annotate roundup/backends/indexer_xapian.py @ 8511:b91d8a4e5850 permission-performance

close incorporated branch -- hg diff -r "ancestor(default,permission-performance)" -r permission-performance; returns nothing
author John Rouillard <rouilj@ieee.org>
date Tue, 27 Jan 2026 21:22:47 -0500
parents 1505f6ab86ec
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
1 ''' This implements the full-text indexer using the Xapian indexer.
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
2 '''
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
3 import os
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
4 import re
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
5 import time
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
6 import xapian
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
7
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
8 from roundup.anypy.strings import b2s, s2b
3544
5cd1c83dea50 Features and fixes.
Richard Jones <richard@users.sourceforge.net>
parents: 3295
diff changeset
9 from roundup.backends.indexer_common import Indexer as IndexerBase
6353
9d209d2b34ae Add indexer_language to change stemmer for xapian FTS indexer
John Rouillard <rouilj@ieee.org>
parents: 5964
diff changeset
10 from roundup.i18n import _
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
11
5491
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
12 # Note that Xapian always uses UTF-8 encoded string, see
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
13 # https://xapian.org/docs/bindings/python3/introduction.html#strings:
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
14 # "Where std::string is returned, it's always mapped to bytes in
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
15 # Python..."
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
16
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
17
3544
5cd1c83dea50 Features and fixes.
Richard Jones <richard@users.sourceforge.net>
parents: 3295
diff changeset
18 class Indexer(IndexerBase):
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
19 def __init__(self, db):
3544
5cd1c83dea50 Features and fixes.
Richard Jones <richard@users.sourceforge.net>
parents: 3295
diff changeset
20 IndexerBase.__init__(self, db)
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
21 self.db_path = db.config.DATABASE
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
22 self.reindex = 0
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
23 self.transaction_active = False
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
24
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
25 def _get_database(self):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
26 index = os.path.join(self.db_path, 'text-index')
5142
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
27 for n in range(10):
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
28 try:
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
29 # if successful return
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
30 return xapian.WritableDatabase(index, xapian.DB_CREATE_OR_OPEN)
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
31 except xapian.DatabaseLockError:
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
32 # adaptive sleep. Get longer as count increases.
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
33 time_to_sleep = 0.01 * (2 << min(5, n))
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
34 time.sleep(time_to_sleep)
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
35 # we are back to the for loop
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
36
93832cec4c31 issue2550839: Xapian, DatabaseLockError: Unable to get write lock on
John Rouillard <rouilj@ieee.org>
parents: 5108
diff changeset
37 # Get here only if we dropped out of the for loop.
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
38 raise xapian.DatabaseLockError(_(
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
39 "Unable to get lock after 10 retries on %s.") % index)
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
40
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
41 def save_index(self):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
42 '''Save the changes to the index.'''
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
43 if not self.transaction_active:
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
44 return
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
45 database = self._get_database()
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
46 database.commit_transaction()
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
47 self.transaction_active = False
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
48
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
49 def close(self):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
50 '''close the indexing database'''
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
51 pass
3887
c7363442cdbb change xapian stemmer to use "new" API
Justus Pendleton <jpend@users.sourceforge.net>
parents: 3555
diff changeset
52
3555
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
53 def rollback(self):
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
54 if not self.transaction_active:
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
55 return
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
56 database = self._get_database()
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
57 database.cancel_transaction()
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
58 self.transaction_active = False
91c495476db3 pre-release stuff and test fix
Richard Jones <richard@users.sourceforge.net>
parents: 3547
diff changeset
59
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
60 def force_reindex(self):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
61 '''Force a reindexing of the database. This essentially
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
62 empties the tables ids and index and sets a flag so
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
63 that the databases are reindexed'''
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
64 self.reindex = 1
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
65
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
66 def should_reindex(self):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
67 '''returns True if the indexes need to be rebuilt'''
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
68 return self.reindex
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
69
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
70 def add_text(self, identifier, text, mime_type='text/plain'):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
71 ''' "identifier" is (classname, itemid, property) '''
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
72 if mime_type != 'text/plain':
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
73 return
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
74 if not text:
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
75 text = ''
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
76
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
77 # open the database and start a transaction if needed
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
78 database = self._get_database()
4378
477f2a47cbca - Indexer Xapian, made Xapian 1.2 compatible.
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 4252
diff changeset
79
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
80 # XXX: Xapian now supports transactions,
4378
477f2a47cbca - Indexer Xapian, made Xapian 1.2 compatible.
