def overlay_template_pdf():
"""
The api app calls this with a PDF as the POST body, expecting to receive a PDF back with the red overlay applied.
This endpoint will raise an error if you try and include a page number because it assumes you meant to ask for a png
in that case.
"""
encoded_string = request.get_data()
if not encoded_string:
raise InvalidRequest('no data received in POST')
if request.args:
raise InvalidRequest(
f'Did not expect any args but received {request.args}. Did you mean to call overlay.png?'
)
pdf = PdfFileReader(BytesIO(encoded_string))
for i in range(pdf.numPages):
_colour_no_print_areas_of_page_in_red(pdf.getPage(i),
is_first_page=(i == 0))
return send_file(filename_or_fp=bytesio_from_pdf(pdf),
mimetype='application/pdf')
def overlay_template_png_for_page():
"""
The admin app calls this multiple times to get pngs of each separate page to show on the front end.
This endpoint expects a "page_number" param that _must_ be included. It also includes as the HTTP POST body the
binary data of that individual page of the PDF.
"""
encoded_string = request.get_data()
if not encoded_string:
raise InvalidRequest('no data received in POST')
file_data = BytesIO(encoded_string)
if 'is_first_page' in request.args:
is_first_page = request.args.get('is_first_page', '').lower() == 'true'
elif 'page_number' in request.args:
page = int(request.args.get('page_number'))
is_first_page = page == 1 # page_number arg is one-indexed
else:
raise InvalidRequest(
f'page_number or is_first_page must be specified in request params {request.args}'
)
return send_file(
filename_or_fp=png_from_pdf(
_colour_no_print_areas_of_single_page_pdf_in_red(
file_data, is_first_page=is_first_page),
# the pdf is only one page, so this is always 1.
page_number=1),
mimetype='image/png',
)
def _colour_no_print_areas_of_single_page_pdf_in_red(src_pdf, is_first_page):
"""
Overlays the non-printable areas onto the src PDF, this is so users know which parts of they letter fail validation.
This function expects that src_pdf only represents a single page. It adds red areas (if `is_first_page` is set, then
it'll add red areas around the address window too) and returns a single page pdf.
:param BytesIO src_pdf: A file-like representing a single page pdf
:param bool is_first_page: true if we should overlay the address block red area too.
"""
pdf = PdfFileReader(src_pdf)
if pdf.numPages != 1:
# this function is used to render images, which call template-preview separately for each page. This function
# should be colouring a single page pdf (which might be any individual page of an original precompiled letter)
raise InvalidRequest(
'_colour_no_print_areas_of_page_in_red should only be called for a one-page-pdf'
)
page = pdf.getPage(0)
_colour_no_print_areas_of_page_in_red(page, is_first_page)
out = bytesio_from_pdf(pdf)
# it's a good habit to put things back exactly the way we found them
src_pdf.seek(0)
return out
def sanitise_precompiled_letter():
encoded_string = request.get_data()
allow_international_letters = (
request.args.get('allow_international_letters') == 'true')
if not encoded_string:
raise InvalidRequest('no-encoded-string')
sanitise_json = sanitise_file_contents(
encoded_string,
allow_international_letters=allow_international_letters,
)
status_code = 400 if sanitise_json.get('message') else 200
return jsonify(sanitise_json), status_code
def _does_pdf_contain_colorspace(colourspace, data):
doc = fitz.open(stream=data, filetype="pdf")
for i in range(len(doc)):
try:
page = doc.getPageImageList(i)
except RuntimeError:
current_app.logger.warning("Fitz couldn't read page info for page {}".format(i + 1))
raise InvalidRequest("Invalid PDF on page {}".format(i + 1))
for img in page:
xref = img[0]
pix = fitz.Pixmap(doc, xref)
if colourspace in pix.colorspace.__str__():
data.seek(0)
return True
data.seek(0)
return False