##// END OF EJS Templates
goto-switcher: optimized performance and query capabilities....
goto-switcher: optimized performance and query capabilities. - Previous implementation had on significant bug. The use of LIMIT 20 was limiting results BEFORE auth checks. In case of large ammount of similarly named repositories user didn't had access too, the result goto search was empty. This was becuase first 20 items fetched didn't pass permission checks and final auth list was empty. To fix this we now use proper filtering for auth using SQL. It means we first check user allowed repositories, and add this as a filter so end result from database is already to only the accessible repositories.

File last commit:

r1684:21d2623d default
r2038:2bdf9d4d default
Show More
whoosh.py
281 lines | 9.7 KiB | text/x-python | PythonLexer
# -*- coding: utf-8 -*-
# Copyright (C) 2012-2017 RhodeCode GmbH
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License, version 3
# (only), as published by the Free Software Foundation.
#
# 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 Affero General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
#
# This program is dual-licensed. If you wish to learn more about the
# RhodeCode Enterprise Edition, including its added features, Support services,
# and proprietary license terms, please see https://rhodecode.com/licenses/
"""
Index schema for RhodeCode
"""
from __future__ import absolute_import
import logging
import os
import re
from pylons.i18n.translation import _
from whoosh import query as query_lib, sorting
from whoosh.highlight import HtmlFormatter, ContextFragmenter
from whoosh.index import create_in, open_dir, exists_in, EmptyIndexError
from whoosh.qparser import QueryParser, QueryParserError
import rhodecode.lib.helpers as h
from rhodecode.lib.index import BaseSearch
log = logging.getLogger(__name__)
try:
# we first try to import from rhodecode tools, fallback to copies if
# we're unable to
from rhodecode_tools.lib.fts_index.whoosh_schema import (
ANALYZER, FILE_INDEX_NAME, FILE_SCHEMA, COMMIT_INDEX_NAME,
COMMIT_SCHEMA)
except ImportError:
log.warning('rhodecode_tools schema not available, doing a fallback '
'import from `rhodecode.lib.index.whoosh_fallback_schema`')
from rhodecode.lib.index.whoosh_fallback_schema import (
ANALYZER, FILE_INDEX_NAME, FILE_SCHEMA, COMMIT_INDEX_NAME,
COMMIT_SCHEMA)
FORMATTER = HtmlFormatter('span', between='\n<span class="break">...</span>\n')
FRAGMENTER = ContextFragmenter(200)
log = logging.getLogger(__name__)
class Search(BaseSearch):
# this also shows in UI
query_lang_doc = 'http://whoosh.readthedocs.io/en/latest/querylang.html'
name = 'whoosh'
def __init__(self, config):
super(Search, self).__init__()
self.config = config
if not os.path.isdir(self.config['location']):
os.makedirs(self.config['location'])
opener = create_in
if exists_in(self.config['location'], indexname=FILE_INDEX_NAME):
opener = open_dir
file_index = opener(self.config['location'], schema=FILE_SCHEMA,
indexname=FILE_INDEX_NAME)
opener = create_in
if exists_in(self.config['location'], indexname=COMMIT_INDEX_NAME):
opener = open_dir
changeset_index = opener(self.config['location'], schema=COMMIT_SCHEMA,
indexname=COMMIT_INDEX_NAME)
self.commit_schema = COMMIT_SCHEMA
self.commit_index = changeset_index
self.file_schema = FILE_SCHEMA
self.file_index = file_index
self.searcher = None
def cleanup(self):
if self.searcher:
self.searcher.close()
def _extend_query(self, query):
hashes = re.compile('([0-9a-f]{5,40})').findall(query)
if hashes:
hashes_or_query = ' OR '.join('commit_id:%s*' % h for h in hashes)
query = u'(%s) OR %s' % (query, hashes_or_query)
return query
def search(self, query, document_type, search_user,
repo_name=None, requested_page=1, page_limit=10, sort=None,
raise_on_exc=True):
original_query = query
query = self._extend_query(query)
log.debug(u'QUERY: %s on %s', query, document_type)
result = {
'results': [],
'count': 0,
'error': None,
'runtime': 0
}
search_type, index_name, schema_defn = self._prepare_for_search(
document_type)
self._init_searcher(index_name)
try:
qp = QueryParser(search_type, schema=schema_defn)
allowed_repos_filter = self._get_repo_filter(
search_user, repo_name)
try:
query = qp.parse(unicode(query))
log.debug('query: %s (%s)' % (query, repr(query)))
reverse, sortedby = False, None
if search_type == 'message':
if sort == 'oldfirst':
sortedby = 'date'
reverse = False
elif sort == 'newfirst':
sortedby = 'date'
reverse = True
whoosh_results = self.searcher.search(
query, filter=allowed_repos_filter, limit=None,
sortedby=sortedby, reverse=reverse)
# fixes for 32k limit that whoosh uses for highlight
whoosh_results.fragmenter.charlimit = None
res_ln = whoosh_results.scored_length()
result['runtime'] = whoosh_results.runtime
result['count'] = res_ln
result['results'] = WhooshResultWrapper(
search_type, res_ln, whoosh_results)
except QueryParserError:
result['error'] = _('Invalid search query. Try quoting it.')
