Show More
@@ -1,188 +1,199 b'' | |||
|
1 | 1 | #!/usr/bin/env python |
|
2 | 2 | # encoding: utf-8 |
|
3 | 3 | # whoosh indexer daemon for hg-app |
|
4 | 4 | # Copyright (C) 2009-2010 Marcin Kuzminski <marcin@python-works.com> |
|
5 | 5 | # |
|
6 | 6 | # This program is free software; you can redistribute it and/or |
|
7 | 7 | # modify it under the terms of the GNU General Public License |
|
8 | 8 | # as published by the Free Software Foundation; version 2 |
|
9 | 9 | # of the License or (at your opinion) any later version of the license. |
|
10 | 10 | # |
|
11 | 11 | # This program is distributed in the hope that it will be useful, |
|
12 | 12 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
13 | 13 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
14 | 14 | # GNU General Public License for more details. |
|
15 | 15 | # |
|
16 | 16 | # You should have received a copy of the GNU General Public License |
|
17 | 17 | # along with this program; if not, write to the Free Software |
|
18 | 18 | # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, |
|
19 | 19 | # MA 02110-1301, USA. |
|
20 | 20 | """ |
|
21 | 21 | Created on Jan 26, 2010 |
|
22 | 22 | |
|
23 | 23 | @author: marcink |
|
24 | 24 | A deamon will read from task table and run tasks |
|
25 | 25 | """ |
|
26 | 26 | import sys |
|
27 | 27 | import os |
|
28 | 28 | from os.path import dirname as dn |
|
29 | 29 | from os.path import join as jn |
|
30 | 30 | |
|
31 | 31 | #to get the pylons_app import |
|
32 | 32 | project_path = dn(dn(dn(dn(os.path.realpath(__file__))))) |
|
33 | 33 | sys.path.append(project_path) |
|
34 | 34 | |
|
35 | 35 | from pidlock import LockHeld, DaemonLock |
|
36 | 36 | import traceback |
|
37 | 37 | from pylons_app.config.environment import load_environment |
|
38 | 38 | from pylons_app.model.hg_model import HgModel |
|
39 | 39 | from whoosh.index import create_in, open_dir |
|
40 | 40 | from shutil import rmtree |
|
41 | 41 | from pylons_app.lib.indexers import ANALYZER, INDEX_EXTENSIONS, IDX_LOCATION, \ |
|
42 | 42 | SCHEMA, IDX_NAME |
|
43 | 43 | |
|
44 | 44 | import logging |
|
45 | 45 | import logging.config |
|
46 | 46 | logging.config.fileConfig(jn(project_path, 'development.ini')) |
|
47 | 47 | log = logging.getLogger('whooshIndexer') |
|
48 | 48 | |
|
49 | 49 | def scan_paths(root_location): |
|
50 | 50 | return HgModel.repo_scan('/', root_location, None, True) |
|
51 | 51 | |
|
52 | 52 | class WhooshIndexingDaemon(object): |
|
53 | 53 | """Deamon for atomic jobs""" |
|
54 | 54 | |
|
55 | 55 | def __init__(self, indexname='HG_INDEX', repo_location=None): |
|
56 | 56 | self.indexname = indexname |
|
57 | 57 | self.repo_location = repo_location |
|
58 | 58 | |
|
59 | 59 | def get_paths(self, root_dir): |
|
60 | 60 | """recursive walk in root dir and return a set of all path in that dir |
|
61 | 61 | excluding files in .hg dir""" |
|
62 | 62 | index_paths_ = set() |
|
63 | 63 | for path, dirs, files in os.walk(root_dir): |
|
64 | 64 | if path.find('.hg') == -1: |
|
65 | 65 | for f in files: |
|
66 | 66 | index_paths_.add(jn(path, f)) |
|
67 | 67 | |
|
68 | 68 | return index_paths_ |
|
69 | 69 | |
|
70 | 70 | def add_doc(self, writer, path, repo): |
|
71 | 71 | """Adding doc to writer""" |
|
72 | 72 | |
|
73 | 73 | ext = unicode(path.split('/')[-1].split('.')[-1].lower()) |
|
74 | 74 | #we just index the content of choosen files |
|
75 | 75 | if ext in INDEX_EXTENSIONS: |
|
76 | 76 | log.debug(' >> %s [WITH CONTENT]' % path) |
|
77 | 77 | fobj = open(path, 'rb') |
|
78 | 78 | content = fobj.read() |
|
79 | 79 | fobj.close() |
|
80 | 80 | try: |
|
81 | 81 | u_content = unicode(content) |
|
82 | 82 | except UnicodeDecodeError: |
|
83 | 83 | #incase we have a decode error just represent as byte string |
|
84 | 84 | u_content = unicode(str(content).encode('string_escape')) |
|
85 | 85 | else: |
|
86 | 86 | log.debug(' >> %s' % path) |
|
87 | 87 | #just index file name without it's content |
|
88 | 88 | u_content = u'' |
|
89 | 89 |
|
|
90 | ||
|
91 | ||
|
92 | try: | |
|
93 | os.stat(path) | |
|
90 | 94 | writer.add_document(owner=unicode(repo.contact), |
|
91 | 95 | repository=u"%s" % repo.name, |
|
92 | 96 | path=u"%s" % path, |
|
93 | 97 | content=u_content, |
|
