kallithea Changeset - 1f92cded1bd2

Changeset - 1f92cded1bd2

Parent rev.

Child rev.

[Not reviewed]

default

0 5 0

Mads Kiilerich (mads) - 5 years ago 2020-12-29 22:07:41
mads@kiilerich.com

celery: move Whoosh indexing task to whoosh library where it belongs

Avoid bundling everything from many different layers in one big task library.

This is more feasible now when we don't need kallithea.CELERY_APP set at import
time.

5 files changed with 18 insertions and 14 deletions:

kallithea/controllers/admin/settings.py

kallithea/lib/celery_app.py

kallithea/lib/indexers/daemon.py

kallithea/model/async_tasks.py

scripts/deps.py

0 comments (0 inline, 0 general)

kallithea/controllers/admin/settings.py

➞

Show inline comments

 # -*- coding: utf-8 -*-
 # 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/>.
 """
 kallithea.controllers.admin.settings
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 settings controller for Kallithea admin
 This file was forked by the Kallithea project in July 2014.
 Original author and date, and relevant copyright and licensing information is below:
 :created_on: Jul 14, 2010
 :author: marcink
 :copyright: (c) 2013 RhodeCode GmbH, and others.
 :license: GPLv3, see LICENSE.md for more details.
 """
 import logging
 import traceback
 import formencode
 from formencode import htmlfill
 from tg import config, request
 from tg import tmpl_context as c
 from tg.i18n import ugettext as _
 from webob.exc import HTTPFound
 import kallithea
 import kallithea.lib.indexers.daemon
 from kallithea.controllers import base
 from kallithea.lib import webutils
 from kallithea.lib.auth import HasPermissionAnyDecorator, LoginRequired
 from kallithea.lib.utils import repo2db_mapper, set_app_settings
 from kallithea.lib.utils2 import safe_str
 from kallithea.lib.vcs import VCSError
 from kallithea.lib.webutils import url
 from kallithea.model import async_tasks, db, meta
 from kallithea.model.forms import ApplicationSettingsForm, ApplicationUiSettingsForm, ApplicationVisualisationForm
 from kallithea.model.notification import EmailNotificationModel
 from kallithea.model.scm import ScmModel
 log = logging.getLogger(__name__)
 class SettingsController(base.BaseController):
     @LoginRequired(allow_default_user=True)
     def _before(self, *args, **kwargs):
         super(SettingsController, self)._before(*args, **kwargs)
     def _get_hg_ui_settings(self):
         ret = db.Ui.query().all()
         settings = {}
         for each in ret:
             k = each.ui_section + '_' + each.ui_key
             v = each.ui_value
             if k == 'paths_/':
                 k = 'paths_root_path'
             k = k.replace('.', '_')
             if each.ui_section in ['hooks', 'extensions']:
                 v = each.ui_active
             settings[k] = v
         return settings
     @HasPermissionAnyDecorator('hg.admin')
     def settings_vcs(self):
         c.active = 'vcs'
         if request.POST:
             application_form = ApplicationUiSettingsForm()()
             try:
                 form_result = application_form.to_python(dict(request.POST))
             except formencode.Invalid as errors:
                 return htmlfill.render(
                      base.render('admin/settings/settings.html'),
                      defaults=errors.value,
                      errors=errors.error_dict or {},
                      prefix_error=False,
                      encoding="UTF-8",
                      force_defaults=False)
             try:
                 if c.visual.allow_repo_location_change:
                     sett = db.Ui.get_by_key('paths', '/')
                     sett.ui_value = form_result['paths_root_path']
                 # HOOKS
                 sett = db.Ui.get_by_key('hooks', db.Ui.HOOK_UPDATE)
                 sett.ui_active = form_result['hooks_changegroup_update']
                 sett = db.Ui.get_by_key('hooks', db.Ui.HOOK_REPO_SIZE)
                 sett.ui_active = form_result['hooks_changegroup_repo_size']
                 ## EXTENSIONS
                 sett = db.Ui.get_or_create('extensions', 'largefiles')
                 sett.ui_active = form_result['extensions_largefiles']
 #                sett = db.Ui.get_or_create('extensions', 'hggit')
 #                sett.ui_active = form_result['extensions_hggit']
                 meta.Session().commit()
                 webutils.flash(_('Updated VCS settings'), category='success')
             except Exception:
                 log.error(traceback.format_exc())
                 webutils.flash(_('Error occurred while updating '
                           'application settings'), category='error')
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_mapping(self):
         c.active = 'mapping'
         if request.POST:
             rm_obsolete = request.POST.get('destroy', False)
             install_git_hooks = request.POST.get('hooks', False)
             overwrite_git_hooks = request.POST.get('hooks_overwrite', False)
             invalidate_cache = request.POST.get('invalidate', False)
             log.debug('rescanning repo location with destroy obsolete=%s, '
                       'install git hooks=%s and '
                       'overwrite git hooks=%s' % (rm_obsolete, install_git_hooks, overwrite_git_hooks))
             filesystem_repos = ScmModel().repo_scan()
             added, removed = repo2db_mapper(filesystem_repos, rm_obsolete,
                                             install_git_hooks=install_git_hooks,
                                             user=request.authuser.username,
                                             overwrite_git_hooks=overwrite_git_hooks)
             added_msg = webutils.HTML(', ').join(
                 webutils.link_to(safe_str(repo_name), webutils.url('summary_home', repo_name=repo_name)) for repo_name in added
             ) or '-'
             removed_msg = webutils.HTML(', ').join(
                 safe_str(repo_name) for repo_name in removed
             ) or '-'
             webutils.flash(webutils.HTML(_('Repositories successfully rescanned. Added: %s. Removed: %s.')) %
                     (added_msg, removed_msg), category='success')
             if invalidate_cache:
                 log.debug('invalidating all repositories cache')
                 i = 0
                 for repo in db.Repository.query():
                     try:
                         ScmModel().mark_for_invalidation(repo.repo_name)
                         i += 1
                     except VCSError as e:
                         log.warning('VCS error invalidating %s: %s', repo.repo_name, e)
                 webutils.flash(_('Invalidated %s repositories') % i, category='success')
             raise HTTPFound(location=url('admin_settings_mapping'))
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_global(self):
         c.active = 'global'
         if request.POST:
             application_form = ApplicationSettingsForm()()
             try:
                 form_result = application_form.to_python(dict(request.POST))
             except formencode.Invalid as errors:
                 return htmlfill.render(
                     base.render('admin/settings/settings.html'),
                     defaults=errors.value,
                     errors=errors.error_dict or {},
                     prefix_error=False,
                     encoding="UTF-8",
                     force_defaults=False)
             try:
                 for setting in (
                     'title',
                     'realm',
                     'ga_code',
