View GitHub Repository Annif

Schedule Inspection
Subscribe
Completed
Push — master ( 131b99...8a6527 )
by Osma
28s queued 10s
created

annif.vocab.AnnifVocabulary.subjects()   A

↳ Parent: annif.vocab

Complexity

Conditions 3

Size

Total Lines 11
Code Lines 10

Duplication

Lines 0
Ratio 0 %

Importance

Changes 0
Metric Value
eloc 10
dl 0
loc 11
rs 9.9
c 0
b 0
f 0
cc 3
nop 1
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):

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    41
                

                                    
                                                    
                
                
                
        """load subjects from a subject corpus"""

            


            
                                                                                                            
                            
            
                                    
            
            

                
                    42
                

                                    
                                                    
                
                
                

            


            
                                                                                                            
                                                                
            
                                    
            
            

                
                    43
                

                                    
                                                    
                
                
                
        self._create_subject_index(subject_corpus)

            


            
                                                        
            
                                    
            
            

                
                    44