Mercurial > kallithea
annotate rhodecode/lib/indexers/daemon.py @ 561:5f3b967d9d10
fixed reindexing, and made some optimizations to reuse repo instances from repo scann list.
author | Marcin Kuzminski <marcin@python-works.com> |
---|---|
date | Sat, 09 Oct 2010 01:11:44 +0200 |
parents | 3072935bdeed |
children | 80dc0a23edf7 |
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 |
547
1e757ac98988
renamed project to rhodecode
Marcin Kuzminski <marcin@python-works.com>
parents:
497
diff
changeset
|
35 from rhodecode.lib.pidlock import LockHeld, DaemonLock |
1e757ac98988
renamed project to rhodecode
Marcin Kuzminski <marcin@python-works.com>
parents:
497
diff
changeset
|
36 from rhodecode.model.hg_model import HgModel |
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 |
547
1e757ac98988
renamed project to rhodecode
Marcin Kuzminski <marcin@python-works.com>
parents:
497
diff
changeset
|
40 from rhodecode.lib.indexers import INDEX_EXTENSIONS, IDX_LOCATION, 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 |
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
|
43 from vcs.exceptions import ChangesetError |
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 def scan_paths(root_location): |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
65 return HgModel.repo_scan('/', root_location, None, True) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
66 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
67 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
|
68 """ |
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
69 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
|
70 """ |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
71 |
411
9b67cebe6609
some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents:
407
diff
changeset
|
72 def __init__(self, indexname='HG_INDEX', repo_location=None): |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
73 self.indexname = indexname |
411
9b67cebe6609
some fixes to whoosh indexer daemon
Marcin Kuzminski <marcin@python-works.com>
parents:
407
diff
changeset
|
74 self.repo_location = repo_location |
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
|
75 self.repo_paths = scan_paths(self.repo_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
|
76 self.initial = False |
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
|
77 if not os.path.isdir(IDX_LOCATION): |
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
|
78 os.mkdir(IDX_LOCATION) |
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
|
79 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
|
80 ' 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
|
81 self.initial = True |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
82 |
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
|
83 def get_paths(self, repo): |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
84 """ |
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
85 recursive walk in root dir and return a set of all path in that dir |
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
86 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
|
87 """ |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
88 index_paths_ = set() |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
89 for topnode, dirs, files in repo.walk('/', 'tip'): |
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
90 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
|
91 index_paths_.add(jn(repo.path, f.path)) |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
92 for dir in dirs: |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
93 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
|
94 index_paths_.add(jn(repo.path, f.path)) |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
95 |
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
96 return index_paths_ |
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 |
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
|
98 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
|
99 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
|
100 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
|
101 return 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
|
102 |
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
|
103 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
|
104 return mktime(node.last_changeset.date.timetuple()) |
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
|
105 |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
106 def add_doc(self, writer, path, repo): |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
107 """Adding doc to writer""" |
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
|
108 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
|
109 |
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
110 #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
|
111 if node.extension in INDEX_EXTENSIONS: |
436
28f19fa562df
updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents:
411
diff
changeset
|
112 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
|
113 u_content = node.content |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
114 else: |
436
28f19fa562df
updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents:
411
diff
changeset
|
115 log.debug(' >> %s' % path) |
28f19fa562df
updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents:
411
diff
changeset
|
116 #just index file name without it's content |
28f19fa562df
updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents:
411
diff
changeset
|
117 u_content = u'' |
441
c59c4d4323e7
added support for broken symlinks in whoosh indexer
Marcin Kuzminski <marcin@python-works.com>
parents:
436
diff
changeset
|
118 |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
119 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
|
120 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
|
121 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
|
122 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
|
123 modtime=self.get_node_mtime(node), |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
124 extension=node.extension) |
441
c59c4d4323e7
added support for broken symlinks in whoosh indexer
Marcin Kuzminski <marcin@python-works.com>
parents:
436
diff
changeset
|
125 |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
126 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
127 def build_index(self): |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
128 if os.path.exists(IDX_LOCATION): |
560
3072935bdeed
rewrote whoosh indexing to run internal repository.walk() instead of filesystem.
