django_pdf_utils.py

"""
A bunch of PDF utilities for Django, wkhtmltopdf and pdftk

Note: These have been cobbled together from various projects,
      and probably can't be trusted. Make sure you read over
      and full understand the code fully before attempting to use!
"""

import sys, os, csv
import subprocess
import datetime
from tempfile import NamedTemporaryFile
from pipes import quote

from django.conf import settings
from django.template import RequestContext
from django.template.loader import render_to_string
from django.template import loader
from django.contrib.sites.models import Site
from django.core.files.base import ContentFile
from django.http import HttpResponse



def get_file_location_from_template_name(template_name):
    template_dirs = getattr(settings, 'TEMPLATE_DIRS')
    for template_dir in template_dirs:
        location = os.path.join(template_dir, template_name)
        if os.path.exists(location):
            return location

    return None


def render_pdf(file_contents, file_name='output.pdf'):
    response = HttpResponse(mimetype='application/pdf')

    if not file_name.endswith('.pdf'):
        file_name = file_name + '.pdf'

    response['Content-Disposition'] = 'filename=%s' % file_name

    response.write(file_contents)
    
    return response


def render_png(file_contents, file_name='output.png'):
    response = HttpResponse(mimetype='image/png')

    if not file_name.endswith('.png'):
        file_name = file_name + '.png'

    response['Content-Disposition'] = 'filename=%s' % file_name

    response.write(file_contents)

    return response


def process_value_for_fdf(val):
    """
    Due to a weird bug in PDFTK 1.44, we can't handle parens
    in pdf forms. For the moment, we just replace them with
    square brackets.
    """
    if not val:
        return val

    val = val.replace("(", "[")
    val = val.replace(")", "]")

    return val


def create_pdf_from_pdf(template_name, context, flatten=True):
    """
    Given a PDF template and context returns a PDF file to the user.
    The PDF should contain a form, and the context should refer to fields in that form.

    Prerequisites: fdfgen (available on PyPI) and pdftk (http://www.pdflabs.com/docs/install-pdftk/)
    """
    from fdfgen import forge_fdf

    template_location = get_file_location_from_template_name(template_name)

    if not template_location:
        raise Exception("PDF Template %s does not exist!" % template_name)

    fields = [(key, process_value_for_fdf(val)) for key, val in context.iteritems()]
    fdf = forge_fdf("", fields, [], [], [])

    fdf_file = NamedTemporaryFile(delete=False, suffix='.fdf')
    fdf_file.write(fdf)
    fdf_file.close()

    command = 'pdftk %s fill_form %s output -' % (template_location, fdf_file.name)
    if flatten:
        command = command + ' flatten'

    popen = subprocess.Popen(command, bufsize=4096,
                             stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                             shell=True)
    pdf_contents = popen.stdout.read()
    popen.wait()

    os.unlink(fdf_file.name)

    return pdf_contents


def create_pdf_from_html(template_name, context, toc=False, include_footer=False, \
                           margin_bottom=0, margin_top=0, margin_left=0, \
                           margin_right=0, javascript_delay=0, zoom=1.0, page_size='A4'):
    """
    Given a template, context and filename, returns a PDF file to the user.
    Prerequisites: wkhtmltopdf (http://wkhtmltopdf.org/)
    """

    context['base_url'] = 'file:///' + settings.STATIC_ROOT + '/'
    rendered = render_to_string(template_name, context)

    rendered = rendered.replace('src="' + settings.MEDIA_URL, 'src="file:///' + settings.MEDIA_ROOT)

    rendered = rendered.encode('ascii', 'xmlcharrefreplace')

    html_file = NamedTemporaryFile(delete=False, suffix='.html')
    html_file.write(rendered)
    html_file.close()

    margin_args = ''
    if type(margin_bottom) in (int, float):
        margin_args += ' --margin-bottom %s' % margin_bottom

    if type(margin_top) in (int, float):
        margin_args += ' --margin-top %s' % margin_top

    if type(margin_left) in (int, float):
        margin_args += ' --margin-left %s' % margin_left

    if type(margin_right) in (int, float):
        margin_args += ' --margin-right %s' % margin_right

    zoom_arg = ''
    if type(zoom) in (int, float):
        zoom_arg = ' --zoom %s' % zoom

    page_size_arg = ''
    if page_size:
        page_size_arg = ' --page-size %s' % page_size

    if toc:
        extra_args = 'toc'
    else:
        extra_args = ''

    if include_footer:
        left_footer = "Page [page] of [toPage]"
        left_footer_arg = quote(left_footer)

        right_footer = "PDF Generated by on %s" % (
            datetime.datetime.today().strftime("%b %d %Y"),
        )

        # Remove square brackets so wkhtml doesn't get confuzzled
        right_footer = right_footer.replace('[', '')
        right_footer = right_footer.replace(']', '')

