regluit/core/models/bibmodels.py

import logging
import math
import re
import urllib
import urllib2
import uuid

from decimal import Decimal
import unicodedata
from urlparse import urlparse

from sorl.thumbnail import get_thumbnail
from PIL import ImageFile

from django.conf import settings
from django.contrib.auth.models import User
from django.contrib.contenttypes.fields import GenericRelation
from django.core.files.base import ContentFile
from django.urls import reverse
from django.db import models
from django.db.models import F
from django.db.models.signals import post_save, pre_delete
from django.utils.timezone import now

from django_comments.models import Comment

import regluit
from regluit.marc.models import MARCRecord as NewMARC
from questionnaire.models import Landing

from regluit.core import mobi
import regluit.core.cc as cc
from regluit.core.epub import test_epub
from regluit.core.links import id_url
from regluit.core.validation import valid_subject

from regluit.core.parameters import (
    AGE_LEVEL_CHOICES,
    BORROWED,
    BUY2UNGLUE,
    ID_CHOICES_MAP,
    INDIVIDUAL,
    LIBRARY,
    OFFER_CHOICES,
    TESTING,
    TEXT_RELATION_CHOICES,
    THANKED,
    THANKS,
    WORK_IDENTIFIERS,
)

# fix truncated file problems per https://stackoverflow.com/questions/12984426/python-pil-ioerror-image-file-truncated-with-big-images
ImageFile.LOAD_TRUNCATED_IMAGES = True

logger = logging.getLogger(__name__)
good_providers = ('Internet Archive', 'Unglue.it', 'Github', 'OAPEN Library')

def id_for(obj, type):
    if not obj.pk:
        return ''
    try:
        return obj.identifiers.filter(type=type)[0].value
    except IndexError:
        return ''

class Identifier(models.Model):
    # olib, ltwk, goog, gdrd, thng, isbn, oclc, olwk, doab, gtbg, glue, doi
    type = models.CharField(max_length=4, null=False)
    value = models.CharField(max_length=250, null=False)
    work = models.ForeignKey("Work", on_delete=models.CASCADE, related_name="identifiers", null=False)
    edition = models.ForeignKey("Edition", on_delete=models.CASCADE, related_name="identifiers", null=True)

    class Meta:
        unique_together = ("type", "value")

    @staticmethod
    def set(type=None, value=None, edition=None, work=None):
        # if there's already an id of this type for this work and edition, change it
        # if not, create it. if the id exists and points to something else, change it.
        try:
            identifier = Identifier.objects.filter(type=type, value=value)[0]
        except IndexError:
            if type in WORK_IDENTIFIERS:
                identifier = Identifier.objects.create(type=type, value=value, work=work)
            else:
                identifier = Identifier.objects.create(type=type, value=value, work=work, edition=edition)
        if identifier.work_id != work.id:
            identifier.work = work
            identifier.save()
        if identifier.edition and edition:
            if identifier.edition_id != edition.id:
                identifier.edition = edition
                identifier.save()
            others = Identifier.objects.filter(type=type, work=work, edition=edition).exclude(value=value)
            if others.count() > 0:
                for other in others:
                    other.delete()
        return identifier

    @staticmethod
    def get_or_add(type='goog', value=None, edition=None, work=None):
        try:
            return Identifier.objects.get(type=type, value=value)
        except Identifier.DoesNotExist:
            i = Identifier(type=type, value=value, edition=edition, work=work)
            i.save()
            return i

    def __unicode__(self):
        return u'{0}:{1}'.format(self.type, self.value)
        
    def label(self):
        return ID_CHOICES_MAP.get(self.type, self.type)
        
    def url(self):
        return id_url(self.type, self.value)

class Work(models.Model):
    created = models.DateTimeField(auto_now_add=True, db_index=True,)
    title = models.CharField(max_length=1000)
    language = models.CharField(max_length=5, default="en", null=False, db_index=True,)
    openlibrary_lookup = models.DateTimeField(null=True, blank=True)
    num_wishes = models.IntegerField(default=0, db_index=True)
    description = models.TextField(default='', null=True, blank=True)
    selected_edition = models.ForeignKey("Edition", on_delete=models.CASCADE, related_name='selected_works', null=True)
    # repurposed earliest_publication to actually be publication range
    publication_range = models.CharField(max_length=50, null=True, blank=True)
    featured = models.DateTimeField(null=True, blank=True, db_index=True,)
    is_free = models.BooleanField(default=False)
    landings = GenericRelation(Landing, related_query_name='works')
    related = models.ManyToManyField('self', symmetrical=False, blank=True, through='WorkRelation', related_name='reverse_related')
    age_level = models.CharField(max_length=5, choices=AGE_LEVEL_CHOICES, default='', blank=True) 

    class Meta:
        ordering = ['title']
    
    def __unicode__(self):
        return self.title

    def __init__(self, *args, **kwargs):
        self._last_campaign = None
        super(Work, self).__init__(*args, **kwargs)