                     'captcha_public_key',
                     'captcha_private_key',
                 ):
                     db.Setting.create_or_update(setting, form_result[setting])
                 meta.Session().commit()
                 set_app_settings(config)
                 webutils.flash(_('Updated application settings'), category='success')
             except Exception:
                 log.error(traceback.format_exc())
                 webutils.flash(_('Error occurred while updating '
                           'application settings'),
                           category='error')
             raise HTTPFound(location=url('admin_settings_global'))
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_visual(self):
         c.active = 'visual'
         if request.POST:
             application_form = ApplicationVisualisationForm()()
             try:
                 form_result = application_form.to_python(dict(request.POST))
             except formencode.Invalid as errors:
                 return htmlfill.render(
                     base.render('admin/settings/settings.html'),
                     defaults=errors.value,
                     errors=errors.error_dict or {},
                     prefix_error=False,
                     encoding="UTF-8",
                     force_defaults=False)
             try:
                 settings = [
                     ('show_public_icon', 'show_public_icon', 'bool'),
                     ('show_private_icon', 'show_private_icon', 'bool'),
                     ('stylify_metalabels', 'stylify_metalabels', 'bool'),
                     ('repository_fields', 'repository_fields', 'bool'),
                     ('dashboard_items', 'dashboard_items', 'int'),
                     ('admin_grid_items', 'admin_grid_items', 'int'),
                     ('show_version', 'show_version', 'bool'),
                     ('use_gravatar', 'use_gravatar', 'bool'),
                     ('gravatar_url', 'gravatar_url', 'unicode'),
                     ('clone_uri_tmpl', 'clone_uri_tmpl', 'unicode'),
                     ('clone_ssh_tmpl', 'clone_ssh_tmpl', 'unicode'),
+                ]
                 for setting, form_key, type_ in settings:
                     db.Setting.create_or_update(setting, form_result[form_key], type_)
                 meta.Session().commit()
                 set_app_settings(config)
                 webutils.flash(_('Updated visualisation settings'),
                         category='success')
             except Exception:
                 log.error(traceback.format_exc())
                 webutils.flash(_('Error occurred during updating '
                           'visualisation settings'),
                         category='error')
             raise HTTPFound(location=url('admin_settings_visual'))
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_email(self):
         c.active = 'email'
         if request.POST:
             test_email = request.POST.get('test_email')
             test_email_subj = 'Kallithea test email'
             test_body = ('Kallithea Email test, '
                                'Kallithea version: %s' % c.kallithea_version)
             if not test_email:
                 webutils.flash(_('Please enter email address'), category='error')
                 raise HTTPFound(location=url('admin_settings_email'))
             test_email_txt_body = EmailNotificationModel() \
                 .get_email_tmpl(EmailNotificationModel.TYPE_DEFAULT,
                                 'txt', body=test_body)
             test_email_html_body = EmailNotificationModel() \
                 .get_email_tmpl(EmailNotificationModel.TYPE_DEFAULT,
                                 'html', body=test_body)
             recipients = [test_email] if test_email else None
             async_tasks.send_email(recipients, test_email_subj,
                              test_email_txt_body, test_email_html_body)
             webutils.flash(_('Send email task created'), category='success')
             raise HTTPFound(location=url('admin_settings_email'))
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         c.ini = kallithea.CONFIG
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_hooks(self):
         c.active = 'hooks'
         if request.POST:
             if c.visual.allow_custom_hooks_settings:
                 ui_key = request.POST.get('new_hook_ui_key')
                 ui_value = request.POST.get('new_hook_ui_value')
                 hook_id = request.POST.get('hook_id')
                 try:
                     ui_key = ui_key and ui_key.strip()
                     if ui_key in (x.ui_key for x in db.Ui.get_custom_hooks()):
                         webutils.flash(_('Hook already exists'), category='error')
                     elif ui_key in (x.ui_key for x in db.Ui.get_builtin_hooks()):
                         webutils.flash(_('Builtin hooks are read-only. Please use another hook name.'), category='error')
                     elif ui_value and ui_key:
                         db.Ui.create_or_update_hook(ui_key, ui_value)
                         webutils.flash(_('Added new hook'), category='success')
                     elif hook_id:
                         db.Ui.delete(hook_id)
                         meta.Session().commit()
                     # check for edits
                     update = False
                     _d = request.POST.dict_of_lists()
                     for k, v, ov in zip(_d.get('hook_ui_key', []),
                                         _d.get('hook_ui_value_new', []),
                                         _d.get('hook_ui_value', [])):
                         if v != ov:
                             db.Ui.create_or_update_hook(k, v)
                             update = True
                     if update:
                         webutils.flash(_('Updated hooks'), category='success')
                     meta.Session().commit()
                 except Exception:
                     log.error(traceback.format_exc())
                     webutils.flash(_('Error occurred during hook creation'),
                             category='error')
                 raise HTTPFound(location=url('admin_settings_hooks'))
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         c.hooks = db.Ui.get_builtin_hooks()
         c.custom_hooks = db.Ui.get_custom_hooks()
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_search(self):
         c.active = 'search'
         if request.POST:
             repo_location = self._get_hg_ui_settings()['paths_root_path']
             full_index = request.POST.get('full_index', False)
-            async_tasks.whoosh_index(repo_location, full_index)
+            kallithea.lib.indexers.daemon.whoosh_index(repo_location, full_index)
             webutils.flash(_('Whoosh reindex task scheduled'), category='success')
             raise HTTPFound(location=url('admin_settings_search'))
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)
     @HasPermissionAnyDecorator('hg.admin')
     def settings_system(self):
         c.active = 'system'
         defaults = db.Setting.get_app_settings()
         defaults.update(self._get_hg_ui_settings())
         c.ini = kallithea.CONFIG
         server_info = db.Setting.get_server_info()
         for key, val in server_info.items():
             setattr(c, key, val)
         return htmlfill.render(
             base.render('admin/settings/settings.html'),
             defaults=defaults,
             encoding="UTF-8",
             force_defaults=False)

