def test_string_content():
"""FileBackup PDFs should be parsable."""
pdf_path = Path(__file__).parent / 'data' / 'matmod_exam_des_2017.pdf'
pdf_content = ContentFile(pdf_path.read_bytes())
sha1 = '0000000000000000000000000000000000000000'
pdf_backup = Pdf(sha1_hash=sha1)
pdf_backup.file.save(name=sha1 + '.pdf', content=pdf_content)
pdf_backup.read_text()
pdf_backup.save()
pdf_backup.refresh_from_db()
# Ensure unicode string
assert isinstance(pdf_backup.text, str)
# Check content with text property
assert len(pdf_backup.text.split('\f')) == 6
assert 'Rottman' in pdf_backup.text
assert 'population model' in pdf_backup.text
assert 'this is not in the exam' not in pdf_backup.text
# Check associated PdfPage model objects
pages = pdf_backup.pages.all()
assert pages.count() == 6
# The ordering should be based on page number
for page_num, page in enumerate(pages):
assert page.number == page_num
assert page.confidence is None
# And each page should containt content
assert 'Rottman' in pages[0].text
assert 'Rottman' not in pages[2].text
assert 'population model' in pages[2].text
assert 'population model' not in pages[0].text
def test_classify_pdf(self):
"""Exam type should be determinable from pdf content."""
# The PDF contains the following content
sha1_hash = '0000000000000000000000000000000000000000'
pdf = Pdf(sha1_hash=sha1_hash)
text = """
NTNU TMA4115 Matematikk 3
Institutt for matematiske f*g
eksamen 11.08.05
Eksamenssettet har 12 punkter.
"""
content = ContentFile(text)
pdf.file.save(content=content, name=sha1_hash + '.pdf', save=True)
# No errors should be raised when no pages has been saved yet, but
# False should be returned to indicate a lack of success.
assert pdf.classify(
allow_ocr=True) is False # Malformed plain text PDF
assert pdf.classify(allow_ocr=False) is False
assert pdf.exams.count() == 0
# But now we add a cover page and classify its content
PdfPage.objects.create(text=text, pdf=pdf, number=0)
pdf.refresh_from_db()
assert pdf.classify() is True
# And all metadata should be saved
pdf = Pdf.objects.get(id=pdf.id)
assert pdf.exams.count() == 1
exam = pdf.exams.first()
assert exam.language == Language.BOKMAL
assert exam.course_code == 'TMA4115'
assert exam.solutions is False
assert exam.year == 2005
assert exam.season == Season.CONTINUATION
assert exam.content_type == DocumentInfo.EXAM
# When the classification method changes it result, old results
# should be removed. This is simulated here by mutating the exam.
exam.year == 1999
exam.save()
pdf.classify()
pdf = Pdf.objects.get(id=pdf.id)
assert pdf.exams.count() == 1
assert pdf.exams.first().year == 2005
# But verified exams should NOT be deleted
verified_exam = DocumentInfo.objects.create(
year=1999,
course_code=exam.course_code,
language=exam.language,
solutions=exam.solutions,
)
user = UserFactory.create(username='******')
verified_exam_pdf = DocumentInfoSource.objects.create(
document_info=verified_exam,
pdf=pdf,
)
verified_exam_pdf.verified_by.add(user)
pdf.classify()
pdf = Pdf.objects.get(id=pdf.id)
assert pdf.exams.count() == 2
# If a PDF has already verified document info, classify should by a
# no-op.
DocumentInfoSource.objects.all().delete()
verified_exam_pdf = DocumentInfoSource.objects.create(
document_info=verified_exam,
pdf=pdf,
)
verified_exam_pdf.verified_by.add(user)
pdf.classify(save=True)
assert pdf.exams.count() == 1
assert pdf.exams.first() == verified_exam