Changeset - eaa887c6c0af
[Not reviewed]
beta
0 2 0
Marcin Kuzminski - 13 years ago 2013-01-30 22:30:52
marcin@python-works.com
added recursion limit for stats gathering, sometimes it did >1000 loops which lead to python throwing max recursion depth exceeded error.
ref #642
2 files changed with 10 insertions and 4 deletions:
0 comments (0 inline, 0 general)
rhodecode/controllers/summary.py
Show inline comments
 
@@ -47,193 +47,195 @@ from rhodecode.lib.utils2 import safe_un
 
from rhodecode.lib.auth import LoginRequired, HasRepoPermissionAnyDecorator,\
 
    NotAnonymous
 
from rhodecode.lib.base import BaseRepoController, render
 
from rhodecode.lib.vcs.backends.base import EmptyChangeset
 
from rhodecode.lib.markup_renderer import MarkupRenderer
 
from rhodecode.lib.celerylib import run_task
 
from rhodecode.lib.celerylib.tasks import get_commits_stats
 
from rhodecode.lib.helpers import RepoPage
 
from rhodecode.lib.compat import json, OrderedDict
 
from rhodecode.lib.vcs.nodes import FileNode
 

	
 
log = logging.getLogger(__name__)
 

	
 
README_FILES = [''.join([x[0][0], x[1][0]]) for x in
 
                    sorted(list(product(ALL_READMES, ALL_EXTS)),
 
                           key=lambda y:y[0][1] + y[1][1])]
 

	
 

	
 
class SummaryController(BaseRepoController):
 

	
 
    @LoginRequired()
 
    @HasRepoPermissionAnyDecorator('repository.read', 'repository.write',
 
                                   'repository.admin')
 
    def __before__(self):
 
        super(SummaryController, self).__before__()
 

	
 
    def index(self, repo_name):
 
        c.dbrepo = dbrepo = c.rhodecode_db_repo
 
        c.following = self.scm_model.is_following_repo(repo_name,
 
                                                self.rhodecode_user.user_id)
 

	
 
        def url_generator(**kw):
 
            return url('shortlog_home', repo_name=repo_name, size=10, **kw)
 

	
 
        c.repo_changesets = RepoPage(c.rhodecode_repo, page=1,
 
                                     items_per_page=10, url=url_generator)
 
        page_revisions = [x.raw_id for x in list(c.repo_changesets)]
 
        c.statuses = c.rhodecode_db_repo.statuses(page_revisions)
 

	
 
        if self.rhodecode_user.username == 'default':
 
            # for default(anonymous) user we don't need to pass credentials
 
            username = ''
 
            password = ''
 
        else:
 
            username = str(self.rhodecode_user.username)
 
            password = '@'
 

	
 
        parsed_url = urlparse(url.current(qualified=True))
 

	
 
        default_clone_uri = '{scheme}://{user}{pass}{netloc}{path}'
 

	
 
        uri_tmpl = config.get('clone_uri', default_clone_uri)
 
        uri_tmpl = uri_tmpl.replace('{', '%(').replace('}', ')s')
 
        decoded_path = safe_unicode(urllib.unquote(parsed_url.path))
 
        uri_dict = {
 
           'user': urllib.quote(username),
 
           'pass': password,
 
           'scheme': parsed_url.scheme,
 
           'netloc': parsed_url.netloc,
 
           'path': decoded_path
 
        }
 

	
 
        uri = uri_tmpl % uri_dict
 
        # generate another clone url by id
 
        uri_dict.update(
 
         {'path': decoded_path.replace(repo_name, '_%s' % c.dbrepo.repo_id)}
 
        )
 
        uri_id = uri_tmpl % uri_dict
 

	
 
        c.clone_repo_url = uri
 
        c.clone_repo_url_id = uri_id
 
        c.repo_tags = OrderedDict()
 
        for name, hash_ in c.rhodecode_repo.tags.items()[:10]:
 
            try:
 
                c.repo_tags[name] = c.rhodecode_repo.get_changeset(hash_)
 
            except ChangesetError:
 
                c.repo_tags[name] = EmptyChangeset(hash_)
 

	
 
        c.repo_branches = OrderedDict()
 
        for name, hash_ in c.rhodecode_repo.branches.items()[:10]:
 
            try:
 
                c.repo_branches[name] = c.rhodecode_repo.get_changeset(hash_)
 
            except ChangesetError:
 
                c.repo_branches[name] = EmptyChangeset(hash_)
 