except (EmptyIndexError, IOError, OSError):
msg = _('There is no index to search in. '
'Please run whoosh indexer')
log.exception(msg)
result['error'] = msg
except Exception:
msg = _('An error occurred during this search operation')
log.exception(msg)
result['error'] = msg
return result
def statistics(self):
stats = [
{'key': _('Index Type'), 'value': 'Whoosh'},
{'key': _('File Index'), 'value': str(self.file_index)},
{'key': _('Indexed documents'),
'value': self.file_index.doc_count()},
{'key': _('Last update'),
'value': h.time_to_datetime(self.file_index.last_modified())},
{'key': _('Commit index'), 'value': str(self.commit_index)},
{'key': _('Indexed documents'),
'value': str(self.commit_index.doc_count())},
{'key': _('Last update'),
'value': h.time_to_datetime(self.commit_index.last_modified())}
]
return stats
def _get_repo_filter(self, auth_user, repo_name):
allowed_to_search = [
repo for repo, perm in
auth_user.permissions['repositories'].items()
if perm != 'repository.none']
if repo_name:
repo_filter = [query_lib.Term('repository', repo_name)]
elif 'hg.admin' in auth_user.permissions.get('global', []):
return None
else:
repo_filter = [query_lib.Term('repository', _rn)
for _rn in allowed_to_search]
# in case we're not allowed to search anywhere, it's a trick
# to tell whoosh we're filtering, on ALL results
repo_filter = repo_filter or [query_lib.Term('repository', '')]
return query_lib.Or(repo_filter)
def _prepare_for_search(self, cur_type):
search_type = {
'content': 'content',
'commit': 'message',
'path': 'path',
'repository': 'repository'
}.get(cur_type, 'content')
index_name = {
'content': FILE_INDEX_NAME,
'commit': COMMIT_INDEX_NAME,
'path': FILE_INDEX_NAME
}.get(cur_type, FILE_INDEX_NAME)
schema_defn = {
'content': self.file_schema,
'commit': self.commit_schema,
'path': self.file_schema
}.get(cur_type, self.file_schema)
log.debug('IDX: %s' % index_name)
log.debug('SCHEMA: %s' % schema_defn)
return search_type, index_name, schema_defn
def _init_searcher(self, index_name):
idx = open_dir(self.config['location'], indexname=index_name)
self.searcher = idx.searcher()
return self.searcher
class WhooshResultWrapper(object):
def __init__(self, search_type, total_hits, results):
self.search_type = search_type
self.results = results
self.total_hits = total_hits
def __str__(self):
return '<%s at %s>' % (self.__class__.__name__, len(self))
def __repr__(self):
return self.__str__()
def __len__(self):
return self.total_hits
def __iter__(self):
"""
Allows Iteration over results,and lazy generate content
*Requires* implementation of ``__getitem__`` method.
"""
for hit in self.results:
yield self.get_full_content(hit)
def __getitem__(self, key):
"""
Slicing of resultWrapper
"""
i, j = key.start, key.stop
for hit in self.results[i:j]:
yield self.get_full_content(hit)
def get_full_content(self, hit):
# TODO: marcink: this feels like an overkill, there's a lot of data
# inside hit object, and we don't need all
res = dict(hit)
f_path = '' # noqa
if self.search_type in ['content', 'path']:
f_path = res['path'][len(res['repository']):]
f_path = f_path.lstrip(os.sep)
if self.search_type == 'content':
res.update({'content_short_hl': hit.highlights('content'),
'f_path': f_path})
elif self.search_type == 'path':
res.update({'f_path': f_path})
elif self.search_type == 'message':
res.update({'message_hl': hit.highlights('message')})
return res