kallithea/lib/celery_app.py

➞

Show inline comments

 # -*- coding: utf-8 -*-
 """
 Kallithea wrapper of Celery
 The Celery configuration is in the Kallithea ini file but must be converted to an
 entirely different format before Celery can use it.
 We read the configuration from tg.config at module import time. This module can
 thus not be imported in global scope but must be imported on demand in function
 scope after tg.config has been initialized.
 To make sure that the config really has been initialized, we check one of the
 mandatory settings.
 """
 import logging
 class CeleryConfig(object):
     imports = ['kallithea.model.async_tasks']
     imports = [
         'kallithea.lib.indexers.daemon',
         'kallithea.model.async_tasks',
+    ]
     task_always_eager = False
 list_config_names = {'imports', 'accept_content'}
 desupported = set([
     'broker.url',
     'celery.accept.content',
     'celery.always.eager',
     'celery.amqp.task.result.expires',
     'celeryd.concurrency',
     'celeryd.max.tasks.per.child',
     'celery.result.backend',  # Note: the .ini template used this instead of 'celery.result_backend' in 0.6
     'celery.result.dburi',
     'celery.result.serialier',
     'celery.result.serializer',
     'celery.send.task.error.emails',
     'celery.task_always_eager',  # still a valid configuration in celery, but not supported in Kallithea
     'celery.task.serializer',
 ])
 log = logging.getLogger(__name__)
 def make_celery_config(config):
     """Return Celery config object populated from relevant settings in a config dict, such as tg.config"""
     celery_config = CeleryConfig()
     for config_key, config_value in sorted(config.items()):
         if config_key in desupported and config_value:
             log.error('Celery configuration setting %r is no longer supported', config_key)
         parts = config_key.split('.', 1)
         if parts[0] == 'celery' and len(parts) == 2:  # Celery 4 config key
             celery_key = parts[1]
         else:
             continue
         if not isinstance(config_value, str):
             continue
         if celery_key in list_config_names:
             celery_value = config_value.split()
         elif config_value.isdigit():
             celery_value = int(config_value)
         elif config_value.lower() in ['true', 'false']:
             celery_value = config_value.lower() == 'true'
         else:
             celery_value = config_value
         setattr(celery_config, celery_key, celery_value)
     return celery_config

