paradizelost/paperless-ngx
1
0
Fork 0
mirror of https://github.com/paperless-ngx/paperless-ngx.git synced 2026-01-28 22:59:03 -06:00
Code Issues Packages Projects Releases Wiki Activity

Hooks up the class and fixes up the old testing. Includes ocr to date parser conversion we now do

Browse Source
This commit is contained in:
Trenton H
2026-01-28 14:13:29 -08:00
parent 9b7ae1c8ea
commit 32771391ad
6 changed files with 15 additions and 620 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
src/documents/consumer.py
View File

@@ -32,12 +32,12 @@ from documents.models import WorkflowTrigger
from documents.parsers import DocumentParser
from documents.parsers import ParseError
from documents.parsers import get_parser_class_for_mime_type
from documents.parsers import parse_date
from documents.permissions import set_permissions_for_object
from documents.plugins.base import AlwaysRunPluginMixin
from documents.plugins.base import ConsumeTaskPlugin
from documents.plugins.base import NoCleanupPluginMixin
from documents.plugins.base import NoSetupPluginMixin
from documents.plugins.date_parsing import get_date_parser
from documents.plugins.helpers import ProgressManager
from documents.plugins.helpers import ProgressStatusOptions
from documents.signals import document_consumption_finished
@@ -426,7 +426,8 @@ class ConsumerPlugin(
ProgressStatusOptions.WORKING,
ConsumerStatusShortMessage.PARSE_DATE,
)
date = parse_date(self.filename, text)
date_parser = get_date_parser()
date = next(date_parser.parse(self.filename, text), None)
archive_path = document_parser.get_archive_path()
page_count = document_parser.get_page_count(self.working_copy, mime_type)

74
src/documents/parsers.py
View File

@@ -9,22 +9,17 @@ import subprocess
import tempfile
from functools import lru_cache
from pathlib import Path
from re import Match
from typing import TYPE_CHECKING
from django.conf import settings
from django.utils import timezone
from documents.loggers import LoggingMixin
from documents.signals import document_consumer_declaration
from documents.utils import copy_file_with_basic_stats
from documents.utils import run_subprocess
from paperless.config import OcrConfig
from paperless.utils import ocr_to_dateparser_languages
if TYPE_CHECKING:
import datetime
from collections.abc import Iterator
# This regular expression will try to find dates in the document at
# hand and will match the following formats:
@@ -259,75 +254,6 @@ def make_thumbnail_from_pdf(in_path: Path, temp_dir: Path, logging_group=None) -
return out_path
def parse_date(filename, text) -> datetime.datetime | None:
return next(parse_date_generator(filename, text), None)
def parse_date_generator(filename, text) -> Iterator[datetime.datetime]:
"""
Returns the date of the document.
"""
def __parser(ds: str, date_order: str) -> datetime.datetime:
"""
Call dateparser.parse with a particular date ordering
"""
import dateparser
ocr_config = OcrConfig()
languages = settings.DATE_PARSER_LANGUAGES or ocr_to_dateparser_languages(
ocr_config.language,
)
return dateparser.parse(
ds,
settings={
"DATE_ORDER": date_order,
"PREFER_DAY_OF_MONTH": "first",
"RETURN_AS_TIMEZONE_AWARE": True,
"TIMEZONE": settings.TIME_ZONE,
},
locales=languages,
)
def __filter(date: datetime.datetime) -> datetime.datetime | None:
if (
date is not None
and date.year > 1900
and date <= timezone.now()
and date.date() not in settings.IGNORE_DATES
):
return date
return None
def __process_match(
match: Match[str],
date_order: str,
) -> datetime.datetime | None:
date_string = match.group(0)
try:
date = __parser(date_string, date_order)
except Exception:
# Skip all matches that do not parse to a proper date
date = None
return __filter(date)
def __process_content(content: str, date_order: str) -> Iterator[datetime.datetime]:
for m in re.finditer(DATE_REGEX, content):
date = __process_match(m, date_order)
if date is not None:
yield date
# if filename date parsing is enabled, search there first:
if settings.FILENAME_DATE_ORDER:
yield from __process_content(filename, settings.FILENAME_DATE_ORDER)
# Iterate through all regex matches in text and try to parse the date
yield from __process_content(text, settings.DATE_ORDER)
class ParseError(Exception):
pass

6
src/documents/plugins/date_parsing/__init__.py
View File

@@ -10,6 +10,7 @@ from django.utils import timezone
from documents.plugins.date_parsing.base import DateParserConfig
from documents.plugins.date_parsing.base import DateParserPluginBase
from documents.plugins.date_parsing.regex_parser import RegexDateParserPlugin
from paperless.utils import ocr_to_dateparser_languages
logger = logging.getLogger(__name__)
@@ -73,7 +74,10 @@ def get_date_parser() -> DateParserPluginBase:
# 2. Load configuration from settings
# TODO: Get the language from the settings and/or configuration object, depending
languages = settings.DATE_PARSER_LANGUAGES
languages = languages = (
settings.DATE_PARSER_LANGUAGES
or ocr_to_dateparser_languages(settings.OCR_LANGUAGE)
)
config = DateParserConfig(
languages=languages,

7
src/documents/tests/test_api_documents.py
View File

@@ -1978,11 +1978,11 @@ class TestDocumentApi(DirectoriesMixin, DocumentConsumeDelayMixin, APITestCase):
response = self.client.get(f"/api/documents/{doc.pk}/suggestions/")
self.assertEqual(response.status_code, status.HTTP_200_OK)
@mock.patch("documents.parsers.parse_date_generator")
@mock.patch("documents.views.get_date_parser")
@override_settings(NUMBER_OF_SUGGESTED_DATES=0)
def test_get_suggestions_dates_disabled(
self,
parse_date_generator,
mock_get_date_parser: mock.MagicMock,
):
"""
GIVEN:
@@ -1999,7 +1999,8 @@ class TestDocumentApi(DirectoriesMixin, DocumentConsumeDelayMixin, APITestCase):
)
self.client.get(f"/api/documents/{doc.pk}/suggestions/")
self.assertFalse(parse_date_generator.called)
mock_get_date_parser.assert_not_called()
def test_saved_views(self):
u1 = User.objects.create_superuser("user1")

538
src/documents/tests/test_date_parsing.py
View File

@@ -1,538 +0,0 @@
import datetime
from zoneinfo import ZoneInfo
import pytest
from pytest_django.fixtures import SettingsWrapper
from documents.parsers import parse_date
from documents.parsers import parse_date_generator
@pytest.mark.django_db()
class TestDate:
def test_date_format_1(self):
text = "lorem ipsum 130218 lorem ipsum"
assert parse_date("", text) is None
def test_date_format_2(self):
text = "lorem ipsum 2018 lorem ipsum"
assert parse_date("", text) is None
def test_date_format_3(self):
text = "lorem ipsum 20180213 lorem ipsum"
assert parse_date("", text) is None
def test_date_format_4(self, settings_timezone: ZoneInfo):
text = "lorem ipsum 13.02.2018 lorem ipsum"
date = parse_date("", text)
assert date == datetime.datetime(2018, 2, 13, 0, 0, tzinfo=settings_timezone)
def test_date_format_5(self, settings_timezone: ZoneInfo):
text = "lorem ipsum 130218, 2018, 20180213 and lorem 13.02.2018 lorem ipsum"
date = parse_date("", text)
assert date == datetime.datetime(2018, 2, 13, 0, 0, tzinfo=settings_timezone)
def test_date_format_6(self):
text = (
"lorem ipsum\n"
"Wohnort\n"
"3100\n"
"IBAN\n"
"AT87 4534\n"
"1234\n"
"1234 5678\n"
"BIC\n"
"lorem ipsum"
)
assert parse_date("", text) is None
def test_date_format_7(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
settings.DATE_PARSER_LANGUAGES = ["de"]
text = "lorem ipsum\nMärz 2019\nlorem ipsum"
date = parse_date("", text)
assert date == datetime.datetime(2019, 3, 1, 0, 0, tzinfo=settings_timezone)
def test_date_format_8(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
settings.DATE_PARSER_LANGUAGES = ["de"]
text = (
"lorem ipsum\n"
"Wohnort\n"
"3100\n"
"IBAN\n"
"AT87 4534\n"
"1234\n"
"1234 5678\n"
"BIC\n"
"lorem ipsum\n"
"März 2020"
)
assert parse_date("", text) == datetime.datetime(
2020,
3,
1,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_9(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
settings.DATE_PARSER_LANGUAGES = ["de"]
text = "lorem ipsum\n27. Nullmonth 2020\nMärz 2020\nlorem ipsum"
assert parse_date("", text) == datetime.datetime(
2020,
3,
1,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_10(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 22-MAR-2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
22,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_11(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 22 MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
22,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_12(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 22/MAR/2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
22,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_13(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 22.MAR.2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
22,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_14(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 22.MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
22,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_15(self):
text = "Customer Number Currency 22.MAR.22 Credit Card 1934829304"
assert parse_date("", text) is None
def test_date_format_16(self):
text = "Customer Number Currency 22.MAR,22 Credit Card 1934829304"
assert parse_date("", text) is None
def test_date_format_17(self):
text = "Customer Number Currency 22,MAR,2022 Credit Card 1934829304"
assert parse_date("", text) is None
def test_date_format_18(self):
text = "Customer Number Currency 22 MAR,2022 Credit Card 1934829304"
assert parse_date("", text) is None
def test_date_format_19(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 21st MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
21,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_20(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 22nd March 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
22,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_21(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 2nd MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
2,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_22(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 23rd MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
23,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_23(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 24th MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
24,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_24(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 21-MAR-2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
21,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_25(self, settings_timezone: ZoneInfo):
text = "Customer Number Currency 25TH MAR 2022 Credit Card 1934829304"
assert parse_date("", text) == datetime.datetime(
2022,
3,
25,
0,
0,
tzinfo=settings_timezone,
)
def test_date_format_26(self, settings_timezone: ZoneInfo):
text = "CHASE 0 September 25, 2019 JPMorgan Chase Bank, NA. P0 Box 182051"
assert parse_date("", text) == datetime.datetime(
2019,
9,
25,
0,
0,
tzinfo=settings_timezone,
)
def test_crazy_date_past(self):
assert parse_date("", "01-07-0590 00:00:00") is None
def test_crazy_date_future(self):
assert parse_date("", "01-07-2350 00:00:00") is None
def test_crazy_date_with_spaces(self):
assert parse_date("", "20 408000l 2475") is None
def test_utf_month_names(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
settings.DATE_PARSER_LANGUAGES = ["fr", "de", "hr", "cs", "pl", "tr"]
assert parse_date("", "13 décembre 2023") == datetime.datetime(
2023,
12,
13,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "13 août 2022") == datetime.datetime(
2022,
8,
13,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "11 März 2020") == datetime.datetime(
2020,
3,
11,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "17. ožujka 2018.") == datetime.datetime(