        # Escape and surround the right footer with quotes
        right_footer_arg = quote(right_footer)

        footer_args = '--footer-left %s --footer-right %s' % (
            left_footer_arg,
            right_footer_arg,
        )

    else:
        footer_args = ''

    command_args = 'wkhtmltopdf %s --javascript-delay %s %s %s %s %s %s -' % (
        margin_args,
        javascript_delay,
        page_size_arg,
        zoom_arg,
        footer_args,
        extra_args,
        html_file.name
    )

    popen = subprocess.Popen(command_args, bufsize=4096,
                             stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                             shell=True)
    pdf_contents = popen.stdout.read()
    popen.wait()

    os.unlink(html_file.name)

    return pdf_contents


def create_png_from_html(template_name, context, width=None, height=None, crop_h=None, crop_w=None, crop_x=None, crop_y=None, javascript_delay=0):
    """
    Given a template, context and filename, returns a PNG file
    Prerequisites: wkhtmltopdf (http://wkhtmltopdf.org/)
    """

    context['base_url'] = 'file:///' + settings.STATIC_ROOT + '/'
    rendered = render_to_string(template_name, context)

    rendered = rendered.replace('src="' + settings.MEDIA_URL, 'src="file:///' + settings.MEDIA_ROOT)

    rendered = rendered.encode('ascii', 'xmlcharrefreplace')

    html_file = NamedTemporaryFile(delete=False, suffix='.html')
    html_file.write(rendered)
    html_file.close()

    crop_args = ''
    if type(crop_h) in (int, float):
        crop_args += ' --crop-h %s' % crop_h

    if type(crop_w) in (int, float):
        crop_args += ' --crop-w %s' % crop_w

    if type(crop_x) in (int, float):
        crop_args += ' --crop-x %s' % crop_x

    if type(crop_y) in (int, float):
        crop_args += ' --crop-y %s' % crop_y

    width_arg = ''
    if type(width) in (int, float):
        width_arg = ' --width %s' % width

    height_arg = ''
    if type(height) in (int, float):
        height_arg = ' --height %s' % height

    command = 'wkhtmltoimage --format png --javascript-delay %s %s %s %s %s -' % (
        javascript_delay,
        crop_args,
        width_arg,
        height_arg,
        html_file.name
    )

    popen = subprocess.Popen(command, bufsize=4096,
                             stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                             shell=True)
    image_contents = popen.stdout.read()
    popen.wait()

    os.unlink(html_file.name)

    return image_contents


def overlay_pdfs(overlay, background, overlay_is_template=False, background_is_template=False):
  """
  Overlays one pdf on top of another.
  Prerequisites: pdftk (http://www.pdflabs.com/docs/install-pdftk/)
  """
    if overlay_is_template:
        overlay_file_name = get_file_location_from_template_name(overlay)

        if not overlay_file_name:
            raise Exception("PDF Template %s does not exist!" % overlay)
    else:
        overlay_file = NamedTemporaryFile(delete=False, suffix='.pdf')
        overlay_file.write(overlay)
        overlay_file.close()
        overlay_file_name = overlay_file.name

    if background_is_template:
        background_file_name = get_file_location_from_template_name(background)

        if not background_file_name:
            raise Exception("PDF Template %s does not exist!" % background)
    else:
        background_file = NamedTemporaryFile(delete=False, suffix='.pdf')
        background_file.write(background)
        background_file.close()
        background_file_name = background_file.name


    command = "pdftk %s multibackground %s output -" % (overlay_file_name, background_file_name)

    popen = subprocess.Popen(command, bufsize=4096,
                             stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                             shell=True)

    pdf_contents = popen.stdout.read()
    popen.wait()

    if not overlay_is_template:
        os.unlink(overlay_file.name)

    if not background_is_template:
        os.unlink(background_file.name)

    return pdf_contents


def join_pdfs(*pdfs):
    """
    Joins multiple pdf files
    Prerequisites: pdftk (http://www.pdflabs.com/docs/install-pdftk/)
    """
    file_names = []
    temporary_files = []

    for pdf in pdfs:
        if hasattr(pdf, 'name'):
            pdf_path = settings.MEDIA_ROOT + '/' + pdf.name
            file_names.append(pdf_path)
        elif pdf.endswith('.pdf'):
            file_names.append(get_file_location_from_template_name(pdf))
        else:
            pdf_file = NamedTemporaryFile(delete=False, suffix='.pdf')
            pdf_file.write(pdf)
            pdf_file.close()
            temporary_files.append(pdf_file)
            file_names.append(pdf_file.name)

    command = "pdftk %s cat output -" % ' '.join(file_names)

    popen = subprocess.Popen(command, bufsize=4096,
                             stdout=subprocess.PIPE, stderr=subprocess.PIPE,
                             shell=True)

    pdf_contents = popen.stdout.read()
    popen.wait()

    for temporary_file in temporary_files:
        os.unlink(temporary_file.name)

    return pdf_contents

Breaking change with wkhtmltopdf 0.12.6

If images from the filesystem are embedded then the updated version of wkhtmltopdf will cause an error unless ether the
--enable-local-file-access flag is passed in or --allow {path} is used.