matthias/mayan-edms
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Finish the document parsing app.

Browse Source 
Signed-off-by: Roberto Rosario <roberto.rosario.gonzalez@gmail.com>
This commit is contained in:
Roberto Rosario
2017-08-24 03:54:07 -04:00
parent e9591c92f9
commit a7eaf6b368
25 changed files with 423 additions and 639 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
mayan/apps/document_parsing/tests/test_parsers.py
View File

@@ -5,12 +5,9 @@ from django.test import override_settings
from common.tests import BaseTestCase
from documents.models import DocumentType
from documents.tests import (
TEST_DOCUMENT_PATH, TEST_DOCUMENT_TYPE_LABEL, TEST_HYBRID_DOCUMENT_PATH
)
from documents.tests import TEST_DOCUMENT_PATH, TEST_DOCUMENT_TYPE_LABEL
from ..classes import TextExtractor
from ..parsers import PDFMinerParser, PopplerParser
from ..parsers import PopplerParser
@override_settings(OCR_AUTO_OCR=False)
@@ -30,54 +27,11 @@ class ParserTestCase(BaseTestCase):
self.document_type.delete()
super(ParserTestCase, self).tearDown()
def test_pdfminer_parser(self):
parser = PDFMinerParser()
parser.process_document_version(self.document.latest_version)
self.assertTrue(
'Mayan EDMS Documentation' in self.document.pages.first().ocr_content.content
)
def test_poppler_parser(self):
parser = PopplerParser()
parser.process_document_version(self.document.latest_version)
self.assertTrue(
'Mayan EDMS Documentation' in self.document.pages.first().ocr_content.content
)
@override_settings(OCR_AUTO_OCR=False)
class TextExtractorTestCase(BaseTestCase):
def setUp(self):
super(TextExtractorTestCase, self).setUp()
self.document_type = DocumentType.objects.create(
label=TEST_DOCUMENT_TYPE_LABEL
)
with open(TEST_HYBRID_DOCUMENT_PATH) as file_object:
self.document = self.document_type.new_document(
file_object=File(file_object)
)
def tearDown(self):
self.document_type.delete()
super(TextExtractorTestCase, self).tearDown()
def test_text_extractor(self):
TextExtractor.process_document_version(
document_version=self.document.latest_version
)
self.assertEqual(
self.document.latest_version.pages.first().ocr_content.content,
'Sample text',
)
self.assertEqual(
self.document.latest_version.pages.last().ocr_content.content,
'Sample text in image form',
'Mayan EDMS Documentation' in self.document.pages.first().content.content
)