prefix, _org_key)
         if inv and inv.cache_active is False:
             return inv
     @classmethod
     def set_invalidate(cls, key):
         """
         Mark this Cache key for invalidation
         :param key:
         """
         key, _prefix, _org_key = cls._get_key(key)
         inv_objs = Session.query(cls).filter(cls.cache_args == _org_key).all()
         log.debug('marking %s key[s] %s for invalidation' % (len(inv_objs),
                                                              _org_key))
         try:
             for inv_obj in inv_objs:
                 if inv_obj:
                     inv_obj.cache_active = False
                 Session.add(inv_obj)
             Session.commit()
         except Exception:
             log.error(traceback.format_exc())
             Session.rollback()
     @classmethod
     def set_valid(cls, key):
         """
         Mark this cache key as active and currently cached
         :param key:
         """
         inv_obj = cls.get_by_key(key)
         inv_obj.cache_active = True
         Session.add(inv_obj)
         Session.commit()
     @classmethod
     def get_cache_map(cls):
         class cachemapdict(dict):
             def __init__(self, *args, **kwargs):
                 fixkey = kwargs.get('fixkey')
                 if fixkey:
                     del kwargs['fixkey']
                 self.fixkey = fixkey
                 super(cachemapdict, self).__init__(*args, **kwargs)
             def __getattr__(self, name):
                 key = name
                 if self.fixkey:
                     key, _prefix, _org_key = cls._get_key(key)
                 if key in self.__dict__:
                     return self.__dict__[key]
                 else:
                     return self[key]
             def __getitem__(self, key):
                 if self.fixkey:
                     key, _prefix, _org_key = cls._get_key(key)
                 try:
                     return super(cachemapdict, self).__getitem__(key)
                 except KeyError:
                     return
         cache_map = cachemapdict(fixkey=True)
         for obj in cls.query().all():
             cache_map[obj.cache_key] = cachemapdict(obj.get_dict())
         return cache_map
 class ChangesetComment(Base, BaseModel):
     __tablename__ = 'changeset_comments'
     __table_args__ = (
         {'extend_existing': True, 'mysql_engine':'InnoDB',
          'mysql_charset': 'utf8'},
+    )
     comment_id = Column('comment_id', Integer(), nullable=False, primary_key=True)
     repo_id = Column('repo_id', Integer(), ForeignKey('repositories.repo_id'), nullable=False)
     revision = Column('revision', String(40), nullable=False)
     line_no = Column('line_no', Unicode(10), nullable=True)
     f_path = Column('f_path', Unicode(1000), nullable=True)
     user_id = Column('user_id', Integer(), ForeignKey('users.user_id'), nullable=False)
     text = Column('text', Unicode(25000), nullable=False)
     modified_at = Column('modified_at', DateTime(), nullable=False, default=datetime.datetime.now)
     author = relationship('User', lazy='joined')
     repo = relationship('Repository')
     @classmethod
     def get_users(cls, revision):
         """
         Returns user associated with this changesetComment. ie those
         who actually commented
         :param cls:
         :param revision:
         """
         return Session.query(User)\
                 .filter(cls.revision == revision)\
                 .join(ChangesetComment.author).all()
 class Notification(Base, BaseModel):
     __tablename__ = 'notifications'
     __table_args__ = (
         {'extend_existing': True, 'mysql_engine':'InnoDB',
          'mysql_charset': 'utf8'},
+    )
     TYPE_CHANGESET_COMMENT = u'cs_comment'
     TYPE_MESSAGE = u'message'
     TYPE_MENTION = u'mention'
     TYPE_REGISTRATION = u'registration'
     notification_id = Column('notification_id', Integer(), nullable=False, primary_key=True)
     subject = Column('subject', Unicode(512), nullable=True)
     body = Column('body', Unicode(50000), nullable=True)

rhodecode/model/scm.py

➞

Show inline comments

@@ @@ -32,98 +32,102 @@ from rhodecode.lib.vcs import get_backen @@
 from rhodecode.lib.vcs.exceptions import RepositoryError
 from rhodecode.lib.vcs.utils.lazy import LazyProperty
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode import BACKENDS
 from rhodecode.lib import helpers as h
 from rhodecode.lib.utils2 import safe_str, safe_unicode
 from rhodecode.lib.auth import HasRepoPermissionAny, HasReposGroupPermissionAny
 from rhodecode.lib.utils import get_repos as get_filesystem_repos, make_ui, \
     action_logger, EmptyChangeset, REMOVED_REPO_PAT
 from rhodecode.model import BaseModel
 from rhodecode.model.db import Repository, RhodeCodeUi, CacheInvalidation, \
     UserFollowing, UserLog, User, RepoGroup
 log = logging.getLogger(__name__)
 class UserTemp(object):
     def __init__(self, user_id):
         self.user_id = user_id
     def __repr__(self):
         return "<%s('id:%s')>" % (self.__class__.__name__, self.user_id)
 class RepoTemp(object):
     def __init__(self, repo_id):
         self.repo_id = repo_id
     def __repr__(self):
         return "<%s('id:%s')>" % (self.__class__.__name__, self.repo_id)
 class CachedRepoList(object):
     def __init__(self, db_repo_list, repos_path, order_by=None):
         self.db_repo_list = db_repo_list
         self.repos_path = repos_path
         self.order_by = order_by
         self.reversed = (order_by or '').startswith('-')
     def __len__(self):
         return len(self.db_repo_list)
     def __repr__(self):
         return '<%s (%s)>' % (self.__class__.__name__, self.__len__())
     def __iter__(self):
         # pre-propagated cache_map to save executing select statements
         # for each repo
         cache_map = CacheInvalidation.get_cache_map()
         for dbr in self.db_repo_list:
             scmr = dbr.scm_instance_cached
+            scmr = dbr.scm_instance_cached(cache_map)
             # check permission at this level
             if not HasRepoPermissionAny(
                 'repository.read', 'repository.write', 'repository.admin'
             )(dbr.repo_name, 'get repo check'):
                 continue
             if scmr is None:
                 log.error(
                     '%s this repository is present in database but it '
                     'cannot be created as an scm instance' % dbr.repo_name
+                )
                 continue
             last_change = scmr.last_change
             tip = h.get_changeset_safe(scmr, 'tip')
             tmp_d = {}
             tmp_d['name'] = dbr.repo_name
             tmp_d['name_sort'] = tmp_d['name'].lower()
             tmp_d['description'] = dbr.description
             tmp_d['description_sort'] = tmp_d['description']
             tmp_d['last_change'] = last_change
             tmp_d['last_change_sort'] = time.mktime(last_change.timetuple())
             tmp_d['tip'] = tip.raw_id
             tmp_d['tip_sort'] = tip.revision
             tmp_d['rev'] = tip.revision
             tmp_d['contact'] = dbr.user.full_contact
             tmp_d['contact_sort'] = tmp_d['contact']
             tmp_d['owner_sort'] = tmp_d['contact']
             tmp_d['repo_archives'] = list(scmr._get_archives())
             tmp_d['last_msg'] = tip.message
             tmp_d['author'] = tip.author
             tmp_d['dbrepo'] = dbr.get_dict()
             tmp_d['dbrepo_fork'] = dbr.fork.get_dict() if dbr.fork else {}
             yield tmp_d
 class GroupList(object):
     def __init__(self, db_repo_group_list):
         self.db_repo_group_list = db_repo_group_list
     def __len__(self):
         return len(self.db_repo_group_list)
     def __repr__(self):
         return '<%s (%s)>' % (self.__class__.__name__, self.__len__())

0 comments (0 inline, 0 general)