|
1
|
|
|
"""Utility functions for Annif""" |
|
2
|
|
|
|
|
3
|
|
|
import glob |
|
4
|
|
|
import os |
|
5
|
|
|
import tempfile |
|
6
|
|
|
import numpy as np |
|
7
|
|
|
from annif import logger |
|
8
|
|
|
from annif.hit import VectorAnalysisResult |
|
9
|
|
|
|
|
10
|
|
|
|
|
11
|
|
|
def atomic_save(obj, dirname, filename, method=None): |
|
12
|
|
|
"""Save the given object (which must have a .save() method, unless the |
|
13
|
|
|
method parameter is given) into the given directory with the given |
|
14
|
|
|
filename, using a temporary file and renaming the temporary file to the |
|
15
|
|
|
final name.""" |
|
16
|
|
|
|
|
17
|
|
|
tempfd, tempfilename = tempfile.mkstemp(prefix=filename, dir=dirname) |
|
18
|
|
|
os.close(tempfd) |
|
19
|
|
|
logger.debug('saving %s to temporary file %s', str(obj), tempfilename) |
|
20
|
|
|
if method is not None: |
|
21
|
|
|
method(obj, tempfilename) |
|
22
|
|
|
else: |
|
23
|
|
|
obj.save(tempfilename) |
|
24
|
|
|
for fn in glob.glob(tempfilename + '*'): |
|
25
|
|
|
newname = fn.replace(tempfilename, os.path.join(dirname, filename)) |
|
26
|
|
|
logger.debug('renaming temporary file %s to %s', fn, newname) |
|
27
|
|
|
os.rename(fn, newname) |
|
28
|
|
|
|
|
29
|
|
|
|
|
30
|
|
|
def cleanup_uri(uri): |
|
31
|
|
|
"""remove angle brackets from a URI, if any""" |
|
32
|
|
|
if uri.startswith('<') and uri.endswith('>'): |
|
33
|
|
|
return uri[1:-1] |
|
34
|
|
|
return uri |
|
35
|
|
|
|
|
36
|
|
|
|
|
37
|
|
|
def merge_hits(weighted_hits, subject_index): |
|
38
|
|
|
"""Merge hits from multiple sources. Input is a sequence of WeightedHits |
|
39
|
|
|
objects. A SubjectIndex is needed to convert between subject IDs and URIs. |
|
40
|
|
|
Returns an AnalysisResult object.""" |
|
41
|
|
|
|
|
42
|
|
|
weights = [whit.weight for whit in weighted_hits] |
|
43
|
|
|
scores = [whit.hits.vector for whit in weighted_hits] |
|
44
|
|
|
result = np.average(scores, axis=0, weights=weights) |
|
45
|
|
|
return VectorAnalysisResult(result, subject_index) |
|
46
|
|
|
|
|
47
|
|
|
|
|
48
|
|
|
def parse_sources(sourcedef): |
|
49
|
|
|
"""parse a source definition such as 'src1:1.0,src2' into a sequence of |
|
50
|
|
|
tuples (src_id, weight)""" |
|
51
|
|
|
|
|
52
|
|
|
sources = [] |
|
53
|
|
|
for srcdef in sourcedef.strip().split(','): |
|
54
|
|
|
srcval = srcdef.strip().split(':') |
|
55
|
|
|
src_id = srcval[0] |
|
56
|
|
|
if len(srcval) > 1: |
|
57
|
|
|
weight = float(srcval[1]) |
|
58
|
|
|
else: |
|
59
|
|
|
weight = 1.0 |
|
60
|
|
|
sources.append((src_id, weight)) |
|
61
|
|
|
return sources |
|
62
|
|
|
|
NatLibFi /
Annif
