annotate src/memory/cache.py @ 2320:f4a0723042ee

jp (pubsub/search): fixed encoding for stdin in external commands
author Goffi <goffi@goffi.org>
date Sun, 09 Jul 2017 15:20:31 +0200
parents 1b3fbb76984b
children 8b37a62336c3
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
2109
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
1 #!/usr/bin/env python2
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
2 # -*- coding: utf-8 -*-
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
3
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
4 # SAT: a jabber client
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
5 # Copyright (C) 2009-2016 Jérôme Poisson (goffi@goffi.org)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
6
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
7 # This program is free software: you can redistribute it and/or modify
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
8 # it under the terms of the GNU Affero General Public License as published by
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
9 # the Free Software Foundation, either version 3 of the License, or
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
10 # (at your option) any later version.
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
11
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
12 # This program is distributed in the hope that it will be useful,
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
15 # GNU Affero General Public License for more details.
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
16
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
17 # You should have received a copy of the GNU Affero General Public License
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
18 # along with this program. If not, see <http://www.gnu.org/licenses/>.
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
19
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
20 from sat.core.log import getLogger
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
21 log = getLogger(__name__)
2166
1b3fbb76984b core (memory/cache): escape profile when creating cache dir
Goffi <goffi@goffi.org>
parents: 2116
diff changeset
22 from sat.tools.common import regex
2109
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
23 from sat.core import exceptions
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
24 from sat.core.constants import Const as C
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
25 import cPickle as pickle
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
26 import mimetypes
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
27 import os.path
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
28 import time
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
29
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
30
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
31 class Cache(object):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
32 """generic file caching"""
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
33
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
34 def __init__(self, host, profile=None):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
35 host = host
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
36 self.profile = profile
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
37 self.cache_dir = os.path.join(
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
38 host.memory.getConfig('', 'local_dir'),
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
39 C.CACHE_DIR,
2166
1b3fbb76984b core (memory/cache): escape profile when creating cache dir
Goffi <goffi@goffi.org>
parents: 2116
diff changeset
40 regex.pathEscape(profile or ''))
2109
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
41 if not os.path.exists(self.cache_dir):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
42 os.makedirs(self.cache_dir)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
43
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
44 def getPath(self, filename):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
45 """return cached file URL
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
46
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
47 @param filename(unicode): cached file name (cache data or actual file)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
48 """
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
49 if not filename or u'/' in filename:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
50 log.error(u"invalid char found in file name, hack attempt? name:{}".format(filename))
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
51 raise exceptions.DataError(u"Invalid char found")
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
52 return os.path.join(self.cache_dir, filename)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
53
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
54 def getFilePath(self, uid):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
55 """retrieve absolute path to file
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
56
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
57 @param uid(unicode): unique identifier of file
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
58 @return (unicode, None): absolute path to cached file
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
59 None if file is not in cache (or cache is invalid)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
60 """
2116
766dbbec56f2 core (memory/cache): geFilePath now return None when uid is empty
Goffi <goffi@goffi.org>
parents: 2109
diff changeset
61 uid = uid.strip()
766dbbec56f2 core (memory/cache): geFilePath now return None when uid is empty
Goffi <goffi@goffi.org>
parents: 2109
diff changeset
62 if not uid:
766dbbec56f2 core (memory/cache): geFilePath now return None when uid is empty
Goffi <goffi@goffi.org>
parents: 2109
diff changeset
63 return None
2109
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
64 cache_url = self.getPath(uid)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
65 if not os.path.exists(cache_url):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
66 return None
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
67
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
68 try:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
69 with open(cache_url, 'rb') as f:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
70 cache_data = pickle.load(f)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
71 except IOError:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
72 log.warning(u"can't read cache at {}".format(cache_url))
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
73 return None
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
74
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
75 except pickle.UnpicklingError:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
76 log.warning(u'invalid cache found at {}'.format(cache_url))
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
77 return None
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
78
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
79 try:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
80 eol = cache_data['eol']
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
81 except KeyError:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
82 log.warning(u'no End Of Life found for cached file {}'.format(uid))
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
83 eol = 0
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
84 if eol < time.time():
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
85 log.debug(u"removing expired cache (expired for {}s)".format(
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
86 time.time() - eol))
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
87 return None
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
88
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
89 return self.getPath(cache_data['filename'])
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
90
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
91 def cacheData(self, source, uid, mime_type=u'', max_age=None, filename=None):
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
92 """create cache metadata and file object to use for actual data
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
93
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
94 @param source(unicode): source of the cache (should be plugin's import_name)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
95 @param uid(unicode): an identifier of the file which must be unique
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
96 @param mime_type(unicode): MIME type of the file to cache
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
97 it will be used notably to guess file extension
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
98 @param max_age(int, None): maximum age in seconds
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
99 the cache metadata will have an "eol" (end of life)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
100 None to use default value
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
101 0 to ignore cache (file will be re-downloaded on each access)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
102 @param filename: if not None, will be used as filename
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
103 else one will be generated from uid and guessed extension
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
104 @return(file): file object opened in write mode
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
105 you have to close it yourself (hint: use with statement)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
106 """
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
107 # FIXME: is it needed to use a separate thread?
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
108 # probably not with the little data expected with BoB
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
109 cache_url = self.getPath(uid)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
110 ext = mimetypes.guess_extension(mime_type, strict=False)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
111 if ext is None:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
112 log.warning(u"can't find extension for MIME type {}".format(mime_type))
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
113 ext = '.dump'
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
114 if filename is None:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
115 filename = uid + ext
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
116 if max_age is None:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
117 max_age = C.DEFAULT_MAX_AGE
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
118 cache_data = {'source': source,
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
119 'filename': filename,
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
120 'eol': int(time.time()) + max_age,
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
121 'mime_type': mime_type,
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
122 }
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
123 file_path = self.getPath(filename)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
124
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
125 with open(cache_url, 'wb') as f:
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
126 pickle.dump(cache_data, f, protocol=2)
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
127
85f3e12e984d core (memory/cache): file caching handling, first draft:
Goffi <goffi@goffi.org>
parents:
diff changeset
128 return open(file_path, 'wb')