	
 
        td = date.today() + timedelta(days=1)
 
        td_1m = td - timedelta(days=calendar.mdays[td.month])
 
        td_1y = td - timedelta(days=365)
 

	
 
        ts_min_m = mktime(td_1m.timetuple())
 
        ts_min_y = mktime(td_1y.timetuple())
 
        ts_max_y = mktime(td.timetuple())
 

	
 
        if dbrepo.enable_statistics:
 
            c.show_stats = True
 
            c.no_data_msg = _('No data loaded yet')
 
            run_task(get_commits_stats, c.dbrepo.repo_name, ts_min_y, ts_max_y)
 
            recurse_limit = 500  # don't recurse more than 500 times when parsing
 
            run_task(get_commits_stats, c.dbrepo.repo_name, ts_min_y,
 
                     ts_max_y, recurse_limit)
 
        else:
 
            c.show_stats = False
 
            c.no_data_msg = _('Statistics are disabled for this repository')
 
        c.ts_min = ts_min_m
 
        c.ts_max = ts_max_y
 

	
 
        stats = self.sa.query(Statistics)\
 
            .filter(Statistics.repository == dbrepo)\
 
            .scalar()
 

	
 
        c.stats_percentage = 0
 

	
 
        if stats and stats.languages:
 
            c.no_data = False is dbrepo.enable_statistics
 
            lang_stats_d = json.loads(stats.languages)
 
            c.commit_data = stats.commit_activity
 
            c.overview_data = stats.commit_activity_combined
 

	
 
            lang_stats = ((x, {"count": y,
 
                               "desc": LANGUAGES_EXTENSIONS_MAP.get(x)})
 
                          for x, y in lang_stats_d.items())
 

	
 
            c.trending_languages = json.dumps(
 
                sorted(lang_stats, reverse=True, key=lambda k: k[1])[:10]
 
            )
 
            last_rev = stats.stat_on_revision + 1
 
            c.repo_last_rev = c.rhodecode_repo.count()\
 
                if c.rhodecode_repo.revisions else 0
 
            if last_rev == 0 or c.repo_last_rev == 0:
 
                pass
 
            else:
 
                c.stats_percentage = '%.2f' % ((float((last_rev)) /
 
                                                c.repo_last_rev) * 100)
 
        else:
 
            c.commit_data = json.dumps({})
 
            c.overview_data = json.dumps([[ts_min_y, 0], [ts_max_y, 10]])
 
            c.trending_languages = json.dumps({})
 
            c.no_data = True
 

	
 
        c.enable_downloads = dbrepo.enable_downloads
 
        if c.enable_downloads:
 
            c.download_options = self._get_download_links(c.rhodecode_repo)
 

	
 
        c.readme_data, c.readme_file = \
 
            self.__get_readme_data(c.rhodecode_db_repo)
 
        return render('summary/summary.html')
 

	
 
    @NotAnonymous()
 
    @jsonify
 
    def repo_size(self, repo_name):
 
        if request.is_xhr:
 
            return _('repository size: %s') % c.rhodecode_db_repo._repo_size()
 
        else:
 
            raise HTTPBadRequest()
 

	
 
    def __get_readme_data(self, db_repo):
 
        repo_name = db_repo.repo_name
 

	
 
        @cache_region('long_term')
 
        def _get_readme_from_cache(key):
 
            readme_data = None
 
            readme_file = None
 
            log.debug('Looking for README file')
 
            try:
 
                # get's the landing revision! or tip if fails
 
                cs = db_repo.get_landing_changeset()
 
                if isinstance(cs, EmptyChangeset):
 
                    raise EmptyRepositoryError()
 
                renderer = MarkupRenderer()
 
                for f in README_FILES:
 
                    try:
 
                        readme = cs.get_node(f)
 
                        if not isinstance(readme, FileNode):
 
                            continue
 
                        readme_file = f
 
                        log.debug('Found README file `%s` rendering...' %
 
                                  readme_file)
 
                        readme_data = renderer.render(readme.content, f)
 
                        break
 
                    except NodeDoesNotExistError:
 
                        continue
 
            except ChangesetError:
 
                log.error(traceback.format_exc())
 
                pass
 
            except EmptyRepositoryError:
 
                pass
 
            except Exception:
 
                log.error(traceback.format_exc())
 

