def generate_reference_user_status(user,references):
"""Generate reference user status instances for a given set of references.
WARNING: the new instances are not saved in the database!
"""
new_ref_status = []
for ref in references:
rust = ReferenceUserStatus()
rust.owner = user
rust.reference = ref
rust.reference_pub_date = ref.pub_date
source_query = ref.sources.filter(userprofile=user.userprofile)\
.distinct().order_by("pub_date")
try:
rust.main_source = source_query.get()
except MultipleObjectsReturned:
rust.main_source = source_query.all()[0]
except ObjectDoesNotExist:
try:
rust.main_source = Reference.objects.get(url="<unknown>")
except ObjectDoesNotExist:
s = Reference(url="<unknown>",title="<unknown>",
save_count=1,
pub_date=datetime.utcfromtimestamp(0)\
.replace(tzinfo=timezone.utc))
s.save()
rust.main_source = s
new_ref_status.append(rust)
return new_ref_status
def create_reference_from_feedparser_entry(entry, date, previous_ref):
"""Takes a FeedParser entry and create a reference from it and
attributing it the publication date given in argument.
If the corresponding Reference already exists, it must be given as
the previous_ref argument, and if previous_ref is None, it will be
assumed that there is no matching Rerefence in the db.
Note: Enforce Dave Winer's recommendation for linkblog:
http://scripting.com/2014/04/07/howToDisplayTitlelessFeedItems.html
with a little twist: if a feed item has no title we will use the
(possibly truncated) description as a title and if there is no
description the link will be used. In any case the description of a
reference is set even if this description is also used for the
title.
Return a tuple with the unsaved reference and a list of tag names.
"""
url = entry.link
info = ""
tags = set()
if entry.get("tags", None):
tags = set([t.term for t in entry.tags])
if previous_ref is None:
url_truncated, did_truncate = sanitize_url(url)
if did_truncate:
# Save the full url in info to limit the loss of information
info = u"<WOM had to truncate the following URL: %s>" % url
logger.warning("Found an url of length %d (>%d) \
when importing references from feed." % (len(url), URL_MAX_LENGTH))
url = url_truncated
# set the title only for new ref (should avoid weird behaviour
# from the user point of view)
title = truncate_reference_title(
HTMLUnescape(entry.get("title") \
or strip_tags(entry.get("description")) \
or url))
ref = Reference(url=url, title=title)
else:
ref = previous_ref
ref.description = " ".join((info, entry.get("description", "")))
ref.pub_date = date
return (ref, tags)
def save(self):
"""Warning: the source will be saved as well as the related objects
(no commit options).
Returns the source.
"""
form_url, _ = sanitize_url(self.cleaned_data["url"])
form_title = self.cleaned_data["title"]
form_feed_url, _ = sanitize_url(self.cleaned_data["feed_url"])
if self.user.userprofile.web_feeds.filter(
source__url=form_url).exists():
# nothing to do
return
# try a bigger look-up anyway
same_sources = WebFeed.objects.filter(source__url=form_url).all()
# url are unique for sources
if same_sources:
new_feed = same_sources[0]
else:
if form_title:
source_title = form_title
else:
source_title = build_reference_title_from_url(form_url)
try:
source_ref = Reference.objects.get(url=form_url)
except ObjectDoesNotExist:
source_ref = Reference(url=form_url,
title=source_title,
pub_date=datetime.now(timezone.utc))
source_ref.save()
new_feed = WebFeed(source=source_ref)
# assume that either form_feed_url or form_url have been
# validated as a valid feed url
new_feed.xmlURL = form_feed_url or form_url
new_feed.last_update_check = datetime.utcfromtimestamp(0)\
.replace(tzinfo=timezone.utc)
new_feed.save()
self.user.userprofile.sources.add(source_ref)
self.user.userprofile.public_sources.add(source_ref)
self.user.userprofile.web_feeds.add(new_feed)
return new_feed
def import_feedsources_from_opml(opml_txt):
"""
Save in the db the FeedSources found in the OPML-formated text.
opml_txt: a unicode string representing the content of a full OPML file.
Return a dictionary assiociating each feed with a set of tags {feed:tagSet,...).
"""
collected_feeds, _ = parse_opml(opml_txt, False)
db_new_feedsources = []
feeds_and_tags = []
newly_referenced_source = []
for current_feed in collected_feeds:
try:
feed_source = WebFeed.objects.get(xmlURL=current_feed.xmlUrl)
except MultipleObjectsReturned:
feed_source = WebFeed.objects.all()[0]
except ObjectDoesNotExist:
url_id = current_feed.htmlUrl or current_feed.xmlUrl
try:
ref = Reference.objects.get(url=url_id)
except ObjectDoesNotExist:
ref = Reference(url=url_id,
title=HTMLUnescape(current_feed.title),
pub_date=datetime.now(timezone.utc))
ref.save()
feed_source = WebFeed(source=ref, xmlURL=current_feed.xmlUrl)
feed_source.last_update_check = datetime.utcfromtimestamp(0)\
.replace(tzinfo=timezone.utc)
newly_referenced_source.append(ref)
db_new_feedsources.append(feed_source)
feeds_and_tags.append((feed_source, current_feed.tags))
with transaction.commit_on_success():
for f in db_new_feedsources:
f.save()
# make sure to record the fact
for r in newly_referenced_source:
r.save_count += 1
r.save()
return dict(feeds_and_tags)
def save(self):
"""Warning: the bookmark will be saved as well as the related objects
(no commit options).
