|
1
|
|
|
"""Vocabulary management functionality for Annif""" |
|
2
|
|
|
|
|
3
|
|
|
import os.path |
|
4
|
|
|
import annif |
|
5
|
|
|
import annif.corpus |
|
6
|
|
|
import annif.util |
|
7
|
|
|
from annif.datadir import DatadirMixin |
|
8
|
|
|
from annif.exception import NotInitializedException |
|
9
|
|
|
|
|
10
|
|
|
logger = annif.logger |
|
11
|
|
|
|
|
12
|
|
|
|
|
13
|
|
|
class AnnifVocabulary(DatadirMixin): |
|
14
|
|
|
"""Class representing a subject vocabulary which can be used by multiple |
|
15
|
|
|
Annif projects.""" |
|
16
|
|
|
|
|
17
|
|
|
# defaults for uninitialized instances |
|
18
|
|
|
_subjects = None |
|
19
|
|
|
|
|
20
|
|
|
def __init__(self, vocab_id, datadir): |
|
21
|
|
|
DatadirMixin.__init__(self, datadir, 'vocabs', vocab_id) |
|
22
|
|
|
self.vocab_id = vocab_id |
|
23
|
|
|
|
|
24
|
|
|
def _create_subject_index(self, subject_corpus): |
|
25
|
|
|
self._subjects = annif.corpus.SubjectIndex(subject_corpus) |
|
26
|
|
|
annif.util.atomic_save(self._subjects, self.datadir, 'subjects') |
|
27
|
|
|
|
|
28
|
|
|
@property |
|
29
|
|
|
def subjects(self): |
|
30
|
|
|
if self._subjects is None: |
|
31
|
|
|
path = os.path.join(self.datadir, 'subjects') |
|
32
|
|
|
if os.path.exists(path): |
|
33
|
|
|
logger.debug('loading subjects from %s', path) |
|
34
|
|
|
self._subjects = annif.corpus.SubjectIndex.load(path) |
|
35
|
|
|
else: |
|
36
|
|
|
raise NotInitializedException( |
|
37
|
|
|
"subject file {} not found".format(path)) |
|
38
|
|
|
return self._subjects |
|
39
|
|
|
|
|
40
|
|
|
def load_vocabulary(self, subject_corpus, language): |
|
41
|
|
|
"""load subjects from a subject corpus and save them into a |
|
42
|
|
|
SKOS/Turtle file for later use""" |
|
43
|
|
|
|
|
44
|
|
|
self._create_subject_index(subject_corpus) |
|
45
|
|
|
subject_corpus.save_skos(os.path.join(self.datadir, 'subjects.ttl'), |
|
46
|
|
|
language) |
|
47
|
|
|
|
|
48
|
|
|
def as_skos(self): |
|
49
|
|
|
"""return the vocabulary as a file object, in SKOS/Turtle syntax""" |
|
50
|
|
|
return open(os.path.join(self.datadir, 'subjects.ttl'), 'rb') |
|
51
|
|
|
|
NatLibFi /
Annif