kallithea/lib/indexers/daemon.py

➞

Show inline comments

 # -*- coding: utf-8 -*-
 # 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/>.
 """
 kallithea.lib.indexers.daemon
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 A daemon will read from task table and run tasks
 This file was forked by the Kallithea project in July 2014.
 Original author and date, and relevant copyright and licensing information is below:
 :created_on: Jan 26, 2010
 :author: marcink
 :copyright: (c) 2013 RhodeCode GmbH, and others.
 :license: GPLv3, see LICENSE.md for more details.
 """
 import logging
 import os
 import traceback
 from shutil import rmtree
 from time import mktime
 from tg import config
 from whoosh.index import create_in, exists_in, open_dir
 from whoosh.qparser import QueryParser
 from kallithea.lib import celerylib
 from kallithea.lib.conf import INDEX_EXTENSIONS, INDEX_FILENAMES
 from kallithea.lib.indexers import CHGSET_IDX_NAME, CHGSETS_SCHEMA, IDX_NAME, SCHEMA
 from kallithea.lib.utils2 import safe_str
 from kallithea.lib.vcs.exceptions import ChangesetDoesNotExistError, ChangesetError, NodeDoesNotExistError, RepositoryError
 from kallithea.model import db
 from kallithea.model.scm import ScmModel
 log = logging.getLogger('whoosh_indexer')
 class WhooshIndexingDaemon(object):
     """
     Daemon for atomic indexing jobs
     """
     def __init__(self, indexname=IDX_NAME, index_location=None,
                  repo_location=None, repo_list=None,
                  repo_update_list=None):
         self.indexname = indexname
         self.index_location = index_location
         if not index_location:
             raise Exception('You have to provide index location')
         self.repo_location = repo_location
         if not repo_location:
             raise Exception('You have to provide repositories location')
         self.repo_paths = ScmModel().repo_scan(self.repo_location)
         # filter repo list
         if repo_list:
             repo_list = set(repo_list)
             self.filtered_repo_paths = {}
             for repo_name, repo in self.repo_paths.items():
                 if repo_name in repo_list:
                     self.filtered_repo_paths[repo_name] = repo
             self.repo_paths = self.filtered_repo_paths
         # filter update repo list
         self.filtered_repo_update_paths = {}
         if repo_update_list:
             self.filtered_repo_update_paths = {}
             for repo_name, repo in self.repo_paths.items():
                 if repo_name in repo_update_list:
                     self.filtered_repo_update_paths[repo_name] = repo
             self.repo_paths = self.filtered_repo_update_paths
         self.initial = True
         if not os.path.isdir(self.index_location):
             os.makedirs(self.index_location)
             log.info('Cannot run incremental index since it does not '
                      'yet exist - running full build')
         elif not exists_in(self.index_location, IDX_NAME):
             log.info('Running full index build, as the file content '
                      'index does not exist')
         elif not exists_in(self.index_location, CHGSET_IDX_NAME):
             log.info('Running full index build, as the changeset '
                      'index does not exist')
         else:
             self.initial = False
     def _get_index_revision(self, repo):
         db_repo = db.Repository.get_by_repo_name(repo.name)
         landing_rev = 'tip'
         if db_repo:
             _rev_type, _rev = db_repo.landing_rev
             landing_rev = _rev
         return landing_rev
     def _get_index_changeset(self, repo, index_rev=None):
         if not index_rev:
             index_rev = self._get_index_revision(repo)
         cs = repo.get_changeset(index_rev)
         return cs
     def get_paths(self, repo):
         """
         recursive walk in root dir and return a set of all path in that dir
         based on repository walk function
         """
         index_paths_ = set()
         try:
             cs = self._get_index_changeset(repo)
             for _topnode, _dirs, files in cs.walk('/'):
                 for f in files:
                     index_paths_.add(os.path.join(repo.path, f.path))
         except RepositoryError:
             log.debug(traceback.format_exc())
             pass
         return index_paths_
     def get_node(self, repo, path, index_rev=None):
         """
         gets a filenode based on given full path.
         :param repo: scm repo instance
         :param path: full path including root location
         :return: FileNode
         """
         # FIXME: paths should be normalized ... or even better: don't include repo.path
         assert path.startswith(repo.path)
         assert path[len(repo.path)] in (os.path.sep, os.path.altsep)
         node_path = path[len(repo.path) + 1:]
         cs = self._get_index_changeset(repo, index_rev=index_rev)
         node = cs.get_node(node_path)
         return node
     def is_indexable_node(self, node):
         """
         Just index the content of chosen files, skipping binary files
         """
         return (node.extension in INDEX_EXTENSIONS or node.name in INDEX_FILENAMES) and \
                not node.is_binary
     def get_node_mtime(self, node):
         return mktime(node.last_changeset.date.timetuple())
     def add_doc(self, writer, path, repo, repo_name, index_rev=None):
         """
         Adding doc to writer this function itself fetches data from
         the instance of vcs backend
         """
         try:
             node = self.get_node(repo, path, index_rev)
         except (ChangesetError, NodeDoesNotExistError):
             log.debug("    >> %s - not found in %s %s", path, repo, index_rev)
             return 0, 0
         indexed = indexed_w_content = 0
         if self.is_indexable_node(node):
             bytes_content = node.content
             if b'\0' in bytes_content:
                 log.warning('    >> %s - no text content', path)
                 u_content = ''
             else:
                 log.debug('    >> %s', path)
                 u_content = safe_str(bytes_content)
                 indexed_w_content += 1
         else:
             log.debug('    >> %s - not indexable', path)
             # just index file name without it's content
             u_content = ''
             indexed += 1
         writer.add_document(
             fileid=path,
             repository_rawname=repo_name,
             repository=repo_name,
             path=path,
             content=u_content,
             modtime=self.get_node_mtime(node),
             extension=node.extension
+        )
         return indexed, indexed_w_content
     def index_changesets(self, writer, repo_name, repo, start_rev=None):
         """
         Add all changeset in the vcs repo starting at start_rev
         to the index writer
         :param writer: the whoosh index writer to add to
         :param repo_name: name of the repository from whence the
           changeset originates including the repository group
         :param repo: the vcs repository instance to index changesets for,
           the presumption is the repo has changesets to index
         :param start_rev=None: the full sha id to start indexing from
           if start_rev is None then index from the first changeset in
           the repo
         """
         if start_rev is None:
             start_rev = repo[0].raw_id
         log.debug('Indexing changesets in %s, starting at rev %s',
                   repo_name, start_rev)
         indexed = 0
         cs_iter = repo.get_changesets(start=start_rev)
         total = len(cs_iter)
         for cs in cs_iter:
             indexed += 1
             log.debug('    >> %s %s/%s', cs, indexed, total)
             writer.add_document(
                 raw_id=cs.raw_id,
                 date=cs._timestamp,
                 repository_rawname=repo_name,
                 repository=repo_name,
                 author=cs.author,
                 message=cs.message,
                 last=cs.last,
                 added=' '.join(node.path for node in cs.added).lower(),
                 removed=' '.join(node.path for node in cs.removed).lower(),
                 changed=' '.join(node.path for node in cs.changed).lower(),
                 parents=' '.join(cs.raw_id for cs in cs.parents),
+            )
         return indexed
     def index_files(self, file_idx_writer, repo_name, repo):
         """
         Index files for given repo_name
         :param file_idx_writer: the whoosh index writer to add to
         :param repo_name: name of the repository we're indexing
         :param repo: instance of vcs repo
         """
         i_cnt = iwc_cnt = 0
         log.debug('Building file index for %s @revision:%s', repo_name,
                                                 self._get_index_revision(repo))
         index_rev = self._get_index_revision(repo)
         for idx_path in self.get_paths(repo):
