paradizelost/paperless-ngx
1
0
Fork 0
mirror of https://github.com/paperless-ngx/paperless-ngx.git synced 2025-07-28 18:24:38 -05:00
Code Issues Packages Projects Releases Wiki Activity

Feature: Upgrade Gotenberg to v8 (#7094)

Browse Source
This commit is contained in:
Trenton H
2024-06-26 19:37:50 -07:00
committed by GitHub
parent 80c2d90e74
commit 29e6371cd1
8 changed files with 45 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
src/paperless_mail/parsers.py
View File

@@ -9,7 +9,9 @@ from django.conf import settings
from django.utils.timezone import is_naive
from django.utils.timezone import make_aware
from gotenberg_client import GotenbergClient
from gotenberg_client.options import Margin
from gotenberg_client.options import MarginType
from gotenberg_client.options import MarginUnitType
from gotenberg_client.options import PageMarginsType
from gotenberg_client.options import PageSize
from gotenberg_client.options import PdfAFormat
from humanize import naturalsize
@@ -20,6 +22,7 @@ from tika_client import TikaClient
from documents.parsers import DocumentParser
from documents.parsers import ParseError
from documents.parsers import make_thumbnail_from_pdf
from paperless.models import OutputTypeChoices
class MailDocumentParser(DocumentParser):
@@ -30,17 +33,22 @@ class MailDocumentParser(DocumentParser):
logging_name = "paperless.parsing.mail"
@staticmethod
def _settings_to_gotenberg_pdfa() -> Optional[PdfAFormat]:
def _settings_to_gotenberg_pdfa(self) -> Optional[PdfAFormat]:
"""
Converts our requested PDF/A output into the Gotenberg API
format
"""
if settings.OCR_OUTPUT_TYPE in {"pdfa", "pdfa-2"}:
if settings.OCR_OUTPUT_TYPE in {
OutputTypeChoices.PDF_A,
OutputTypeChoices.PDF_A2,
}:
return PdfAFormat.A2b
elif settings.OCR_OUTPUT_TYPE == "pdfa-1": # pragma: no cover
return PdfAFormat.A1a
elif settings.OCR_OUTPUT_TYPE == "pdfa-3": # pragma: no cover
elif settings.OCR_OUTPUT_TYPE == OutputTypeChoices.PDF_A1: # pragma: no cover
self.log.warn(
"Gotenberg does not support PDF/A-1a, choosing PDF/A-2b instead",
)
return PdfAFormat.A2b
elif settings.OCR_OUTPUT_TYPE == OutputTypeChoices.PDF_A3: # pragma: no cover
return PdfAFormat.A3b
return None
@@ -329,7 +337,14 @@ class MailDocumentParser(DocumentParser):
response = (
route.index(email_html_file)
.resource(css_file)
.margins(Margin(top=0.1, bottom=0.1, left=0.1, right=0.1))
.margins(
PageMarginsType(
top=MarginType(0.1, MarginUnitType.Inches),
bottom=MarginType(0.1, MarginUnitType.Inches),
left=MarginType(0.1, MarginUnitType.Inches),
right=MarginType(0.1, MarginUnitType.Inches),
),
)
.size(PageSize(height=11.7, width=8.27))
.scale(1.0)
.run()
@@ -404,7 +419,14 @@ class MailDocumentParser(DocumentParser):
route.index(html_clean_file)
# Set page size, margins
route.margins(Margin(top=0.1, bottom=0.1, left=0.1, right=0.1)).size(
route.margins(
PageMarginsType(
top=MarginType(0.1, MarginUnitType.Inches),
bottom=MarginType(0.1, MarginUnitType.Inches),
left=MarginType(0.1, MarginUnitType.Inches),
right=MarginType(0.1, MarginUnitType.Inches),
),
).size(
PageSize(height=11.7, width=8.27),
).scale(1.0)

5
src/paperless_tika/parsers.py
View File

@@ -102,7 +102,10 @@ class TikaDocumentParser(DocumentParser):
}:
route.pdf_format(PdfAFormat.A2b)
elif settings.OCR_OUTPUT_TYPE == OutputTypeChoices.PDF_A1:
route.pdf_format(PdfAFormat.A1a)
self.log.warn(
"Gotenberg does not support PDF/A-1a, choosing PDF/A-2b instead",
)
route.pdf_format(PdfAFormat.A2b)
elif settings.OCR_OUTPUT_TYPE == OutputTypeChoices.PDF_A3:
route.pdf_format(PdfAFormat.A3b)

5
src/paperless_tika/tests/test_tika_parser.py
View File

@@ -109,7 +109,7 @@ class TestTikaParser(HttpxMockMixin, TestCase):
for setting, expected_key in [
("pdfa", "PDF/A-2b"),
("pdfa-2", "PDF/A-2b"),
("pdfa-1", "PDF/A-1a"),
("pdfa-1", "PDF/A-2b"),
("pdfa-3", "PDF/A-3b"),
]:
with override_settings(OCR_OUTPUT_TYPE=setting):
@@ -124,9 +124,10 @@ class TestTikaParser(HttpxMockMixin, TestCase):
request = self.httpx_mock.get_request()
found = False
for field in request.stream.fields:
if isinstance(field, DataField) and field.name == "pdfFormat":
if isinstance(field, DataField) and field.name == "pdfa":
self.assertEqual(field.value, expected_key)
found = True
break
self.assertTrue(found)
self.httpx_mock.reset(assert_all_responses_were_requested=False)