94 | 98 | modtime=os.path.getmtime(path), |
|
95 | 99 | extension=ext) |
|
100 | except OSError, e: | |
|
101 | import errno | |
|
102 | if e.errno == errno.ENOENT: | |
|
103 | log.debug('path %s does not exist or is a broken symlink' % path) | |
|
104 | else: | |
|
105 | raise e | |
|
106 | ||
|
96 | 107 | |
|
97 | 108 | def build_index(self): |
|
98 | 109 | if os.path.exists(IDX_LOCATION): |
|
99 | 110 | log.debug('removing previos index') |
|
100 | 111 | rmtree(IDX_LOCATION) |
|
101 | 112 | |
|
102 | 113 | if not os.path.exists(IDX_LOCATION): |
|
103 | 114 | os.mkdir(IDX_LOCATION) |
|
104 | 115 | |
|
105 | 116 | idx = create_in(IDX_LOCATION, SCHEMA, indexname=IDX_NAME) |
|
106 | 117 | writer = idx.writer() |
|
107 | 118 | |
|
108 | 119 | for cnt, repo in enumerate(scan_paths(self.repo_location).values()): |
|
109 | 120 | log.debug('building index @ %s' % repo.path) |
|
110 | 121 | |
|
111 | 122 | for idx_path in self.get_paths(repo.path): |
|
112 | 123 | self.add_doc(writer, idx_path, repo) |
|
113 | 124 | writer.commit(merge=True) |
|
114 | 125 | |
|
115 | 126 | log.debug('>>> FINISHED BUILDING INDEX <<<') |
|
116 | 127 | |
|
117 | 128 | |
|
118 | 129 | def update_index(self): |
|
119 | 130 | log.debug('STARTING INCREMENTAL INDEXING UPDATE') |
|
120 | 131 | |
|
121 | 132 | idx = open_dir(IDX_LOCATION, indexname=self.indexname) |
|
122 | 133 | # The set of all paths in the index |
|
123 | 134 | indexed_paths = set() |
|
124 | 135 | # The set of all paths we need to re-index |
|
125 | 136 | to_index = set() |
|
126 | 137 | |
|
127 | 138 | reader = idx.reader() |
|
128 | 139 | writer = idx.writer() |
|
129 | 140 | |
|
130 | 141 | # Loop over the stored fields in the index |
|
131 | 142 | for fields in reader.all_stored_fields(): |
|
132 | 143 | indexed_path = fields['path'] |
|
133 | 144 | indexed_paths.add(indexed_path) |
|
134 | 145 | |
|
135 | 146 | if not os.path.exists(indexed_path): |
|
136 | 147 | # This file was deleted since it was indexed |
|
137 | 148 | log.debug('removing from index %s' % indexed_path) |
|
138 | 149 | writer.delete_by_term('path', indexed_path) |
|
139 | 150 | |
|
140 | 151 | else: |
|
141 | 152 | # Check if this file was changed since it |
|
142 | 153 | # was indexed |
|
143 | 154 | indexed_time = fields['modtime'] |
|
144 | 155 | |
|
145 | 156 | mtime = os.path.getmtime(indexed_path) |
|
146 | 157 | |
|
147 | 158 | if mtime > indexed_time: |
|
148 | 159 | |
|
149 | 160 | # The file has changed, delete it and add it to the list of |
|
150 | 161 | # files to reindex |
|
151 | 162 | log.debug('adding to reindex list %s' % indexed_path) |
|
152 | 163 | writer.delete_by_term('path', indexed_path) |
|
153 | 164 | to_index.add(indexed_path) |
|
154 | 165 | #writer.commit() |
|
155 | 166 | |
|
156 | 167 | # Loop over the files in the filesystem |
|
157 | 168 | # Assume we have a function that gathers the filenames of the |
|
158 | 169 | # documents to be indexed |
|
159 | 170 | for repo in scan_paths(self.repo_location).values(): |
|
160 | 171 | for path in self.get_paths(repo.path): |
|
161 | 172 | if path in to_index or path not in indexed_paths: |
|
162 | 173 | # This is either a file that's changed, or a new file |
|
163 | 174 | # that wasn't indexed before. So index it! |
|
164 | 175 | self.add_doc(writer, path, repo) |
|
165 | 176 | log.debug('reindexing %s' % path) |
|
166 | 177 | |
|
167 | 178 | writer.commit(merge=True) |
|
168 | 179 | #idx.optimize() |
|
169 | 180 | log.debug('>>> FINISHED <<<') |
|
170 | 181 | |
|
171 | 182 | def run(self, full_index=False): |
|
172 | 183 | """Run daemon""" |
|
173 | 184 | if full_index: |
|
174 | 185 | self.build_index() |
|
175 | 186 | else: |
|
176 | 187 | self.update_index() |
|
177 | 188 | |
|
178 | 189 | if __name__ == "__main__": |
|
179 | 190 | repo_location = '/home/marcink/hg_repos/*' |
|
180 | 191 | full_index = True # False means looking just for changes |
|
181 | 192 | try: |
|
182 | 193 | l = DaemonLock() |
|
183 | 194 | WhooshIndexingDaemon(repo_location=repo_location)\ |
|
184 | 195 | .run(full_index=full_index) |
|
185 | 196 | l.release() |
|
186 | 197 | except LockHeld: |
|
187 | 198 | sys.exit(1) |
|
188 | 199 |
General Comments 0
You need to be logged in to leave comments.
Login now