def test_textTrimming (self):
sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '07_trim.docx')
docxProcessor = DocxProcessor(sampleDocxFile)
document = docxProcessor.document()
self.assertEqual('Paragraph with trailing and closing spaces', document.content()[0])
self.assertEqual('Paragraph with trailing and closing spaces', document.formatting()[0].text),
def test_textWithLineBlocks (self): sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '08_line_blocks.docx') docxProcessor = DocxProcessor(sampleDocxFile) document = docxProcessor.document() expectedDocument = Document().initWithFile(os.path.join(os.getcwd(), 'samples', 'expected outcome', 'docx', 'test_08')) self.assertEquals(expectedDocument.content(), document.content()) self.assertEquals(expectedDocument.formatting(), document.formatting())
def test_textWithLineBreak (self):
sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '06_new_line_same_paragraph.docx')
docxProcessor = DocxProcessor(sampleDocxFile)
document = docxProcessor.document()
self.assertEqual('Title', document.content()[0])
self.assertEqual('Regular paragraph.\rNew line on the same paragraph.', document.content()[1])
self.assertEqual('New paragraph.', document.content()[2])
def test_textWithWeirdFormatting (self): sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '11_weird_formatting.docx') docxProcessor = DocxProcessor(sampleDocxFile) document = docxProcessor.document() expectedDocument = Document().initWithFile(os.path.join(os.getcwd(), 'samples', 'expected outcome', 'docx', 'test_11')) self.assertEquals(expectedDocument.content(), document.content()) self.assertEquals(expectedDocument.formatting(), document.formatting())
def test_paragraphStyleMapping (self):
sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '02_paragraph_styles.docx')
docxProcessor = DocxProcessor(sampleDocxFile)
styleMapping = docxProcessor.paragraphStyleMapping()
self.assertEqual(len(styleMapping), 9)
self.assertIn('style0', styleMapping.keys())
self.assertEqual('Default Style', styleMapping['style0'])
self.assertIn('style1', styleMapping.keys())
self.assertEqual('Heading 1', styleMapping['style1'])
def test_paragraphStyleReading (self):
sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '02_paragraph_styles.docx')
docxProcessor = DocxProcessor(sampleDocxFile)
document = docxProcessor._word_document
paragraph = [element for element in document.iter() if element.tag == fullyQualifiedName('w', 'p')][0]
pStyle = paragraph.find('./' + fullyQualifiedName('w','pPr') + '/' + fullyQualifiedName('w', 'pStyle'))
style = pStyle.get(fullyQualifiedName('w', 'val'))
self.assertEqual(style, 'style1')
def test_formattingFragmentSorting (self):
sampleDocxFile = os.path.join(os.getcwd(), 'samples', 'docx', '03_text_formatting.docx')
docxProcessor = DocxProcessor(sampleDocxFile)
document = docxProcessor.document()
expectedFormatting = [
self.fragment(1, None, "I can do that\n", {'style': 'Heading 1'}),
self.fragment(3, None, "I gotta piss\n", {'style': 'Heading 1'}),
self.fragment(4, (133, 158), "Do you believe that shit?", {'formattings': ['underline']}),
self.fragment(5, None, "We happy?\n", {'style': 'Heading 1'}),
self.fragment(6, (118, 269), "Do you see a little Asian child with a blank expression on his face sitting outside on a mechanical helicopter that shakes when you put quarters in it?", {'formattings': ['bold']}),
self.fragment(7, None, "Uuummmm, this is a tasty burger!\n", {'style': 'Heading 1'}),
self.fragment(8, (335, 347), "m**********r", {'formattings': ['italic']}),
self.fragment(8, (443, 455), "m**********r", {'formattings': ['italic']}),
self.fragment(9, None, "I'm serious as a heart attack\n", {'style': 'Heading 1'}),
self.fragment(10, (182, 194), "M**********r", {'formattings': ['italic']})
]
print("DOCUMENT FORMATTING:\n" + str(document.formatting()))
self.assertIsInstance(document.formatting()[0], FragmentIdentifier)
self.assertEqual(expectedFormatting, document.formatting())
def processDocx (docxFile, targetFolder): docxProcessor = DocxProcessor(docxFile) document = docxProcessor.document() document.writeTo(targetFolder)