Marcin Kuzminski <marcin@python-works.com>
parents:
557
diff
changeset
|
129 log.debug('removing previous index') |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
130 rmtree(IDX_LOCATION) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
131 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
132 if not os.path.exists(IDX_LOCATION): |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
133 os.mkdir(IDX_LOCATION) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
134 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
135 idx = create_in(IDX_LOCATION, SCHEMA, indexname=IDX_NAME) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
136 writer = idx.writer() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
137 |
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
|
138 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
|
139 log.debug('building index @ %s' % repo.path) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
140 |
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
|
141 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
|
142 self.add_doc(writer, idx_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
|
143 |
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
|
144 log.debug('>> COMMITING CHANGES <<') |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
145 writer.commit(merge=True) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
146 log.debug('>>> FINISHED BUILDING INDEX <<<') |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
147 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
148 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
149 def update_index(self): |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
150 log.debug('STARTING INCREMENTAL INDEXING UPDATE') |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
151 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
152 idx = open_dir(IDX_LOCATION, indexname=self.indexname) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
153 # 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
|
154 indexed_paths = set() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
155 # 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
|
156 to_index = set() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
157 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
158 reader = idx.reader() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
159 writer = idx.writer() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
160 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
161 # 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
|
162 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
|
163 indexed_path = fields['path'] |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
164 indexed_paths.add(indexed_path) |
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
|
165 |
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
|
166 repo = self.repo_paths[fields['repository']] |
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
|
167 |
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
|
168 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
|
169 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
|
170 except ChangesetError: |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
171 # 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
|
172 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
|
173 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
|
174 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
175 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
|
176 # 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
|
177 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
|
178 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
|
179 if mtime > indexed_time: |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
180 # 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
|
181 # files to reindex |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
182 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
|
183 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
|
184 to_index.add(indexed_path) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
185 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
186 # 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
|
187 # 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
|
188 # 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
|
189 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
|
190 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
|
191 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
|
192 # 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
|
193 # 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
|
194 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
|
195 log.debug('re indexing %s' % 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
|
196 |
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
|
197 log.debug('>> COMMITING CHANGES <<') |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
198 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
|
199 log.debug('>>> FINISHED REBUILDING INDEX <<<') |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
200 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
201 def run(self, full_index=False): |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
202 """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
|
203 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
|
204 self.build_index() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
205 else: |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
206 self.update_index() |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
207 |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
208 if __name__ == "__main__": |
451
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
209 arg = sys.argv[1:] |
452
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
210 if len(arg) != 2: |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
211 sys.stderr.write('Please specify indexing type [full|incremental]' |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
212 'and path to repositories as script args \n') |
451
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
213 sys.exit() |
452
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
214 |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
215 |
451
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
216 if arg[0] == 'full': |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
217 full_index = True |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
218 elif arg[0] == 'incremental': |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
219 # False means looking just for changes |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
220 full_index = False |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
221 else: |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
222 sys.stdout.write('Please use [full|incremental]' |
452
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
223 ' as script first arg \n') |
451
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
224 sys.exit() |
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
225 |
452
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
226 if not os.path.isdir(arg[1]): |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
227 sys.stderr.write('%s is not a valid path \n' % arg[1]) |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
228 sys.exit() |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
229 else: |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
230 if arg[1].endswith('/'): |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
231 repo_location = arg[1] + '*' |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
232 else: |
f19d3ee89335
updated whoosh indexer to take path as second argument
Marcin Kuzminski <marcin@python-works.com>
parents:
451
diff
changeset
|
233 repo_location = arg[1] + '/*' |
451
d726f62f886e
updated whoosh indexer to take index building argument type
Marcin Kuzminski <marcin@python-works.com>
parents:
443
diff
changeset
|
234 |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
235 try: |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
236 l = DaemonLock() |
436
28f19fa562df
updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents:
411
diff
changeset
|
237 WhooshIndexingDaemon(repo_location=repo_location)\ |
28f19fa562df
updated config files,
Marcin Kuzminski <marcin@python-works.com>
parents:
411
diff
changeset
|
238 .run(full_index=full_index) |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
239 l.release() |
483
a9e50dce3081
Removed config names from whoosh and celery,
Marcin Kuzminski <marcin@python-works.com>
parents:
465
diff
changeset
|
240 reload(logging) |
406
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
241 except LockHeld: |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
242 sys.exit(1) |
b153a51b1d3b
Implemented search using whoosh. Still as experimental option.
Marcin Kuzminski <marcin@python-works.com>
parents:
diff
changeset
|
243 |