Mercurial > kallithea

annotate rhodecode/lib/indexers/daemon.py @ 763:0dad296d2a57 beta

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
extended trending languages to more entries, implemented new faster and "fancy" method of extensions recognition. Fixed small bug in indexer when data dir was empty
author Marcin Kuzminski <marcin@python-works.com>
date Fri, 26 Nov 2010 02:25:39 +0100
parents 7486da5f0628
children aac24db58ce8
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
1 #!/usr/bin/env python
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
2 # encoding: utf-8
549
f99075170eb4 more renames for rhode code !!
Marcin Kuzminski <marcin@python-works.com>
parents: 547
diff changeset
3 # whoosh indexer daemon for rhodecode
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
4 # Copyright (C) 2009-2010 Marcin Kuzminski <marcin@python-works.com>
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
5 #
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
6 # This program is free software; you can redistribute it and/or
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
7 # modify it under the terms of the GNU General Public License
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
8 # as published by the Free Software Foundation; version 2
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
9 # of the License or (at your opinion) any later version of the license.
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
10 #
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
11 # This program is distributed in the hope that it will be useful,
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
14 # GNU General Public License for more details.
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
15 #
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
16 # You should have received a copy of the GNU General Public License
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
17 # along with this program; if not, write to the Free Software
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
18 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
19 # MA 02110-1301, USA.
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
20 """
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
21 Created on Jan 26, 2010
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
22
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
23 @author: marcink
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
24 A deamon will read from task table and run tasks
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
25 """
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
26 import sys
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
27 import os
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
28 from os.path import dirname as dn
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
29 from os.path import join as jn
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
30
547
1e757ac98988 renamed project to rhodecode
Marcin Kuzminski <marcin@python-works.com>
parents: 497
diff changeset
31 #to get the rhodecode import
411
9b67cebe6609 some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents: 407
diff changeset
32 project_path = dn(dn(dn(dn(os.path.realpath(__file__)))))
9b67cebe6609 some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents: 407
diff changeset
33 sys.path.append(project_path)
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
34
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
35
691
7486da5f0628 Refactor codes for scm model
Marcin Kuzminski <marcin@python-works.com>
parents: 683
diff changeset
36 from rhodecode.model.scm import ScmModel
547
1e757ac98988 renamed project to rhodecode
Marcin Kuzminski <marcin@python-works.com>
parents: 497
diff changeset
37 from rhodecode.lib.helpers import safe_unicode
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
38 from whoosh.index import create_in, open_dir
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
39 from shutil import rmtree
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
40 from rhodecode.lib.indexers import INDEX_EXTENSIONS, SCHEMA, IDX_NAME
411
9b67cebe6609 some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents: 407
diff changeset
41
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
42 from time import mktime
567
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
43 from vcs.exceptions import ChangesetError, RepositoryError
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
44
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
45 import logging
483
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
46
411
9b67cebe6609 some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents: 407
diff changeset
47 log = logging.getLogger('whooshIndexer')
483
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
48 # create logger
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
49 log.setLevel(logging.DEBUG)
491
fefffd6fd5f4 Added some more tests, rewrite testing schema, to autogenerate fresh db, new index.
Marcin Kuzminski <marcin@python-works.com>
parents: 483
diff changeset
50 log.propagate = False
483
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
51 # create console handler and set level to debug
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
52 ch = logging.StreamHandler()
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
53 ch.setLevel(logging.DEBUG)
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
54
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
55 # create formatter
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
56 formatter = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
57
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
58 # add formatter to ch
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
59 ch.setFormatter(formatter)
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
60
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
61 # add ch to logger
a9e50dce3081 Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents: 465
diff changeset
62 log.addHandler(ch)
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
63
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
64 class WhooshIndexingDaemon(object):
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
65 """
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
66 Deamon for atomic jobs
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
67 """
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
68
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
69 def __init__(self, indexname='HG_INDEX', index_location=None,
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
70 repo_location=None):
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
71 self.indexname = indexname
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
72
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
73 self.index_location = index_location
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
74 if not index_location:
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
75 raise Exception('You have to provide index location')
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
76
411
9b67cebe6609 some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents: 407
diff changeset
77 self.repo_location = repo_location
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
78 if not repo_location:
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
79 raise Exception('You have to provide repositories location')
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
80
763
0dad296d2a57 extended trending languages to more entries, implemented new faster and "fancy"
Marcin Kuzminski <marcin@python-works.com>
parents: 691
diff changeset
81 self.repo_paths = ScmModel().repo_scan(self.repo_location, None)
465
e01a85f9fc90 fixed initial whoosh indexer. Build full index on first run even with incremental flag
Marcin Kuzminski <marcin@python-works.com>
parents: 452
diff changeset
82 self.initial = False
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
83 if not os.path.isdir(self.index_location):
763
0dad296d2a57 extended trending languages to more entries, implemented new faster and "fancy"
Marcin Kuzminski <marcin@python-works.com>
parents: 691
diff changeset
84 os.makedirs(self.index_location)
465
e01a85f9fc90 fixed initial whoosh indexer. Build full index on first run even with incremental flag
Marcin Kuzminski <marcin@python-works.com>
parents: 452
diff changeset
85 log.info('Cannot run incremental index since it does not'
e01a85f9fc90 fixed initial whoosh indexer. Build full index on first run even with incremental flag
Marcin Kuzminski <marcin@python-works.com>
parents: 452
diff changeset
86 ' yet exist running full build')
e01a85f9fc90 fixed initial whoosh indexer. Build full index on first run even with incremental flag
Marcin Kuzminski <marcin@python-works.com>
parents: 452
diff changeset
87 self.initial = True
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
88
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
89 def get_paths(self, repo):
683
341beaa9edba Implemented whoosh index building as paster command.
Marcin Kuzminski <marcin@python-works.com>
parents: 662
diff changeset
90 """recursive walk in root dir and return a set of all path in that dir
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
91 based on repository walk function
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
92 """
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
93 index_paths_ = set()
567
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
94 try:
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
95 for topnode, dirs, files in repo.walk('/', 'tip'):
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
96 for f in files:
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
97 index_paths_.add(jn(repo.path, f.path))
567
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
98 for dir in dirs:
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
99 for f in files:
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
100 index_paths_.add(jn(repo.path, f.path))
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
101
567
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
102 except RepositoryError:
80dc0a23edf7 fixed whoosh failure on new repository
Marcin Kuzminski <marcin@python-works.com>
parents: 561
diff changeset
103 pass
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
104 return index_paths_
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
105
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
106 def get_node(self, repo, path):
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
107 n_path = path[len(repo.path) + 1:]
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
108 node = repo.get_changeset().get_node(n_path)
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
109 return node
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
110
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
111 def get_node_mtime(self, node):
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
112 return mktime(node.last_changeset.date.timetuple())
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
113
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
114 def add_doc(self, writer, path, repo):
683
341beaa9edba Implemented whoosh index building as paster command.
Marcin Kuzminski <marcin@python-works.com>
parents: 662
diff changeset
115 """Adding doc to writer this function itself fetches data from
341beaa9edba Implemented whoosh index building as paster command.