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 4252
diff changeset
81 # but there is a call to save_index() missing.
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
82 # if not self.transaction_active:
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
83 # database.begin_transaction()
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
84 # self.transaction_active = True
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
85
6353
9d209d2b34ae Add indexer_language to change stemmer for xapian FTS indexer
John Rouillard <rouilj@ieee.org>
parents: 5964
diff changeset
86 stemmer = xapian.Stem(self.language)
3547
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
87
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
88 # We use the identifier twice: once in the actual "text" being
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
89 # indexed so we can search on it, and again as the "data" being
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
90 # indexed so we know what we're matching when we get results
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
91 identifier = s2b('%s:%s:%s' % identifier)
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
92
3547
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
93 # create the new document
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
94 doc = xapian.Document()
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
95 doc.set_data(identifier)
4511
931370d96c34 Xapian indexing improved:
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 4470
diff changeset
96 doc.add_term(identifier, 0)
3547
7728ee93efd2 fix reindexing in Xapian
Richard Jones <richard@users.sourceforge.net>
parents: 3544
diff changeset
97
4252
2ff6f39aa391 Indexers behaviour made more consistent regarding length of indexed words...
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 3932
diff changeset
98 for match in re.finditer(r'\b\w{%d,%d}\b'
2ff6f39aa391 Indexers behaviour made more consistent regarding length of indexed words...
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 3932
diff changeset
99 % (self.minlength, self.maxlength),
5964
5bf7b5debb09 Fix xapian indexer for unicode
John Rouillard <rouilj@ieee.org>
parents: 5491
diff changeset
100 text.upper(), re.UNICODE):
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
101 word = match.group(0)
3544
5cd1c83dea50 Features and fixes.
Richard Jones <richard@users.sourceforge.net>
parents: 3295
diff changeset
102 if self.is_stopword(word):
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
103 continue
5491
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
104 term = stemmer(s2b(word.lower()))
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
105 doc.add_posting(term, match.start(0))
4511
931370d96c34 Xapian indexing improved:
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 4470
diff changeset
106
931370d96c34 Xapian indexing improved:
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 4470
diff changeset
107 database.replace_document(identifier, doc)
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
108
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
109 def find(self, wordlist):
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
110 '''look up all the words in the wordlist.
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
111 If none are found return an empty dictionary
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
112 * more rules here
3887
c7363442cdbb change xapian stemmer to use "new" API
Justus Pendleton <jpend@users.sourceforge.net>
parents: 3555
diff changeset
113 '''
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
114 if not wordlist:
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
115 return {}
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
116
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
117 database = self._get_database()
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
118
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
119 enquire = xapian.Enquire(database)
6353
9d209d2b34ae Add indexer_language to change stemmer for xapian FTS indexer
John Rouillard <rouilj@ieee.org>
parents: 5964
diff changeset
120 stemmer = xapian.Stem(self.language)
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
121 terms = []
4252
2ff6f39aa391 Indexers behaviour made more consistent regarding length of indexed words...
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 3932
diff changeset
122 for term in [word.upper() for word in wordlist
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
123 if self.minlength <= len(word) <= self.maxlength]:
4252
2ff6f39aa391 Indexers behaviour made more consistent regarding length of indexed words...
Bernhard Reiter <Bernhard.Reiter@intevation.de>
parents: 3932
diff changeset
124 if not self.is_stopword(term):
5491
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
125 terms.append(stemmer(s2b(term.lower())))
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
126 query = xapian.Query(xapian.Query.OP_AND, terms)
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
127
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
128 enquire.set_query(query)
4841
3ff1a288fb9c issue2550583, issue2550635 Do not limit results with Xapian indexer
Thomas Arendsen Hein <thomas@intevation.de>
parents: 4570
diff changeset
129 matches = enquire.get_mset(0, database.get_doccount())
3295
a615cc230160 added Xapian indexer; replaces standard indexers if Xapian is available
Richard Jones <richard@users.sourceforge.net>
parents:
diff changeset
130
5491
e72573996caf fixed encoding issues for Xapian indexer
Christof Meerwald <cmeerw@cmeerw.org>
parents: 5142
diff changeset
131 return [tuple(b2s(m.document.get_data()).split(':'))
6979
0b6c54893ec5 flake8 - also added translation to an error string.
John Rouillard <rouilj@ieee.org>
parents: 6356
diff changeset
132 for m in matches]

Roundup Issue Tracker: http://roundup-tracker.org/