From: Marcin Koziej Date: Mon, 5 Dec 2011 12:58:40 +0000 (+0100) Subject: customized pdf generation. X-Git-Url: https://git.mdrn.pl/wolnelektury.git/commitdiff_plain/5a605e04abaa18a80e391c5cc26c48e0125d6f3c?ds=inline customized pdf generation. --- diff --git a/apps/catalogue/models.py b/apps/catalogue/models.py index 2e20717fc..b82879ba5 100644 --- a/apps/catalogue/models.py +++ b/apps/catalogue/models.py @@ -8,6 +8,7 @@ from django.db import models from django.db.models import permalink, Q import django.dispatch from django.core.cache import cache +from django.core.files.storage import DefaultStorage from django.utils.translation import ugettext_lazy as _ from django.contrib.auth.models import User from django.template.loader import render_to_string @@ -24,7 +25,8 @@ from newtagging import managers from catalogue.fields import JSONField, OverwritingFileField from catalogue.utils import create_zip from shutil import copy - +from glob import glob +import re from os import path @@ -175,7 +177,7 @@ class Tag(TagBase): def get_dynamic_path(media, filename, ext=None, maxlen=100): from slughifi import slughifi - + # how to put related book's slug here? if not ext: if media.type == 'daisy': @@ -194,6 +196,27 @@ def book_upload_path(ext=None, maxlen=100): return lambda *args: get_dynamic_path(*args, ext=ext, maxlen=maxlen) +def get_customized_pdf_path(book, customizations): + """ + Returns a MEDIA_ROOT relative path for a customized pdf. The name will contain a hash of customization options. + """ + customizations.sort() + h = hash(tuple(customizations)) + pdf_name = '%s-custom-%s' % (book.slug, h) + pdf_file = models.get_dynamic_path(None, pdf_name, ext='pdf') + return pdf_file + + +def get_existing_customized_pdf(book): + """ + Returns a list of paths to generated customized pdf of a book + """ + pdf_glob = '%s-custom-' % (book.slug,) + pdf_glob = get_dynamic_path(None, pdf_glob, ext='pdf') + pdf_glob = re.sub(r"[.]([a-z0-9]+)$", "*.\\1", pdf_glob) + return glob(path.join(settings.MEDIA_ROOT, pdf_glob)) + + class BookMedia(models.Model): type = models.CharField(_('type'), choices=MEDIA_FORMATS, max_length="100") name = models.CharField(_('name'), max_length="100") @@ -480,7 +503,6 @@ class Book(models.Model): from django.core.files import File from librarian import pdf from catalogue.utils import ORMDocProvider, remove_zip - from django.core.files.move import file_move_safe try: pdf_file = NamedTemporaryFile(delete=False) @@ -491,14 +513,22 @@ class Book(models.Model): ) if file_name is None: - self.pdf_file.save('%s.pdf' % self.slug, File(open(pdf_file.name))) + # we'd like to be sure not to overwrite changes happening while + # (timely) pdf generation is taking place (async celery scenario) + current_self = Book.objects.get(id=self.id) + current_self.pdf_file.save('%s.pdf' % self.slug, File(open(pdf_file.name))) + self.pdf_file = current_self.pdf_file else: - copy(pdf_file.name, path.join(settings.MEDIA_ROOT, get_dynamic_path(None, file_name, ext='pdf'))) + print "safing %s" % file_name + print "to: %s" % DefaultStorage().path(file_name) + DefaultStorage().save(file_name, File(open(pdf_file.name))) finally: unlink(pdf_file.name) - # remove zip with all pdf files + # remove cached downloadables remove_zip(settings.ALL_PDF_ZIP) + for customized_pdf in get_existing_customized_pdf(self): + unlink(customized_pdf) def build_mobi(self): """ (Re)builds the MOBI file. diff --git a/apps/catalogue/utils.py b/apps/catalogue/utils.py index 85a9dc274..3ffe9c0f6 100644 --- a/apps/catalogue/utils.py +++ b/apps/catalogue/utils.py @@ -10,6 +10,8 @@ from base64 import urlsafe_b64encode from django.http import HttpResponse, HttpResponseRedirect, Http404, HttpResponsePermanentRedirect from django.core.files.uploadedfile import UploadedFile +from django.core.files.base import File +from django.core.files.storage import DefaultStorage from django.utils.hashcompat import sha_constructor from django.conf import settings from celery.task import task @@ -145,12 +147,19 @@ class AttachmentHttpResponse(HttpResponse): self.file_path = file_path self.file_name = file_name - with open(self.file_path) as f: + with open(DefaultStorage().path(self.file_path)) as f: for chunk in read_chunks(f): self.write(chunk) @task -def create_custom_pdf(book_id, customizations, file_name): +def async_build_pdf(book_id, customizations, file_name): + """ + A celery task to generate pdf files. + Accepts the same args as Book.build_pdf, but with book id as first parameter + instead of Book instance + """ book = catalogue.models.Book.objects.get(id=book_id) - if not path.exists(file_name): + print "will gen %s" % DefaultStorage().path(file_name) + if not DefaultStorage().exists(file_name): book.build_pdf(customizations=customizations, file_name=file_name) + print "done." diff --git a/apps/catalogue/views.py b/apps/catalogue/views.py index 8169ad739..2bc1f10aa 100644 --- a/apps/catalogue/views.py +++ b/apps/catalogue/views.py @@ -28,7 +28,7 @@ from django.views.generic.list_detail import object_list from catalogue import models from catalogue import forms -from catalogue.utils import split_tags, AttachmentHttpResponse, create_custom_pdf +from catalogue.utils import split_tags, AttachmentHttpResponse, async_build_pdf from pdcounter import models as pdcounter_models from pdcounter import views as pdcounter_views from suggest.forms import PublishingSuggestForm @@ -772,12 +772,10 @@ def download_custom_pdf(request, slug): form = forms.CustomPDFForm(request.GET) if form.is_valid(): cust = form.customizations - h = hash(tuple(cust)) - pdf_name = '%s-custom-%s' % (book.slug, h) - pdf_file = path.join(settings.MEDIA_ROOT, models.get_dynamic_path(None, pdf_name, ext='pdf')) - + pdf_file = models.get_customized_pdf_path(book, cust) + if not path.exists(pdf_file): - result = create_custom_pdf.delay(book.id, cust, pdf_name) + result = async_build_pdf.delay(book.id, cust, pdf_file) result.wait() return AttachmentHttpResponse(file_name=("%s.pdf" % book.slug), file_path=pdf_file, mimetype="application/pdf") else: