def uploaded_new_derived_document(file):
"""
Like `uploaded_new_derived_document' this creates the 'DerivedBlob'
object and a 'DerivedDocument' object.
The `DerivedDocument` object is returned, you need to fill in
`derived_from` and `index`.
"""
file_type, md5_sum = identify_and_md5(file)
try:
blob = DerivedBlob.objects.get(md5_sum=md5_sum)
except DerivedBlob.DoesNotExist:
blob = DerivedBlob(md5_sum=md5_sum, file=file, file_type=file_type)
blob.upload_to_url = get_path(file_type)
blob.save()
document = DerivedDocument()
document._blob = blob
return document
def test_auto_derived_blob_creation(self):
"""
Test that a blob is created when using the file attribue on a derived
document.
"""
doc = Document(title='New Doc', file_name='A File', file=self.upload,
author=self.u1)
doc.save()
derived = DerivedDocument(derived_from=doc._blob, index=0,
file=self.upload2)
derived.save()
self.upload2.seek(0)
self.upload.seek(0)
self.assertEqual(derived.file.read(), self.upload2.read())
self.assertEqual(DerivedBlob.objects.count(), 1)
self.upload2.seek(0)
derived.file.seek(0)
self.assertEqual(DerivedBlob.objects.all()[0].file.read(),
self.upload2.read())
def create_pdf(doc):
log.info('PDF Conversion start')
blob = doc._blob
orig_file = blob.file
log.info('Starting conversion of %s to PDF' % blob)
# Check for derived files of PDF type
if blob.documents.filter(file_type='pdf'):
log.info('%s is a PDF, no need to convert' % blob)
return False
elif blob.derived_documents.filter(file_type='pdf'):
log.info('%s has derived PDF, no need to convert' % blob)
return False
tempd_loc = tempfile.mkdtemp()
temp_file, tempf_loc = tempfile.mkstemp(dir=tempd_loc)
temp_file = os.fdopen(temp_file, "wb")
for chunk in orig_file.chunks():
temp_file.write(chunk)
temp_file.close()
try:
port = get_free_port()
log.info("Launching unoconv with port %s" % port)
proc = subprocess.Popen(UNOCONV_CALL % (port, tempd_loc, tempf_loc),
#stderr=subprocess.PIPE,
#stdout=subprocess.PIPE,
shell=True)
# Create and start a watchdog thread
t = threading.Timer(TIMEOUT, timeout, [proc, port])
t.start()
stderr, stdout = None, None#proc.communicate()
proc.wait()
t.cancel()
if proc.returncode != 0:
error = subprocess.CalledProcessError(proc.returncode,
UNOCONV_CALL % (
port, tempd_loc, tempf_loc))
error.output = "%s %s" % (stderr, stdout)
raise error
log.info("unoconv (port %s) output: %s %s" % (port, stderr, stdout))
except subprocess.CalledProcessError as e:
log.error('unoconv (port %s) returned a non-zero exit status: %s' % (port, e.output))
os.unlink(tempf_loc)
files = os.listdir(tempd_loc)
for pdf in files:
if pdf.lower().endswith('.pdf'):
break
else:
shutil.rmtree(tempd_loc, True)
raise ConversionError('Unable to find PDF file')
pdf = os.path.abspath(os.path.join(tempd_loc, pdf))
pdf = open(pdf, 'rb')
filename = os.path.basename(orig_file.name)
filename = os.path.splitext(filename)[0] + '.pdf'
doc = DerivedDocument(derived_from=blob)
doc.file = UploadedFile(pdf, filename, 'application/pdf', 0, None)
doc.index = type_to_priorty('pdf')
# # Do one last check before saving the blob, just incase this task got fired
# # twice in quick succession.
if blob.derived_documents.filter(file_type='pdf'):
return False
else:
doc.save()
pdf.close()
shutil.rmtree(tempd_loc, True)
log.info("Convertion of %s complete" % blob)
return True
def create_pngs(doc, type='pngs'):
blob = doc._blob
log.info('Starting png generation of: %s', doc)
# Check to make sure that we don't already have a pngs pack
# Check for derived files of PDF type
if doc.file_type == 'png':
log.info('%s is a PNG, no need to convert' % blob)
return False
elif blob.derived_documents.filter(file_type='png'):
log.info('%s has derived PNG, no need to convert' % blob)
return False
# Locate a pdf file
if doc.type == 'pdf':
pdf = doc
else:
pdf = doc.get_derived_documents_of_type('pdf')
if pdf:
pdf = pdf[0]
if not pdf:
log.info("No PDF avaliable for %s" % blob)
return
# Create a temp folder
temp_folder = tempfile.mkdtemp()
log.debug('working with: %s', temp_folder)
file = tempfile.NamedTemporaryFile(dir=temp_folder, delete=False)
for data in pdf.file.chunks():
file.write(data)
file.close()
# Now call ghostscript
return_code = subprocess.call(["gs", "-sDEVICE=png16m",
"-sOutputFile=%s/slide-%s.png" % (temp_folder, '%03d'),
"-r600", "-dNOPAUSE", "-dBATCH", "-dMaxBitmap=1000000000",
#"-dFirstPage=1", "-dLastPage=1",
"%s" % file.name])
if return_code != 0:
log.error('Ghostscript error')
# Clean up
shutil.rmtree(temp_folder)
create_pngs.retry()
# Process the generated files with PIL
# First generate a list of file in the tempdir
compiled_regex = re.compile('^slide-(\w+).png$')
scaled_images = {}
for file in os.listdir(temp_folder):
# Check using regex
match = re.match(compiled_regex, file)
if match:
log.debug('scaling image: %s', file)
order = int(match.group(1))
# Resize using PIL
slide = Image.open(os.path.join(temp_folder, file))
slide.thumbnail((1920, 1200), Image.ANTIALIAS)
new_filename = os.path.join(temp_folder, 'slide-scaled-%03d.png' % order)
slide.save(new_filename)
scaled_images[order] = new_filename
# Make sure that the order starts at 0 and has no gaps
new_images = {}
order = 0
sorted_keys = scaled_images.keys()
sorted_keys.sort()
for item in [scaled_images[key] for key in sorted_keys]:
new_images[order] = item
order += 1
scaled_images = new_images
# Before uploading check that there are still no other pngs up there.
if blob.derived_documents.filter(file_type='png'):
log.info('%s has derived PNG now, canceling upload' % blob)
return False
# Now go through all the generated slides and upload
# Create a new derivedfile pack
try:
for order, filename in scaled_images.iteritems():
file = open(filename, 'rb')
parts = os.path.split(filename)
filename = os.path.join(parts[-2], '%s_%s' % (
doc.file_name[0:60], parts[-1]))
upfile = TemporaryUploadedFile(filename, 'image/png', 0, None)
upfile.file = file
derived_doc = DerivedDocument(derived_from=blob)
derived_doc.file = upfile
derived_doc.index = order
derived_doc.save()
except:
log.error(filename)
raise
shutil.rmtree(temp_folder)
return True