             i, iwc = self.add_doc(file_idx_writer, idx_path, repo, repo_name, index_rev)
             i_cnt += i
             iwc_cnt += iwc
         log.debug('added %s files %s with content for repo %s',
                   i_cnt + iwc_cnt, iwc_cnt, repo.path)
         return i_cnt, iwc_cnt
     def update_changeset_index(self):
         idx = open_dir(self.index_location, indexname=CHGSET_IDX_NAME)
         with idx.searcher() as searcher:
             writer = idx.writer()
             writer_is_dirty = False
             try:
                 indexed_total = 0
                 repo_name = None
                 for repo_name, repo in sorted(self.repo_paths.items()):
                     log.debug('Updating changeset index for repo %s', repo_name)
                     # skip indexing if there aren't any revs in the repo
                     num_of_revs = len(repo)
                     if num_of_revs < 1:
                         continue
                     qp = QueryParser('repository', schema=CHGSETS_SCHEMA)
                     q = qp.parse("last:t AND %s" % repo_name)
                     results = searcher.search(q)
                     # default to scanning the entire repo
                     last_rev = 0
                     start_id = None
                     if len(results) > 0:
                         # assuming that there is only one result, if not this
                         # may require a full re-index.
                         start_id = results[0]['raw_id']
                         try:
                             last_rev = repo.get_changeset(revision=start_id).revision
                         except ChangesetDoesNotExistError:
                             log.error('previous last revision %s not found - indexing from scratch', start_id)
                             start_id = None
                     # there are new changesets to index or a new repo to index
                     if last_rev == 0 or num_of_revs > last_rev + 1:
                         # delete the docs in the index for the previous
                         # last changeset(s)
                         for hit in results:
                             q = qp.parse("last:t AND %s AND raw_id:%s" %
                                             (repo_name, hit['raw_id']))
                             writer.delete_by_query(q)
                         # index from the previous last changeset + all new ones
                         indexed_total += self.index_changesets(writer,
                                                 repo_name, repo, start_id)
                         writer_is_dirty = True
                 log.debug('indexed %s changesets for repo %s',
                           indexed_total, repo_name
+                )
             finally:
                 if writer_is_dirty:
                     log.debug('>> COMMITING CHANGES TO CHANGESET INDEX<<')
                     writer.commit(merge=True)
                     log.debug('>>> FINISHED REBUILDING CHANGESET INDEX <<<')
                 else:
                     log.debug('>> NOTHING TO COMMIT TO CHANGESET INDEX<<')
     def update_file_index(self):
         log.debug('STARTING INCREMENTAL INDEXING UPDATE FOR EXTENSIONS %s '
                   'AND REPOS %s', INDEX_EXTENSIONS, ' and '.join(self.repo_paths))
         idx = open_dir(self.index_location, indexname=self.indexname)
         # The set of all paths in the index
         indexed_paths = set()
         # The set of all paths we need to re-index
         to_index = set()
         writer = idx.writer()
         writer_is_dirty = False
         try:
             with idx.reader() as reader:
                 # Loop over the stored fields in the index
                 for fields in reader.all_stored_fields():
                     indexed_path = fields['path']
                     indexed_repo_path = fields['repository']
                     indexed_paths.add(indexed_path)
                     if indexed_repo_path not in self.filtered_repo_update_paths:
                         continue
                     repo = self.repo_paths[indexed_repo_path]
                     try:
                         node = self.get_node(repo, indexed_path)
                         # Check if this file was changed since it was indexed
                         indexed_time = fields['modtime']
                         mtime = self.get_node_mtime(node)
                         if mtime > indexed_time:
                             # The file has changed, delete it and add it to
                             # the list of files to reindex
                             log.debug(
                                 'adding to reindex list %s mtime: %s vs %s',
                                     indexed_path, mtime, indexed_time
+                            )
                             writer.delete_by_term('fileid', indexed_path)
                             writer_is_dirty = True
                             to_index.add(indexed_path)
                     except (ChangesetError, NodeDoesNotExistError):
                         # This file was deleted since it was indexed
                         log.debug('removing from index %s', indexed_path)
                         writer.delete_by_term('path', indexed_path)
                         writer_is_dirty = True
             # Loop over the files in the filesystem
             # Assume we have a function that gathers the filenames of the
             # documents to be indexed
             ri_cnt_total = 0  # indexed
             riwc_cnt_total = 0  # indexed with content
             for repo_name, repo in sorted(self.repo_paths.items()):
                 log.debug('Updating file index for repo %s', repo_name)
                 # skip indexing if there aren't any revisions
                 if len(repo) < 1:
                     continue
                 ri_cnt = 0   # indexed
                 riwc_cnt = 0  # indexed with content
                 for path in self.get_paths(repo):
                     if path in to_index or path not in indexed_paths:
                         # This is either a file that's changed, or a new file
                         # that wasn't indexed before. So index it!
                         i, iwc = self.add_doc(writer, path, repo, repo_name)
                         writer_is_dirty = True
                         ri_cnt += i
                         ri_cnt_total += 1
                         riwc_cnt += iwc
                         riwc_cnt_total += iwc
                 log.debug('added %s files %s with content for repo %s',
                              ri_cnt + riwc_cnt, riwc_cnt, repo.path
+                )
             log.debug('indexed %s files in total and %s with content',
                         ri_cnt_total, riwc_cnt_total
+            )
         finally:
             if writer_is_dirty:
                 log.debug('>> COMMITING CHANGES TO FILE INDEX <<')
                 writer.commit(merge=True)
                 log.debug('>>> FINISHED REBUILDING FILE INDEX <<<')
             else:
                 log.debug('>> NOTHING TO COMMIT TO FILE INDEX <<')
                 writer.cancel()
     def build_indexes(self):
         if os.path.exists(self.index_location):
             log.debug('removing previous index')
             rmtree(self.index_location)
         if not os.path.exists(self.index_location):
             os.mkdir(self.index_location)
         chgset_idx = create_in(self.index_location, CHGSETS_SCHEMA,
                                indexname=CHGSET_IDX_NAME)
         chgset_idx_writer = chgset_idx.writer()
         file_idx = create_in(self.index_location, SCHEMA, indexname=IDX_NAME)
         file_idx_writer = file_idx.writer()
         log.debug('BUILDING INDEX FOR EXTENSIONS %s '
                   'AND REPOS %s', INDEX_EXTENSIONS, ' and '.join(self.repo_paths))
         for repo_name, repo in sorted(self.repo_paths.items()):
             log.debug('Updating indices for repo %s', repo_name)
             # skip indexing if there aren't any revisions
             if len(repo) < 1:
                 continue
             self.index_files(file_idx_writer, repo_name, repo)
             self.index_changesets(chgset_idx_writer, repo_name, repo)
         log.debug('>> COMMITING CHANGES <<')
         file_idx_writer.commit(merge=True)
         chgset_idx_writer.commit(merge=True)
         log.debug('>>> FINISHED BUILDING INDEX <<<')
     def update_indexes(self):
         self.update_file_index()
         self.update_changeset_index()
     def run(self, full_index=False):
         """Run daemon"""
         if full_index or self.initial:
             self.build_indexes()
         else:
             self.update_indexes()
 @celerylib.task
 @celerylib.locked_task
 def whoosh_index(repo_location, full_index):
     index_location = config['index_dir']
     WhooshIndexingDaemon(index_location=index_location,
                          repo_location=repo_location) \
                          .run(full_index=full_index)

