def POST(self):
i = simplejson.loads(web.data())
ocaids = i.get('ocaids', [])
j = web.input(acs='1', restricted='0')
acs, restricted = bool(int(j.acs)), bool(int(j.restricted))
result = lending.is_borrowable(ocaids, acs=acs, restricted=restricted)
return delegate.RawText(simplejson.dumps(result),
content_type="application/json")
def popular_carousel(available_limit=30,
waitlist_limit=18,
loan_check_batch_size=100):
"""Renders a carousel of popular editions, which are available for
reading or borrowing, from user lists (borrowable or downloadable;
excludes daisy only).
Args:
available_limit (int) - Load the popular carousel with how
many items? (preferably divisible by 6; number of books shown
per page)
waitlist_limit (int)) - limit waitlist to how many books
loan_check_batch_size (int) - Bulk submits this many
archive.org itemids at a time to see if they are available to
be borrowed (only considers waitinglist and bookreader
borrows, no acs4)
Selected Lists:
popular.popular is a mapping of OL ids to archive.org identifiers
for popular book editions coming from the following OL lists:
/people/mekBot/lists/OL104041L is a manually curated
collection of popular available books which was constructed by
looking at goodreads
(http://www.goodreads.com/list/show/1.Best_Books_Ever) Best
Ever list. Because this list is more highly curated and has
more overall recognizable and popular books, we prioritize
drawing from this list (shuffled) first and then fallback to
other lists as this one is depleted (i.e. all books become
unavailable for checking out).
/people/openlibrary/lists/OL104411L comes from the "top 2000+
most requested print disabled eBooks in California" displayed
from the /lists page.
Popular List Construction:
https://github.com/internetarchive/openlibrary/pull/406#issuecomment-268090607
The expensive part about automatically checking the list seeds above
for availability is there's no apparent easy way to get ocaids for a
collection of editions at once. Thus, web.ctx.site.get needs be used
on each Edition (which is expensive) before a batch of editions can
be checked for availability. If we had the ocaids of list seeds upfront
and could query them in bulk, this would eliminate the problem.
As a work-around, we periodically create a flatfile cache of
the above list.seed keys mapped ahead of time to their ocaids
(i.e. `popular.popular`).
For steps on (re)generating `popular.popular`, see:
data.py popular.generate_popular_list()
Ideally, solr should be used as cache instead of hard-coded as `popular.popular`.
Returns:
returns a tuple (available_books, waitlisted_books)
"""
available_books = []
waitlisted_books = []
seeds = popular.popular
while seeds and len(available_books) < available_limit:
batch = seeds[:loan_check_batch_size]
seeds = seeds[loan_check_batch_size:]
random.shuffle(batch)
responses = lending.is_borrowable([seed[0] for seed in batch])
for seed in batch:
ocaid, key = seed
if len(available_books) == available_limit:
continue
book_data = web.ctx.site.get(key)
if book_data:
book = format_book_data(book_data)
if ocaid not in responses:
# If book is not accounted for, err on the side of inclusion
available_books.append(book)
elif 'status' in responses[ocaid]:
if responses[ocaid]['status'] == 'available':
available_books.append(book)
elif len(waitlisted_books) < waitlist_limit:
waitlisted_books.append(book)
return storify(available_books), storify(waitlisted_books)
def popular_carousel(available_limit=30, waitlist_limit=18, loan_check_batch_size=100):
"""Renders a carousel of popular editions, which are available for
reading or borrowing, from user lists (borrowable or downloadable;
excludes daisy only).
Args:
available_limit (int) - Load the popular carousel with how
many items? (preferably divisible by 6; number of books shown
per page)
waitlist_limit (int)) - limit waitlist to how many books
loan_check_batch_size (int) - Bulk submits this many
archive.org itemids at a time to see if they are available to
be borrowed (only considers waitinglist and bookreader
borrows, no acs4)
Selected Lists:
popular.popular is a mapping of OL ids to archive.org identifiers
for popular book editions coming from the following OL lists:
/people/mekBot/lists/OL104041L is a manually curated
collection of popular available books which was constructed by
looking at goodreads
(http://www.goodreads.com/list/show/1.Best_Books_Ever) Best
Ever list. Because this list is more highly curated and has
more overall recognizable and popular books, we prioritize
drawing from this list (shuffled) first and then fallback to
other lists as this one is depleted (i.e. all books become
unavailable for checking out).
/people/openlibrary/lists/OL104411L comes from the "top 2000+
most requested print disabled eBooks in California" displayed
from the /lists page.
Popular List Construction:
https://github.com/internetarchive/openlibrary/pull/406#issuecomment-268090607
The expensive part about automatically checking the list seeds above
for availability is there's no apparent easy way to get ocaids for a
collection of editions at once. Thus, web.ctx.site.get needs be used
on each Edition (which is expensive) before a batch of editions can
be checked for availability. If we had the ocaids of list seeds upfront
and could query them in bulk, this would eliminate the problem.
As a work-around, we periodically create a flatfile cache of
the above list.seed keys mapped ahead of time to their ocaids
(i.e. `popular.popular`).
For steps on (re)generating `popular.popular`, see:
data.py popular.generate_popular_list()
Ideally, solr should be used as cache instead of hard-coded as `popular.popular`.
Returns:
returns a tuple (available_books, waitlisted_books)
"""
available_books = []
waitlisted_books = []
seeds = popular.popular
while seeds and len(available_books) < available_limit:
batch = seeds[:loan_check_batch_size]
seeds = seeds[loan_check_batch_size:]
random.shuffle(batch)
responses = lending.is_borrowable([seed[0] for seed in batch])
for seed in batch:
ocaid, key = seed
if len(available_books) == available_limit:
continue
if ocaid not in responses:
# If book is not accounted for, err on the side of inclusion
available_books.append(format_book_data(web.ctx.site.get(key)))
elif 'status' in responses[ocaid]:
if responses[ocaid]['status'] == 'available':
available_books.append(
format_book_data(web.ctx.site.get(key)))
elif len(waitlisted_books) < waitlist_limit:
waitlisted_books.append(
format_book_data(web.ctx.site.get(key)))
return storify(available_books), storify(waitlisted_books)