Returns the bookmark.
"""
url, _ = sanitize_url(self.cleaned_data["url"])
title = self.cleaned_data["title"] \
or build_reference_title_from_url(url)
comment = self.cleaned_data["comment"]
pub_date = self.cleaned_data["pub_date"] \
or datetime.now(timezone.utc)
src_url,_ = sanitize_url(self.cleaned_data["source_url"] \
or build_source_url_from_reference_url(url))
src_title = self.cleaned_data["source_title"] \
or build_reference_title_from_url(src_url)
# Find or create a matching reference
try:
bookmarked_ref = Reference.objects.get(url=url)
# Arbitrarily chose one of the possible sources
src_query = bookmarked_ref.sources
if src_query.count() > 1:
ref_src = src_query.all()[0]
else:
ref_src = src_query.get()
except ObjectDoesNotExist:
try:
ref_src = Reference.objects.get(url=src_url)
except ObjectDoesNotExist:
ref_src = Reference(url=src_url,
title=src_title,
pub_date=pub_date)
ref_src.save()
if src_url == url:
bookmarked_ref = ref_src
else:
bookmarked_ref = Reference(url=url,
title=title,
pub_date=pub_date)
bookmarked_ref.save()
bookmarked_ref.sources.add(ref_src)
with transaction.commit_on_success():
try:
bmk = UserBookmark.objects.get(owner=self.user,
reference=bookmarked_ref)
except ObjectDoesNotExist:
bmk = UserBookmark(owner=self.user,
reference=bookmarked_ref,
saved_date=datetime.now(timezone.utc))
bookmarked_ref.save_count += 1
bmk.save()
bookmarked_ref.save()
# allow the user-specific comment to be changed and also prefix
# it with the user specified title if it differs from the
# existing reference title.
if comment:
new_comment = comment
else:
new_comment = bmk.comment
if self.cleaned_data["title"] and title!=bookmarked_ref.title \
and not new_comment.startswith(title):
new_comment = "%s: %s" % (title, new_comment)
if new_comment != bmk.comment:
bmk.comment = new_comment
bmk.save()
with transaction.commit_on_success():
if ref_src not in self.user.userprofile.sources.all():
self.user.userprofile.sources.add(ref_src)
with transaction.commit_on_success():
for rust in ReferenceUserStatus\
.objects.filter(owner=self.user,
reference=bookmarked_ref).all():
rust.has_been_saved = True
rust.save()
return bmk
def import_references_from_ns_bookmark_list(nsbmk_txt):
"""Extract bookmarks from a Netscape-style bookmark file and save
them as Reference instances in the database.
nsbmk_txt: a unicode string representing the full content of a
Netscape-style bookmark file.
Return a dictionary mapping each reference with the BookmarkMetadata
associated to it according to the input content.
"""
date_now = datetime.datetime.now(timezone.utc)
# Parse the file
collected_bmks = parse_netscape_bookmarks(nsbmk_txt)
if not collected_bmks:
return {}
# Make sure that the source common to all the following import
# exists or create it to be able to link new references to it.
source_url = "#internal-bookmark-import"
try:
common_source = Reference.objects.get(url=source_url)
except ObjectDoesNotExist:
common_source = Reference(url=source_url,
title="Bookmark Import",
pub_date=date_now)
common_source.save()
new_refs = []
ref_and_metadata = []
new_ref_by_url = {}
for bmk_info in collected_bmks:
u = bmk_info["url"]
if not u:
logger.warning("Skipping a bookmark that has an empty URL.")
continue
info = ""
u_truncated, did_truncate = sanitize_url(u)
if did_truncate:
# Save the full url in info to limit the loss of information
info = u"<WOM had to truncate the following URL: %s>" % u
logger.warning("Found an url of length %d (>%d) \
when importing Netscape-style bookmark list." % (len(u),URL_MAX_LENGTH))
u = u_truncated
t = bmk_info.get("title") or build_reference_title_from_url(u)
if "posix_timestamp" in bmk_info:
d = datetime.datetime\
.utcfromtimestamp(float(bmk_info["posix_timestamp"]))\
.replace(tzinfo=timezone.utc)
else:
d = date_now
if u in new_ref_by_url:
ref = new_ref_by_url[u]
else:
try:
ref = Reference.objects.get(url=u)
except ObjectDoesNotExist:
ref = Reference(url=u,title=truncate_reference_title(t),
pub_date=d,description=info)
new_refs.append(ref)
new_ref_by_url[u] = ref
meta = BookmarkMetadata(bmk_info.get("note",""),
set(bmk_info.get("tags","").split(",")),
bmk_info.get("private","0")=="0")
ref_and_metadata.append((ref,meta))
with transaction.commit_on_success():
for ref in new_refs:
ref.save()
ref.sources.add(common_source)
# Note: We have to wait until now to convert the list to a dict,
# because only now will the model instances have their specific ids and
# hashes (before that they would have looked the same for the dict).
return dict(ref_and_metadata)