	
 
            return readme_data, readme_file
 

	
 
        key = repo_name + '_README'
 
        inv = CacheInvalidation.invalidate(key)
 
        if inv is not None:
 
            region_invalidate(_get_readme_from_cache, None, key)
 
            CacheInvalidation.set_valid(inv.cache_key)
rhodecode/lib/celerylib/tasks.py
Show inline comments
 
# -*- coding: utf-8 -*-
 
"""
 
    rhodecode.lib.celerylib.tasks
 
    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

	
 
    RhodeCode task modules, containing all task that suppose to be run
 
    by celery daemon
 

	
 
    :created_on: Oct 6, 2010
 
    :author: marcink
 
    :copyright: (C) 2010-2012 Marcin Kuzminski <marcin@python-works.com>
 
    :license: GPLv3, see COPYING for more details.
 
"""
 
# This program is free software: you can redistribute it and/or modify
 
# it under the terms of the GNU General Public License as published by
 
# the Free Software Foundation, either version 3 of the License, or
 
# (at your option) any later version.
 
#
 
# This program is distributed in the hope that it will be useful,
 
# but WITHOUT ANY WARRANTY; without even the implied warranty of
 
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 
# GNU General Public License for more details.
 
#
 
# You should have received a copy of the GNU General Public License
 
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
from celery.decorators import task
 

	
 
import os
 
import traceback
 
import logging
 
from os.path import join as jn
 

	
 
from time import mktime
 
from operator import itemgetter
 
from string import lower
 

	
 
from pylons import config, url
 
from pylons.i18n.translation import _
 

	
 
from rhodecode.lib.vcs import get_backend
 

	
 
from rhodecode import CELERY_ON, CELERY_EAGER
 
from rhodecode.lib.utils2 import safe_str
 
from rhodecode.lib.celerylib import run_task, locked_task, dbsession, \
 
    str2bool, __get_lockkey, LockHeld, DaemonLock, get_session
 
from rhodecode.lib.helpers import person
 
from rhodecode.lib.rcmail.smtp_mailer import SmtpMailer
 
from rhodecode.lib.utils import add_cache, action_logger
 
from rhodecode.lib.compat import json, OrderedDict
 
from rhodecode.lib.hooks import log_create_repository
 

	
 
from rhodecode.model.db import Statistics, Repository, User
 
from rhodecode.model.scm import ScmModel
 

	
 

	
 
add_cache(config)
 

	
 
__all__ = ['whoosh_index', 'get_commits_stats',
 
           'reset_user_password', 'send_email']
 

	
 

	
 
def get_logger(cls):
 
    if CELERY_ON:
 
        try:
 
            log = cls.get_logger()
 
        except:
 
            log = logging.getLogger(__name__)
 
    else:
 
        log = logging.getLogger(__name__)
 

	
 
    return log
 

	
 

	
 
@task(ignore_result=True)
 
@locked_task
 
@dbsession
 
def whoosh_index(repo_location, full_index):
 
    from rhodecode.lib.indexers.daemon import WhooshIndexingDaemon
 
    log = get_logger(whoosh_index)
 
    DBS = get_session()
 

	
 
    index_location = config['index_dir']
 
    WhooshIndexingDaemon(index_location=index_location,
 
                         repo_location=repo_location, sa=DBS)\
 
                         .run(full_index=full_index)
 

	
 

	
 
@task(ignore_result=True)
 
@dbsession
 
def get_commits_stats(repo_name, ts_min_y, ts_max_y):
 
def get_commits_stats(repo_name, ts_min_y, ts_max_y, recurse_limit=100):
 
    log = get_logger(get_commits_stats)
 
    DBS = get_session()
 
    lockkey = __get_lockkey('get_commits_stats', repo_name, ts_min_y,
 
                            ts_max_y)
 
    lockkey_path = config['app_conf']['cache_dir']
 

	
 
    log.info('running task with lockkey %s' % lockkey)
 

	
 
    try:
 
        lock = l = DaemonLock(file_=jn(lockkey_path, lockkey))
 

	
 
        # for js data compatibility cleans the key for person from '
 
        akc = lambda k: person(k).replace('"', "")
 

	
 
        co_day_auth_aggr = {}
 
        commits_by_day_aggregate = {}
 
        repo = Repository.get_by_repo_name(repo_name)
 
        if repo is None:
 
            return True
 

	
 
        repo = repo.scm_instance
 
        repo_size = repo.count()
 