2018,
3,
17,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "1. veljače 2016.") == datetime.datetime(
2016,
2,
1,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "15. února 1985") == datetime.datetime(
1985,
2,
15,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "30. září 2011") == datetime.datetime(
2011,
9,
30,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "28. května 1990") == datetime.datetime(
1990,
5,
28,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "1. grudzień 1997") == datetime.datetime(
1997,
12,
1,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "17 Şubat 2024") == datetime.datetime(
2024,
2,
17,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "30 Ağustos 2012") == datetime.datetime(
2012,
8,
30,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "17 Eylül 2000") == datetime.datetime(
2000,
9,
17,
0,
0,
tzinfo=settings_timezone,
)
assert parse_date("", "5. október 1992") == datetime.datetime(
1992,
10,
5,
0,
0,
tzinfo=settings_timezone,
)
def test_multiple_dates(self, settings_timezone: ZoneInfo):
text = """This text has multiple dates.
For example 02.02.2018, 22 July 2022 and December 2021.
But not 24-12-9999 because it's in the future..."""
dates = list(parse_date_generator("", text))
assert dates == [
datetime.datetime(2018, 2, 2, 0, 0, tzinfo=settings_timezone),
datetime.datetime(
2022,
7,
22,
0,
0,
tzinfo=settings_timezone,
),
datetime.datetime(
2021,
12,
1,
0,
0,
tzinfo=settings_timezone,
),
]
def test_filename_date_parse_valid_ymd(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
"""
GIVEN:
- Date parsing from the filename is enabled
- Filename date format is with Year Month Day (YMD)
- Filename contains date matching the format
THEN:
- Should parse the date from the filename
"""
settings.FILENAME_DATE_ORDER = "YMD"
assert parse_date(
"/tmp/Scan-2022-04-01.pdf",
"No date in here",
) == datetime.datetime(2022, 4, 1, 0, 0, tzinfo=settings_timezone)
def test_filename_date_parse_valid_dmy(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
"""
GIVEN:
- Date parsing from the filename is enabled
- Filename date format is with Day Month Year (DMY)
- Filename contains date matching the format
THEN:
- Should parse the date from the filename
"""
settings.FILENAME_DATE_ORDER = "DMY"
assert parse_date(
"/tmp/Scan-10.01.2021.pdf",
"No date in here",
) == datetime.datetime(2021, 1, 10, 0, 0, tzinfo=settings_timezone)
def test_filename_date_parse_invalid(self, settings: SettingsWrapper):
"""
GIVEN:
- Date parsing from the filename is enabled
- Filename includes no date
- File content includes no date
THEN:
- No date is parsed
"""
settings.FILENAME_DATE_ORDER = "YMD"
assert parse_date("/tmp/20 408000l 2475 - test.pdf", "No date in here") is None
def test_filename_date_ignored_use_content(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
"""
GIVEN:
- Date parsing from the filename is enabled
- Filename date format is with Day Month Year (YMD)
- Date order is Day Month Year (DMY, the default)
- Filename contains date matching the format
- Filename date is an ignored date
- File content includes a date
THEN:
- Should parse the date from the content not filename
"""
settings.FILENAME_DATE_ORDER = "YMD"
settings.IGNORE_DATES = (datetime.date(2022, 4, 1),)
assert parse_date(
"/tmp/Scan-2022-04-01.pdf",
"The matching date is 24.03.2022",
) == datetime.datetime(2022, 3, 24, 0, 0, tzinfo=settings_timezone)
def test_ignored_dates_default_order(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
"""
GIVEN:
- Ignore dates have been set
- File content includes ignored dates
- File content includes 1 non-ignored date
THEN:
- Should parse the date non-ignored date from content
"""
settings.IGNORE_DATES = (datetime.date(2019, 11, 3), datetime.date(2020, 1, 17))
text = "lorem ipsum 110319, 20200117 and lorem 13.02.2018 lorem ipsum"
assert parse_date("", text) == datetime.datetime(
2018,
2,
13,
0,
0,
tzinfo=settings_timezone,
)
def test_ignored_dates_order_ymd(
self,
settings: SettingsWrapper,
settings_timezone: ZoneInfo,
):
"""
GIVEN:
- Ignore dates have been set
- Date order is Year Month Date (YMD)
- File content includes ignored dates
- File content includes 1 non-ignored date
THEN:
- Should parse the date non-ignored date from content
"""
settings.FILENAME_DATE_ORDER = "YMD"
settings.IGNORE_DATES = (datetime.date(2019, 11, 3), datetime.date(2020, 1, 17))
text = "lorem ipsum 190311, 20200117 and lorem 13.02.2018 lorem ipsum"
assert parse_date("", text) == datetime.datetime(
2018,
2,
13,
0,
0,
tzinfo=settings_timezone,
)

5
src/documents/views.py
View File

@@ -148,7 +148,6 @@ from documents.models import Workflow
from documents.models import WorkflowAction
from documents.models import WorkflowTrigger
from documents.parsers import get_parser_class_for_mime_type
from documents.parsers import parse_date_generator
from documents.permissions import AcknowledgeTasksPermissions
from documents.permissions import PaperlessAdminPermissions
from documents.permissions import PaperlessNotePermissions
@@ -158,6 +157,7 @@ from documents.permissions import get_document_count_filter_for_user
from documents.permissions import get_objects_for_user_owner_aware
from documents.permissions import has_perms_owner_aware
from documents.permissions import set_permissions_for_object
from documents.plugins.date_parsing import get_date_parser
from documents.schema import generate_object_with_permissions_schema
from documents.serialisers import AcknowledgeTasksViewSerializer
from documents.serialisers import BulkDownloadSerializer
@@ -1023,7 +1023,8 @@ class DocumentViewSet(
dates = []
if settings.NUMBER_OF_SUGGESTED_DATES > 0:
gen = parse_date_generator(doc.filename, doc.content)
date_parser = get_date_parser()
gen = date_parser.parse(doc.filename, doc.content)
dates = sorted(
{
i