Merge branch 'tools-cleanup' into 'master'

Cleanup arkindex.project.tools

See merge request !868

arkindex/himanis/tei.py

 from itertools import groupby
 from arkindex_common.tei import TeiElement, TeiParser as BaseTeiParser
 from arkindex_common.enums import MetaType
-from arkindex.project.tools import find_closest
 from arkindex.project.triggers import reindex_start
 from arkindex.documents.models import Element, Entity, DataSource, MLToolType
 import logging
-
+import Levenshtein
 
 logger = logging.getLogger(__name__)
 
 
+def find_closest(name, queryset, min_ratio=0.8):
+    '''
+    Find closest element using Levenshtein distance.
+    '''
+    assert isinstance(name, str)
+
+    scores = sorted([
+        (element, Levenshtein.ratio(name, element.name))
+        for element in queryset
+    ], key=lambda x: x[1], reverse=True)
+
+    best_element, best_score = scores[0]
+    if best_score >= min_ratio:
+        return best_element
+
+
 class TeiParser(BaseTeiParser):
 
     def __init__(self, path, corpus, folder_type, element_type):

arkindex/project/tests/test_tools.py

-from unittest import TestCase
-from arkindex.project.tools import elasticsearch_escape
-
-
-class TestTools(TestCase):
-
-    def test_elasticsearch_escape(self):
-        self.assertEqual(elasticsearch_escape('abcdef'), 'abcdef')
-        self.assertEqual(elasticsearch_escape('aaaaa+b'), 'aaaaa\\\\+b')
-        self.assertEqual(elasticsearch_escape('\\'), '\\\\')
-        self.assertEqual(elasticsearch_escape('[]'), '\\\\[\\\\]')
-        self.assertEqual(elasticsearch_escape(' \" '), ' \\\\\\" ')
-        self.assertEqual(elasticsearch_escape('&|'), '&|')
-        self.assertEqual(elasticsearch_escape('&&||'), '\\\\&&\\\\||')
-        self.assertEqual(elasticsearch_escape('a<a>a'), 'aaa')

arkindex/project/tools.py

 from collections.abc import Sized, Iterable
 from django.urls import reverse
 from arkindex.documents.models import Element
-import Levenshtein
-import random
-import string
-import re
-
-
-def random_string(n):
-    '''
-    Build a random chars string of length n
-    '''
-    return ''.join([
-        random.choice(string.ascii_letters + string.digits)
-        for _ in range(n)
-    ])
-
-
-ES_ESCAPE_REGEX = re.compile(r'([+\-=\\><!(){}\[\]^"~*?:/]|\|\||&&)')
-
-
-def elasticsearch_escape(s):
-    '''
-    Escape a string for use by ElasticSearch
-    https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-query-string-query.html#_reserved_characters # noqa
-    '''
-    def replace(matchobj):
-        if matchobj.group(0) in ['<', '>']:
-            return ''
-        if matchobj.group(0) == '\\':
-            return '\\\\'
-        if matchobj.group(0) == '"':
-            return '\\\\\\"'
-        return '\\\\' + matchobj.group(0)
-
-    return ES_ESCAPE_REGEX.sub(replace, s)
-
-
-def find_closest(name, queryset, min_ratio=0.8):
-    '''
-    Find closest element using Levenshtein distance
-    '''
-    assert isinstance(name, str)
-
-    scores = sorted([
-        (element, Levenshtein.ratio(name, element.name))
-        for element in queryset
-    ], key=lambda x: x[1], reverse=True)
-
-    best_element, best_score = scores[0]
-    if best_score >= min_ratio:
-        return best_element
 
 
 def build_absolute_url(element, request, name, id_argument='pk', **kwargs):
     """
-    Build an absolute URL for a specified view using the element ID.
+    Build an absolute URL for a specified view using the element ID. Used by IIIF serializers.
     """
     kwargs[id_argument] = str(element.id)
     return request.build_absolute_uri(reverse(name, kwargs=kwargs))
 
 
-def read_file_range(path, start, end):
-    """
-    Read a specific range of lines from a file.
-    Line numbers start at 0. Like with Python slices, start is inclusive and end is exclusive.
-    """
-    # TODO: Optimize dis
-    lines = []
-    with open(path) as f:
-        for i, line in enumerate(f):
-            if i < start:
-                continue
-            elif i >= end:
-                break
-            lines.append(line)
-    return ''.join(lines)
-
-
 class disconnect_signal():
     """
     Context manager to temporarily disconnect a signal
@@ -94,22 +27,6 @@ class disconnect_signal():
         self.signal.connect(**self.kwargs)
 
 
-def get_or_instance(model, defaults={}, **filters):
-    """
-    Like model.objects.get_or_create(),
-    except it creates a Python instance that is not saved into DB.
-    """
-    try:
-        return model.objects.get(**filters), False
-    except model.DoesNotExist:
-        filters.update(defaults)
-        kwargs = {  # Filter to remove Django lookups
-            k: v for k, v in filters.items()
-            if '__' not in k
-        }
-        return model(**kwargs), True
-
-
 def build_tree(tree, *, corpus, type):
     """
     Build Element and ElementPath instances from a tree described by