        # return if repo have no revisions
 
        if repo_size < 1:
 
            lock.release()
 
            return True
 

	
 
        skip_date_limit = True
 
        parse_limit = int(config['app_conf'].get('commit_parse_limit'))
 
        last_rev = None
 
        last_cs = None
 
        timegetter = itemgetter('time')
 

	
 
        dbrepo = DBS.query(Repository)\
 
            .filter(Repository.repo_name == repo_name).scalar()
 
        cur_stats = DBS.query(Statistics)\
 
            .filter(Statistics.repository == dbrepo).scalar()
 

	
 
        if cur_stats is not None:
 
            last_rev = cur_stats.stat_on_revision
 

	
 
        if last_rev == repo.get_changeset().revision and repo_size > 1:
 
            # pass silently without any work if we're not on first revision or
 
            # current state of parsing revision(from db marker) is the
 
            # last revision
 
            lock.release()
 
            return True
 

	
 
        if cur_stats:
 
            commits_by_day_aggregate = OrderedDict(json.loads(
 
                                        cur_stats.commit_activity_combined))
 
            co_day_auth_aggr = json.loads(cur_stats.commit_activity)
 

	
 
        log.debug('starting parsing %s' % parse_limit)
 
        lmktime = mktime
 

	
 
        last_rev = last_rev + 1 if last_rev >= 0 else 0
 
        log.debug('Getting revisions from %s to %s' % (
 
             last_rev, last_rev + parse_limit)
 
        )
 
        for cs in repo[last_rev:last_rev + parse_limit]:
 
            log.debug('parsing %s' % cs)
 
            last_cs = cs  # remember last parsed changeset
 
            k = lmktime([cs.date.timetuple()[0], cs.date.timetuple()[1],
 
                          cs.date.timetuple()[2], 0, 0, 0, 0, 0, 0])
 

	
 
            if akc(cs.author) in co_day_auth_aggr:
 
                try:
 
                    l = [timegetter(x) for x in
 
                         co_day_auth_aggr[akc(cs.author)]['data']]
 
                    time_pos = l.index(k)
 
                except ValueError:
 
                    time_pos = False
 

	
 
                if time_pos >= 0 and time_pos is not False:
 

	
 
                    datadict = \
 
                        co_day_auth_aggr[akc(cs.author)]['data'][time_pos]
 

	
 
                    datadict["commits"] += 1
 
                    datadict["added"] += len(cs.added)
 
                    datadict["changed"] += len(cs.changed)
 
                    datadict["removed"] += len(cs.removed)
 

	
 
                else:
 
                    if k >= ts_min_y and k <= ts_max_y or skip_date_limit:
 

	
 
                        datadict = {"time": k,
 
                                    "commits": 1,
 
                                    "added": len(cs.added),
 
                                    "changed": len(cs.changed),
 
                                    "removed": len(cs.removed),
 
                                   }
 
                        co_day_auth_aggr[akc(cs.author)]['data']\
 
                            .append(datadict)
 

	
 
            else:
 
                if k >= ts_min_y and k <= ts_max_y or skip_date_limit:
 
                    co_day_auth_aggr[akc(cs.author)] = {
 
                                        "label": akc(cs.author),
 
                                        "data": [{"time":k,
 
                                                 "commits":1,
 
                                                 "added":len(cs.added),
 
                                                 "changed":len(cs.changed),
 
                                                 "removed":len(cs.removed),
 
                                                 }],
 
                                        "schema": ["commits"],
 
                                        }
 

	
 
            #gather all data by day
 
            if k in commits_by_day_aggregate:
 
                commits_by_day_aggregate[k] += 1
 
            else:
 
                commits_by_day_aggregate[k] = 1
 

	
 
        overview_data = sorted(commits_by_day_aggregate.items(),
 
                               key=itemgetter(0))
 

	
 
        if not co_day_auth_aggr:
 
            co_day_auth_aggr[akc(repo.contact)] = {
 
                "label": akc(repo.contact),
 
                "data": [0, 1],
 
                "schema": ["commits"],
 
            }
 

	
 
        stats = cur_stats if cur_stats else Statistics()
 
        stats.commit_activity = json.dumps(co_day_auth_aggr)
 
        stats.commit_activity_combined = json.dumps(overview_data)
 

	
 
        log.debug('last revison %s' % last_rev)
 
        leftovers = len(repo.revisions[last_rev:])
 
        log.debug('revisions to parse %s' % leftovers)
 

