paradizelost/paperless-ngx
1
0
Fork 0
mirror of https://github.com/paperless-ngx/paperless-ngx.git synced 2025-04-02 13:45:10 -05:00
Code Issues Packages Projects Releases Wiki Activity

#48: make the tag matching smarter

Browse Source
This commit is contained in:
Daniel Quinn 2016-02-19 00:45:02 +00:00
parent 99be40a433
commit ec88ea73f6
2 changed files with 124 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/documents/models.py
View File

@ -105,21 +105,21 @@ class Tag(SluggedModel):
if self.matching_algorithm == self.MATCH_ALL:
for word in self.match.split(" "):
if word not in text:
if not re.search(r"\b{}\b".format(word), text):
return False
return True
if self.matching_algorithm == self.MATCH_ANY:
for word in self.match.split(" "):
if word in text:
if re.search(r"\b{}\b".format(word), text):
return True
return False
if self.matching_algorithm == self.MATCH_LITERAL:
return self.match in text
return bool(re.search(r"\b{}\b".format(self.match), text))
if self.matching_algorithm == self.MATCH_REGEX:
return re.search(re.compile(self.match), text)
return bool(re.search(re.compile(self.match), text))
raise NotImplementedError("Unsupported matching algorithm")

120
src/documents/tests/test_tags.py Normal file
View File

@ -0,0 +1,120 @@
from django.test import TestCase
from ..models import Tag
class TestTagMatching(TestCase):
def test_match_all(self):
t = Tag.objects.create(
name="Test 0",
match="alpha charlie gamma",
matching_algorithm=Tag.MATCH_ALL
)
self.assertFalse(t.matches("I have alpha in me"))
self.assertFalse(t.matches("I have charlie in me"))
self.assertFalse(t.matches("I have gamma in me"))
self.assertFalse(t.matches("I have alpha and charlie in me"))
self.assertTrue(t.matches("I have alpha, charlie, and gamma in me"))
self.assertFalse(t.matches("I have alphas, charlie, and gamma in me"))
self.assertFalse(t.matches("I have alphas in me"))
self.assertFalse(t.matches("I have bravo in me"))
t = Tag.objects.create(
name="Test 1",
match="12 34 56",
matching_algorithm=Tag.MATCH_ALL
)
self.assertFalse(t.matches("I have 12 in me"))
self.assertFalse(t.matches("I have 34 in me"))
self.assertFalse(t.matches("I have 56 in me"))
self.assertFalse(t.matches("I have 12 and 34 in me"))
self.assertTrue(t.matches("I have 12 34, and 56 in me"))
self.assertFalse(t.matches("I have 120, 34, and 56 in me"))
self.assertFalse(t.matches("I have 123456 in me"))
self.assertFalse(t.matches("I have 01234567 in me"))
def test_match_any(self):
t = Tag.objects.create(
name="Test 0",
match="alpha charlie gamma",
matching_algorithm=Tag.MATCH_ANY
)
self.assertTrue(t.matches("I have alpha in me"))
self.assertTrue(t.matches("I have charlie in me"))
self.assertTrue(t.matches("I have gamma in me"))
self.assertTrue(t.matches("I have alpha and charlie in me"))
self.assertFalse(t.matches("I have alphas in me"))
self.assertFalse(t.matches("I have bravo in me"))
t = Tag.objects.create(
name="Test 1",
match="12 34 56",
matching_algorithm=Tag.MATCH_ANY
)
self.assertTrue(t.matches("I have 12 in me"))
self.assertTrue(t.matches("I have 34 in me"))
self.assertTrue(t.matches("I have 56 in me"))
self.assertTrue(t.matches("I have 12 and 34 in me"))
self.assertTrue(t.matches("I have 12 34, and 56 in me"))
self.assertTrue(t.matches("I have 120, 34, and 560 in me"))
self.assertFalse(t.matches("I have 120, 340, and 560 in me"))
self.assertFalse(t.matches("I have 123456 in me"))
self.assertFalse(t.matches("I have 01234567 in me"))
def test_match_literal(self):
t = Tag.objects.create(
name="Test 0",
match="alpha charlie gamma",
matching_algorithm=Tag.MATCH_LITERAL
)
self.assertFalse(t.matches("I have alpha in me"))
self.assertFalse(t.matches("I have charlie in me"))
self.assertFalse(t.matches("I have gamma in me"))
self.assertFalse(t.matches("I have alpha and charlie in me"))
self.assertFalse(t.matches("I have alpha, charlie, and gamma in me"))
self.assertFalse(t.matches("I have alphas, charlie, and gamma in me"))
self.assertTrue(t.matches("I have 'alpha charlie gamma' in me"))
self.assertFalse(t.matches("I have alphas in me"))
self.assertFalse(t.matches("I have bravo in me"))
t = Tag.objects.create(
name="Test 1",
match="12 34 56",
matching_algorithm=Tag.MATCH_LITERAL
)
self.assertFalse(t.matches("I have 12 in me"))
self.assertFalse(t.matches("I have 34 in me"))
self.assertFalse(t.matches("I have 56 in me"))
self.assertFalse(t.matches("I have 12 and 34 in me"))
self.assertFalse(t.matches("I have 12 34, and 56 in me"))
self.assertFalse(t.matches("I have 120, 34, and 560 in me"))
self.assertFalse(t.matches("I have 120, 340, and 560 in me"))
self.assertFalse(t.matches("I have 123456 in me"))
self.assertFalse(t.matches("I have 01234567 in me"))
self.assertTrue(t.matches("I have 12 34 56 in me"))
def test_match_regex(self):
t = Tag.objects.create(
name="Test 0",
match="alpha\w+gamma",
matching_algorithm=Tag.MATCH_REGEX
)
self.assertFalse(t.matches("I have alpha in me"))
self.assertFalse(t.matches("I have gamma in me"))
self.assertFalse(t.matches("I have alpha and charlie in me"))
self.assertTrue(t.matches("I have alpha_and_gamma in me"))
self.assertTrue(t.matches("I have alphas_and_gamma in me"))
self.assertFalse(t.matches("I have alpha,and,gamma in me"))
self.assertFalse(t.matches("I have alpha and gamma in me"))
self.assertFalse(t.matches("I have alpha, charlie, and gamma in me"))
self.assertFalse(t.matches("I have alphas, charlie, and gamma in me"))
self.assertFalse(t.matches("I have alphas in me"))