Marcin Kuzminski <marcin@python-works.com>
parents: 662
diff changeset
116 the instance of vcs backend"""
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
117 node = self.get_node(repo, path)
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
118
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
119 #we just index the content of chosen files
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
120 if node.extension in INDEX_EXTENSIONS:
436
28f19fa562df updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents: 411
diff changeset
121 log.debug(' >> %s [WITH CONTENT]' % path)
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
122 u_content = node.content
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
123 else:
436
28f19fa562df updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents: 411
diff changeset
124 log.debug(' >> %s' % path)
28f19fa562df updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents: 411
diff changeset
125 #just index file name without it's content
28f19fa562df updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents: 411
diff changeset
126 u_content = u''
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
127
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
128 writer.add_document(owner=unicode(repo.contact),
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
129 repository=safe_unicode(repo.name),
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
130 path=safe_unicode(path),
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
131 content=u_content,
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
132 modtime=self.get_node_mtime(node),
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
133 extension=node.extension)
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
134
441
c59c4d4323e7 added support for broken symlinks in whoosh indexer
Marcin Kuzminski <marcin@python-works.com>
parents: 436
diff changeset
135
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
136 def build_index(self):
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
137 if os.path.exists(self.index_location):
560
3072935bdeed rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents: 557
diff changeset
138 log.debug('removing previous index')
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
139 rmtree(self.index_location)
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
140
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
141 if not os.path.exists(self.index_location):
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
142 os.mkdir(self.index_location)
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
143
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
144 idx = create_in(self.index_location, SCHEMA, indexname=IDX_NAME)
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
145 writer = idx.writer()
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
146
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
147 for cnt, repo in enumerate(self.repo_paths.values()):
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
148 log.debug('building index @ %s' % repo.path)
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
149
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
150 for idx_path in self.get_paths(repo):
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
151 self.add_doc(writer, idx_path, repo)
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
152
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
153 log.debug('>> COMMITING CHANGES <<')
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
154 writer.commit(merge=True)
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
155 log.debug('>>> FINISHED BUILDING INDEX <<<')
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
156
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
157
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
158 def update_index(self):
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
159 log.debug('STARTING INCREMENTAL INDEXING UPDATE')
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
160
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
161 idx = open_dir(self.index_location, indexname=self.indexname)
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
162 # The set of all paths in the index
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
163 indexed_paths = set()
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
164 # The set of all paths we need to re-index
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
165 to_index = set()
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
166
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
167 reader = idx.reader()
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
168 writer = idx.writer()
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
169
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
170 # Loop over the stored fields in the index
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
171 for fields in reader.all_stored_fields():
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
172 indexed_path = fields['path']
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
173 indexed_paths.add(indexed_path)
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
174
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
175 repo = self.repo_paths[fields['repository']]
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
176
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
177 try:
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
178 node = self.get_node(repo, indexed_path)
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
179 except ChangesetError:
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
180 # This file was deleted since it was indexed
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
181 log.debug('removing from index %s' % indexed_path)
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
182 writer.delete_by_term('path', indexed_path)
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
183
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
184 else:
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
185 # Check if this file was changed since it was indexed
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
186 indexed_time = fields['modtime']
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
187 mtime = self.get_node_mtime(node)
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
188 if mtime > indexed_time:
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
189 # The file has changed, delete it and add it to the list of
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
190 # files to reindex
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
191 log.debug('adding to reindex list %s' % indexed_path)
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
192 writer.delete_by_term('path', indexed_path)
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
193 to_index.add(indexed_path)
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
194
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
195 # Loop over the files in the filesystem
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
196 # Assume we have a function that gathers the filenames of the
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
197 # documents to be indexed
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
198 for repo in self.repo_paths.values():
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
199 for path in self.get_paths(repo):
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
200 if path in to_index or path not in indexed_paths:
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
201 # This is either a file that's changed, or a new file
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
202 # that wasn't indexed before. So index it!
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
203 self.add_doc(writer, path, repo)
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
204 log.debug('re indexing %s' % path)
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
205
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
206 log.debug('>> COMMITING CHANGES <<')
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
207 writer.commit(merge=True)
561
5f3b967d9d10 fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
Marcin Kuzminski <marcin@python-works.com>
parents: 560
diff changeset
208 log.debug('>>> FINISHED REBUILDING INDEX <<<')
631
05528ad948c4 Hacking for git support,and new faster repo scan
Marcin Kuzminski <marcin@python-works.com>
parents: 629
diff changeset
209
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
210 def run(self, full_index=False):
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
211 """Run daemon"""
465
e01a85f9fc90 fixed initial whoosh indexer. Build full index on first run even with incremental flag
Marcin Kuzminski <marcin@python-works.com>
parents: 452
diff changeset
212 if full_index or self.initial:
406
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
213 self.build_index()
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
214 else:
b153a51b1d3b Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff changeset
215 self.update_index()