	
 
        if last_rev == 0 or leftovers < parse_limit:
 
            log.debug('getting code trending stats')
 
            stats.languages = json.dumps(__get_codes_stats(repo_name))
 

	
 
        try:
 
            stats.repository = dbrepo
 
            stats.stat_on_revision = last_cs.revision if last_cs else 0
 
            DBS.add(stats)
 
            DBS.commit()
 
        except:
 
            log.error(traceback.format_exc())
 
            DBS.rollback()
 
            lock.release()
 
            return False
 

	
 
        # final release
 
        lock.release()
 

	
 
        # execute another task if celery is enabled
 
        if len(repo.revisions) > 1 and CELERY_ON:
 
            run_task(get_commits_stats, repo_name, ts_min_y, ts_max_y)
 
        if len(repo.revisions) > 1 and CELERY_ON and recurse_limit > 0:
 
            recurse_limit -= 1
 
            run_task(get_commits_stats, repo_name, ts_min_y, ts_max_y,
 
                     recurse_limit)
 
        if recurse_limit <= 0:
 
            log.debug('Breaking recursive mode due to reach of recurse limit')
 
        return True
 
    except LockHeld:
 
        log.info('LockHeld')
 
        return 'Task with key %s already running' % lockkey
 

	
 
@task(ignore_result=True)
 
@dbsession
 
def send_password_link(user_email):
 
    from rhodecode.model.notification import EmailNotificationModel
 

	
 
    log = get_logger(send_password_link)
 
    DBS = get_session()
 

	
 
    try:
 
        user = User.get_by_email(user_email)
 
        if user:
 
            log.debug('password reset user found %s' % user)
 
            link = url('reset_password_confirmation', key=user.api_key,
 
                       qualified=True)
 
            reg_type = EmailNotificationModel.TYPE_PASSWORD_RESET
 
            body = EmailNotificationModel().get_email_tmpl(reg_type,
 
                                                **{'user':user.short_contact,
 
                                                   'reset_url':link})
 
            log.debug('sending email')
 
            run_task(send_email, user_email,
 
                     _("password reset link"), body)
 
            log.info('send new password mail to %s' % user_email)
 
        else:
 
            log.debug("password reset email %s not found" % user_email)
 
    except:
 
        log.error(traceback.format_exc())
 
        return False
 

	
 
    return True
 

	
 
@task(ignore_result=True)
 
@dbsession
 
def reset_user_password(user_email):
 
    from rhodecode.lib import auth
 

	
 
    log = get_logger(reset_user_password)
 
    DBS = get_session()
 

	
 
    try:
 
        try:
 
            user = User.get_by_email(user_email)
 
            new_passwd = auth.PasswordGenerator().gen_password(8,
 
                             auth.PasswordGenerator.ALPHABETS_BIG_SMALL)
 
            if user:
 
                user.password = auth.get_crypt_password(new_passwd)
 
                user.api_key = auth.generate_api_key(user.username)
 
                DBS.add(user)
 
                DBS.commit()
 
                log.info('change password for %s' % user_email)
 
            if new_passwd is None:
 
                raise Exception('unable to generate new password')
 
        except:
 
            log.error(traceback.format_exc())
 
            DBS.rollback()
 

	
 
        run_task(send_email, user_email,
 
                 'Your new password',
 
                 'Your new RhodeCode password:%s' % (new_passwd))
 
        log.info('send new password mail to %s' % user_email)
 

	
 
    except:
 
        log.error('Failed to update user password')
 
        log.error(traceback.format_exc())
 

	
 
    return True
 

	
 

	
 
@task(ignore_result=True)
 
@dbsession
 
def send_email(recipients, subject, body, html_body=''):
 
    """
 
    Sends an email with defined parameters from the .ini files.
 

	
 
    :param recipients: list of recipients, it this is empty the defined email
 
        address from field 'email_to' is used instead
 
    :param subject: subject of the mail
 
    :param body: body of the mail
 
    :param html_body: html version of body
 
    """
 
    log = get_logger(send_email)
 
    DBS = get_session()
 

	
 
    email_config = config
 
    subject = "%s %s" % (email_config.get('email_prefix', ''), subject)
 
    if not recipients:
 
        # if recipients are not defined we send to email_config + all admins
 
        admins = [u.email for u in User.query()
 
                  .filter(User.admin == True).all()]
 
        recipients = [email_config.get('email_to')] + admins
 

	
 
    mail_from = email_config.get('app_email_from', 'RhodeCode')
0 comments (0 inline, 0 general)