kallithea/model/async_tasks.py

➞

Show inline comments

 # -*- coding: utf-8 -*-
 # 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/>.
 """
 kallithea.model.async_tasks
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Kallithea task modules, containing all task that suppose to be run
 by celery daemon
 This file was forked by the Kallithea project in July 2014.
 Original author and date, and relevant copyright and licensing information is below:
 :created_on: Oct 6, 2010
 :author: marcink
 :copyright: (c) 2013 RhodeCode GmbH, and others.
 :license: GPLv3, see LICENSE.md for more details.
 """
 import email.message
 import email.utils
 import os
 import smtplib
 import time
 import traceback
 from collections import OrderedDict
 from operator import itemgetter
 from time import mktime
 import celery.utils.log
 from tg import config
 import kallithea
 from kallithea.lib import celerylib, conf, ext_json, hooks
 from kallithea.lib.indexers.daemon import WhooshIndexingDaemon
 from kallithea.lib.utils2 import asbool, ascii_bytes
 from kallithea.lib.vcs.utils import author_email, author_name
 from kallithea.model import db, meta, repo, userlog
-__all__ = ['whoosh_index', 'get_commits_stats', 'send_email']
 __all__ = ['get_commits_stats', 'send_email']
 log = celery.utils.log.get_task_logger(__name__)
 @celerylib.task
 @celerylib.locked_task
 def whoosh_index(repo_location, full_index):
     index_location = config['index_dir']
     WhooshIndexingDaemon(index_location=index_location,
                          repo_location=repo_location) \
                          .run(full_index=full_index)
 def _author_username(author):
     """Return the username of the user identified by the email part of the 'author' string,
     default to the name or email.
     Kind of similar to h.person() ."""
     email = author_email(author)
     if email:
         user = db.User.get_by_email(email)
         if user is not None:
             return user.username
     # Still nothing?  Just pass back the author name if any, else the email
     return author_name(author) or email
 @celerylib.task
 def get_commits_stats(repo_name, ts_min_y, ts_max_y, recurse_limit=100):
     lockkey = celerylib.__get_lockkey('get_commits_stats', repo_name, ts_min_y,
                             ts_max_y)
     log.info('running task with lockkey %s', lockkey)
     try:
         lock = celerylib.DaemonLock(os.path.join(config['cache_dir'], lockkey))
         co_day_auth_aggr = {}
         commits_by_day_aggregate = {}
         db_repo = db.Repository.get_by_repo_name(repo_name)
         if db_repo is None:
             return
         scm_repo = db_repo.scm_instance
         repo_size = scm_repo.count()
         # return if repo have no revisions
         if repo_size < 1:
             lock.release()
             return
         skip_date_limit = True
         parse_limit = int(config.get('commit_parse_limit'))
         last_rev = None
         last_cs = None
         timegetter = itemgetter('time')
         dbrepo = db.Repository.query() \
             .filter(db.Repository.repo_name == repo_name).scalar()
         cur_stats = db.Statistics.query() \
             .filter(db.Statistics.repository == dbrepo).scalar()
         if cur_stats is not None:
             last_rev = cur_stats.stat_on_revision
         if last_rev == scm_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
         if cur_stats:
             commits_by_day_aggregate = OrderedDict(ext_json.loads(
                                         cur_stats.commit_activity_combined))
             co_day_auth_aggr = ext_json.loads(cur_stats.commit_activity)
         log.debug('starting parsing %s', parse_limit)
         last_rev = last_rev + 1 if last_rev and last_rev >= 0 else 0
         log.debug('Getting revisions from %s to %s',
              last_rev, last_rev + parse_limit
+        )
         usernames_cache = {}
         for cs in scm_repo[last_rev:last_rev + parse_limit]:
             log.debug('parsing %s', cs)
             last_cs = cs  # remember last parsed changeset
             tt = cs.date.timetuple()
             k = mktime(tt[:3] + (0, 0, 0, 0, 0, 0))
             # get username from author - similar to what h.person does
             username = usernames_cache.get(cs.author)
             if username is None:
                 username = _author_username(cs.author)
                 usernames_cache[cs.author] = username
             if username in co_day_auth_aggr:
                 try:
                     l = [timegetter(x) for x in
                          co_day_auth_aggr[username]['data']]
                     time_pos = l.index(k)
                 except ValueError:
                     time_pos = None
                 if time_pos is not None and time_pos >= 0:
                     datadict = \
                         co_day_auth_aggr[username]['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[username]['data'] \
                             .append(datadict)
             else:
                 if k >= ts_min_y and k <= ts_max_y or skip_date_limit:
                     co_day_auth_aggr[username] = {
                                         "label": username,
                                         "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))
         stats = cur_stats if cur_stats else db.Statistics()
         stats.commit_activity = ascii_bytes(ext_json.dumps(co_day_auth_aggr))
         stats.commit_activity_combined = ascii_bytes(ext_json.dumps(overview_data))
         log.debug('last revision %s', last_rev)
         leftovers = len(scm_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 = ascii_bytes(ext_json.dumps(__get_codes_stats(repo_name)))
         try:
             stats.repository = dbrepo
             stats.stat_on_revision = last_cs.revision if last_cs else 0
             meta.Session().add(stats)
             meta.Session().commit()
         except:
             log.error(traceback.format_exc())
             meta.Session().rollback()
             lock.release()
             return
         # final release
         lock.release()
         # execute another task if celery is enabled
         if len(scm_repo.revisions) > 1 and asbool(kallithea.CONFIG.get('use_celery')) and recurse_limit > 0:
             get_commits_stats(repo_name, ts_min_y, ts_max_y, recurse_limit - 1)
         elif recurse_limit <= 0:
             log.debug('Not recursing - limit has been reached')
         else:
             log.debug('Not recursing')
     except celerylib.LockHeld:
         log.info('Task with key %s already running', lockkey)
 @celerylib.task
 def send_email(recipients, subject, body='', html_body='', headers=None, from_name=None):
     """
     Sends an email with defined parameters from the .ini files.
     :param recipients: list of recipients, if this is None, the defined email
         address from field 'email_to' and all admins is used instead
     :param subject: subject of the mail
     :param body: plain text body of the mail
     :param html_body: html version of body
     :param headers: dictionary of prepopulated e-mail headers
     :param from_name: full name to be used as sender of this mail - often a
     .full_name_or_username value
     """
     assert isinstance(recipients, list), recipients
     if headers is None:
         headers = {}
     else:
         # do not modify the original headers object passed by the caller
         headers = headers.copy()
     email_config = config
     email_prefix = email_config.get('email_prefix', '')
     if email_prefix:
         subject = "%s %s" % (email_prefix, subject)
     if not recipients:
         # if recipients are not defined we send to email_config + all admins
         recipients = [u.email for u in db.User.query()
                       .filter(db.User.admin == True).all()]
         if email_config.get('email_to') is not None:
             recipients += email_config.get('email_to').split(',')
         # If there are still no recipients, there are no admins and no address
         # configured in email_to, so return.
         if not recipients:
             log.error("No recipients specified and no fallback available.")
             return
         log.warning("No recipients specified for '%s' - sending to admins %s", subject, ' '.join(recipients))
     # SMTP sender
     app_email_from = email_config.get('app_email_from', 'Kallithea')
     # 'From' header
     if from_name is not None:
         # set From header based on from_name but with a generic e-mail address
         # In case app_email_from is in "Some Name <e-mail>" format, we first
         # extract the e-mail address.
         envelope_addr = author_email(app_email_from)
         headers['From'] = '"%s" <%s>' % (
             email.utils.quote('%s (no-reply)' % from_name),
             envelope_addr)
     smtp_server = email_config.get('smtp_server')
     smtp_port = email_config.get('smtp_port')
     smtp_use_tls = asbool(email_config.get('smtp_use_tls'))
     smtp_use_ssl = asbool(email_config.get('smtp_use_ssl'))
     smtp_auth = email_config.get('smtp_auth')  # undocumented - overrule automatic choice of auth mechanism
     smtp_username = email_config.get('smtp_username')
     smtp_password = email_config.get('smtp_password')
     logmsg = ("Mail details:\n"
               "recipients: %s\n"
               "headers: %s\n"
               "subject: %s\n"
               "body:\n%s\n"
               "html:\n%s\n"
               % (' '.join(recipients), headers, subject, body, html_body))
     if smtp_server:
         log.debug("Sending e-mail. " + logmsg)
     else:
         log.error("SMTP mail server not configured - cannot send e-mail.")
         log.warning(logmsg)
         return
     msg = email.message.EmailMessage()
     msg['Subject'] = subject
     msg['From'] = app_email_from  # fallback - might be overridden by a header
     msg['To'] = ', '.join(recipients)
     msg['Date'] = email.utils.formatdate(time.time())
     for key, value in headers.items():
         del msg[key]  # Delete key first to make sure add_header will replace header (if any), no matter the casing
         msg.add_header(key, value)
     msg.set_content(body)
     msg.add_alternative(html_body, subtype='html')
     try:
         if smtp_use_ssl:
             smtp_serv = smtplib.SMTP_SSL(smtp_server, smtp_port)
         else:
             smtp_serv = smtplib.SMTP(smtp_server, smtp_port)
         if smtp_use_tls:
             smtp_serv.starttls()
         if smtp_auth:
             smtp_serv.ehlo()  # populate esmtp_features
             smtp_serv.esmtp_features["auth"] = smtp_auth
         if smtp_username and smtp_password is not None:
             smtp_serv.login(smtp_username, smtp_password)
         smtp_serv.sendmail(app_email_from, recipients, msg.as_string())
         smtp_serv.quit()
         log.info('Mail was sent to: %s' % recipients)
     except:
         log.error('Mail sending failed')
         log.error(traceback.format_exc())
 @celerylib.task
 def create_repo(form_data, cur_user):
     cur_user = db.User.guess_instance(cur_user)
     owner = cur_user
     repo_name = form_data['repo_name']
     repo_name_full = form_data['repo_name_full']
     repo_type = form_data['repo_type']
     description = form_data['repo_description']
     private = form_data['repo_private']
     clone_uri = form_data.get('clone_uri')
     repo_group = form_data['repo_group']
     landing_rev = form_data['repo_landing_rev']
     copy_fork_permissions = form_data.get('copy_permissions')
     copy_group_permissions = form_data.get('repo_copy_permissions')
     fork_of = form_data.get('fork_parent_id')
     state = form_data.get('repo_state', db.Repository.STATE_PENDING)
     # repo creation defaults, private and repo_type are filled in form
     defs = db.Setting.get_default_repo_settings(strip_prefix=True)
     enable_statistics = defs.get('repo_enable_statistics')
     enable_downloads = defs.get('repo_enable_downloads')
     try:
         db_repo = repo.RepoModel()._create_repo(
             repo_name=repo_name_full,
             repo_type=repo_type,
             description=description,
             owner=owner,
             private=private,
             clone_uri=clone_uri,
             repo_group=repo_group,
             landing_rev=landing_rev,
             fork_of=fork_of,
             copy_fork_permissions=copy_fork_permissions,
             copy_group_permissions=copy_group_permissions,
             enable_statistics=enable_statistics,
             enable_downloads=enable_downloads,
             state=state
+        )
         userlog.action_logger(cur_user, 'user_created_repo',
                       form_data['repo_name_full'], '')
         meta.Session().commit()
         # now create this repo on Filesystem
         repo.RepoModel()._create_filesystem_repo(
             repo_name=repo_name,
             repo_type=repo_type,
             repo_group=db.RepoGroup.guess_instance(repo_group),
             clone_uri=clone_uri,
+        )
         db_repo = db.Repository.get_by_repo_name(repo_name_full)
         hooks.log_create_repository(db_repo.get_dict(), created_by=owner.username)
         # update repo changeset caches initially
         db_repo.update_changeset_cache()
         # set new created state
         db_repo.set_state(db.Repository.STATE_CREATED)
         meta.Session().commit()
     except Exception as e:
         log.warning('Exception %s occurred when forking repository, '
                     'doing cleanup...' % e)
         # rollback things manually !
         db_repo = db.Repository.get_by_repo_name(repo_name_full)
         if db_repo:
             db.Repository.delete(db_repo.repo_id)
             meta.Session().commit()
             repo.RepoModel()._delete_filesystem_repo(db_repo)
         raise
 @celerylib.task
 def create_repo_fork(form_data, cur_user):
     """
     Creates a fork of repository using interval VCS methods
     :param form_data:
     :param cur_user:
     """
     base_path = kallithea.CONFIG['base_path']
     cur_user = db.User.guess_instance(cur_user)
     repo_name = form_data['repo_name']  # fork in this case
     repo_name_full = form_data['repo_name_full']
     repo_type = form_data['repo_type']
     owner = cur_user
     private = form_data['private']
     clone_uri = form_data.get('clone_uri')
     repo_group = form_data['repo_group']
     landing_rev = form_data['landing_rev']
     copy_fork_permissions = form_data.get('copy_permissions')
     try:
         fork_of = db.Repository.guess_instance(form_data.get('fork_parent_id'))
         repo.RepoModel()._create_repo(
             repo_name=repo_name_full,
             repo_type=repo_type,
             description=form_data['description'],
             owner=owner,

scripts/deps.py

➞

Show inline comments

 #!/usr/bin/env python3
 import re
 import sys
 ignored_modules = set('''
 argparse
 base64
 bcrypt
 binascii
 bleach
 calendar
 celery
 celery
 chardet
 click
 collections
 configparser
 copy
 csv
 ctypes
 datetime
 dateutil
 decimal
 decorator
 difflib
 distutils
 docutils
 email
 errno
 fileinput
 functools
 getpass
 grp
 hashlib
 hmac
 html
 http
 imp
 importlib
 inspect
 io
 ipaddr
 IPython
 isapi_wsgi
 itertools
 json
 kajiki
 ldap
 logging
 mako
 markdown
 mimetypes
 mock
 msvcrt
 multiprocessing
 operator
 os
 paginate
 paginate_sqlalchemy
 pam
 paste
 pkg_resources
 platform
 posixpath
 pprint
 pwd
 pyflakes
 pytest
 pytest_localserver
 random
 re
 routes
 setuptools
 shlex
 shutil
 smtplib
 socket
 ssl
 stat
 string
 struct
 subprocess
 sys
 tarfile
 tempfile
 textwrap
 tgext
 threading
 time
 traceback
 traitlets
 types
 urllib
 urlobject
 uuid
 warnings
 webhelpers2
 webob
 webtest
 whoosh
 win32traceutil
 zipfile
 '''.split())
 top_modules = set('''
 kallithea.alembic
 kallithea.bin
 kallithea.config
 kallithea.controllers
 kallithea.templates.py
 scripts
 '''.split())
 bottom_external_modules = set('''
 tg
 mercurial
 sqlalchemy
 alembic
 formencode
 pygments
 dulwich
 beaker
 psycopg2
 docs
 setup
 conftest
 '''.split())
 normal_modules = set('''
 kallithea
 kallithea.controllers.base
 kallithea.lib
 kallithea.lib.auth
 kallithea.lib.auth_modules
 kallithea.lib.celerylib
 kallithea.lib.db_manage
 kallithea.lib.helpers
 kallithea.lib.hooks
 kallithea.lib.indexers
 kallithea.lib.utils
 kallithea.lib.utils2
 kallithea.lib.vcs
 kallithea.lib.webutils
 kallithea.model
 kallithea.model.async_tasks
 kallithea.model.scm
 kallithea.templates.py
 '''.split())
 shown_modules = normal_modules | top_modules
 # break the chains somehow - this is a cleanup TODO list
 known_violations = set([
 ('kallithea.lib.auth_modules', 'kallithea.lib.auth'),  # needs base&facade
 ('kallithea.lib.utils', 'kallithea.model'),  # clean up utils
 ('kallithea.lib.utils', 'kallithea.model.db'),
 ('kallithea.lib.utils', 'kallithea.model.scm'),
 ('kallithea.model.async_tasks', 'kallithea.lib.hooks'),
 ('kallithea.model.async_tasks', 'kallithea.lib.indexers'),
 ('kallithea.model.async_tasks', 'kallithea.model'),
 ('kallithea.model', 'kallithea.lib.auth'),  # auth.HasXXX
 ('kallithea.model', 'kallithea.lib.auth_modules'),  # validators
 ('kallithea.model', 'kallithea.lib.hooks'),  # clean up hooks
 ('kallithea.model', 'kallithea.model.scm'),
 ('kallithea.model.scm', 'kallithea.lib.hooks'),
 ])
 extra_edges = [
 ('kallithea.config', 'kallithea.controllers'),  # through TG
 ('kallithea.lib.auth', 'kallithea.lib.auth_modules'),  # custom loader
+]
 def normalize(s):
     """Given a string with dot path, return the string it should be shown as."""
     parts = s.replace('.__init__', '').split('.')
     short_2 = '.'.join(parts[:2])
     short_3 = '.'.join(parts[:3])
     short_4 = '.'.join(parts[:4])
     if parts[0] in ['scripts', 'contributor_data', 'i18n_utils']:
         return 'scripts'
     if short_3 == 'kallithea.model.meta':
         return 'kallithea.model.db'
     if parts[:4] == ['kallithea', 'lib', 'vcs', 'ssh']:
         return 'kallithea.lib.vcs.ssh'
     if short_4 in shown_modules:
         return short_4
     if short_3 in shown_modules:
         return short_3
     if short_2 in shown_modules:
         return short_2
     if short_2 == 'kallithea.tests':
         return None
     if parts[0] in ignored_modules:
         return None
     assert parts[0] in bottom_external_modules, parts
     return parts[0]
 def main(filenames):
     if not filenames or filenames[0].startswith('-'):
         print('''\
 Usage:
     hg files 'set:!binary()&grep("^#!.*python")' 'set:**.py' | xargs scripts/deps.py
     dot -Tsvg deps.dot > deps.svg
         ''')
         raise SystemExit(1)
     files_imports = dict()  # map filenames to its imports
     import_deps = set()  # set of tuples with module name and its imports
     for fn in filenames:
         with open(fn) as f:
             s = f.read()
         dot_name = (fn[:-3] if fn.endswith('.py') else fn).replace('/', '.')
         file_imports = set()
         for m in re.finditer(r'^ *(?:from ([^ ]*) import (?:([a-zA-Z].*)|\(([^)]*)\))|import (.*))$', s, re.MULTILINE):
             m_from, m_from_import, m_from_import2, m_import = m.groups()
             if m_from:
                 pre = m_from + '.'
                 if pre.startswith('.'):
                     pre = dot_name.rsplit('.', 1)[0] + pre
                 importlist = m_from_import or m_from_import2
             else:
                 pre = ''
                 importlist = m_import
             for imp in importlist.split('#', 1)[0].split(','):
                 full_imp = pre + imp.strip().split(' as ', 1)[0]
                 file_imports.add(full_imp)
                 import_deps.add((dot_name, full_imp))
         files_imports[fn] = file_imports
     # dump out all deps for debugging and analysis
     with open('deps.txt', 'w') as f:
         for fn, file_imports in sorted(files_imports.items()):
             for file_import in sorted(file_imports):
                 if file_import.split('.', 1)[0] in ignored_modules:
                     continue
                 f.write('%s: %s\n' % (fn, file_import))
     # find leafs that haven't been ignored - they are the important external dependencies and shown in the bottom row
     only_imported = set(
         set(normalize(b) for a, b in import_deps) -
         set(normalize(a) for a, b in import_deps) -
         set([None, 'kallithea'])
+    )
     normalized_dep_edges = set()
     for dot_name, full_imp in import_deps:
         a = normalize(dot_name)
         b = normalize(full_imp)
         if a is None or b is None or a == b:
             continue
         normalized_dep_edges.add((a, b))
         #print((dot_name, full_imp, a, b))
     normalized_dep_edges.update(extra_edges)
     unseen_shown_modules = shown_modules.difference(a for a, b in normalized_dep_edges).difference(b for a, b in normalized_dep_edges)
     assert not unseen_shown_modules, unseen_shown_modules
     with open('deps.dot', 'w') as f:
         f.write('digraph {\n')
         f.write('subgraph { rank = same; %s}\n' % ''.join('"%s"; ' % s for s in sorted(top_modules)))
         f.write('subgraph { rank = same; %s}\n' % ''.join('"%s"; ' % s for s in sorted(only_imported)))
         for a, b in sorted(normalized_dep_edges):
             f.write('  "%s" -> "%s"%s\n' % (a, b, ' [color=red]' if (a, b) in known_violations else ' [color=green]' if (a, b) in extra_edges else ''))
         f.write('}\n')
     # verify dependencies by untangling dependency chain bottom-up:
     todo = set(normalized_dep_edges)
     unseen_violations = known_violations.difference(todo)
     assert not unseen_violations, unseen_violations
     for x in known_violations:
         todo.remove(x)
     while todo:
         depending = set(a for a, b in todo)
         depended = set(b for a, b in todo)
         drop = depended - depending
         if not drop:
             print('ERROR: cycles:', len(todo))
             for x in sorted(todo):
                 print('%s,' % (x,))
             raise SystemExit(1)
         #for do_b in sorted(drop):
         #    print('Picking', do_b, '- unblocks:', ' '.join(a for a, b in sorted((todo)) if b == do_b))
         todo = set((a, b) for a, b in todo if b in depending)
         #print()
 if __name__ == '__main__':
     main(sys.argv[1:])

0 comments (0 inline, 0 general)