    def delete(self, cascade=True, *args, **kwargs):
        if cascade:
            if self.offers.all() or self.claim.all() or self.campaigns.all() or self.acqs.all() \
                 or self.holds.all() or self.landings.all():
                 return
            for wishlist in self.wishlists.all():
                wishlist.remove_work(self)
            for userprofile in self.contributors.all():
                userprofile.works.remove(self)
            for identifier in self.identifiers.all():
                identifier.delete()
            for comment in Comment.objects.for_model(self):
                comment.delete()
            for edition in self.editions.all():
                for ebook in edition.ebooks.all():
                    ebook.delete()
                for ebookfile in edition.ebook_files.all():
                    ebookfile.delete()
                edition.delete()
            for work_relation in self.works_related_to.all():
                work_relation.delete()
            for work_relation in self.works_related_from.all():
                work_relation.delete()
        super(Work, self).delete(*args, **kwargs)  # Call the "real" save() method.
        
    def id_for(self, type):
        return id_for(self, type)

    @property
    def gtbg(self):
        return id_for(self, 'gtbg')

    @property
    def doab(self):
        return id_for(self, 'doab')

    @property
    def doi(self):
        return self.id_for('doi')

    @property
    def http_id(self):
        return self.id_for('http')

    @property
    def googlebooks_id(self):
        try:
            preferred_id = self.preferred_edition.googlebooks_id
            # note that there should always be a preferred edition
        except AttributeError:
            # this work has no edition.
            return ''
        if preferred_id:
            return preferred_id
        try:
            return self.identifiers.filter(type='goog')[0].value
        except IndexError:
            return ''

    @property
    def googlebooks_url(self):
        return id_url('goog', self.googlebooks_id)

    @property
    def goodreads_id(self):
        preferred_id = self.preferred_edition.goodreads_id
        if preferred_id:
            return preferred_id
        try:
            return self.identifiers.filter(type='gdrd')[0].value
        except IndexError:
            return ''

    @property
    def goodreads_url(self):
        return id_url('gdrd', self.goodreads_id)

    @property
    def librarything_id(self):
        return self.id_for('ltwk')

    @property
    def librarything_url(self):
        return id_url('ltwk', self.librarything_id)

    @property
    def openlibrary_id(self):
        return self.id_for('olwk')

    @property
    def openlibrary_url(self):
        return id_url('olwk', self.openlibrary_id)
        
    def cover_filetype(self):
        if self.uses_google_cover():
            return 'jpeg'
        else:
            # consider the path only and not the params, query, or fragment
            url = urlparse(self.cover_image_small().lower()).path

            if url.endswith('.png'):
                return 'png'
            elif url.endswith('.gif'):
                return 'gif'
            elif url.endswith('.jpg') or url.endswith('.jpeg'):
                return 'jpeg'
            else:
                return 'image'

    def work_ids(self):
        return self.identifiers.filter(edition__isnull=True)

    def uses_google_cover(self):
        if self.preferred_edition and self.preferred_edition.cover_image:
            return False
        else:
            return self.googlebooks_id

    def cover_image_large(self):
        if self.preferred_edition and self.preferred_edition.has_cover_image():
            return self.preferred_edition.cover_image_large()
        return "/static/images/generic_cover_larger.png"

    def cover_image_small(self):
        if self.preferred_edition and self.preferred_edition.has_cover_image():
            return self.preferred_edition.cover_image_small()
        return "/static/images/generic_cover_larger.png"

    def cover_image_thumbnail(self):
        try:
            if self.preferred_edition and self.preferred_edition.has_cover_image():
                return self.preferred_edition.cover_image_thumbnail()
        except IndexError:
            pass
        return "/static/images/generic_cover_larger.png"

    def authors(self):
        # assumes that they come out in the same order they go in!
        if self.preferred_edition and self.preferred_edition.authors.all().count() > 0:
            return  self.preferred_edition.authors.all()
        for edition in self.editions.all():
            if edition.authors.all().count() > 0:
                return edition.authors.all()
        return Author.objects.none()

    def relators(self):
        # assumes that they come out in the same order they go in!
        if self.preferred_edition and self.preferred_edition.relators.all().count() > 0:
            return  self.preferred_edition.relators.all()
        for edition in self.editions.all():
            if edition.relators.all().count() > 0:
                return edition.relators.all()
        return Relator.objects.none()

    def author(self):
        # assumes that they come out in the same order they go in!
        if self.relators().count() > 0:
            return self.relators()[0].name
        return ''

    def authors_short(self):
        # assumes that they come out in the same order they go in!
        if self.relators().count() == 1:
            return self.relators()[0].name
        elif self.relators().count() == 2:
            if self.relators()[0].relation == self.relators()[1].relation:
                if self.relators()[0].relation.code == 'aut':
                    return "%s and %s" % (self.relators()[0].author.name, self.relators()[1].author.name)
                else:
                    return "%s and %s, %ss" % (self.relators()[0].author.name, self.relators()[1].author.name, self.relators()[0].relation.name)
            else:
                return "%s (%s) and %s (%s)" % (self.relators()[0].author.name, self.relators()[0].relation.name, self.relators()[1].author.name, self.relators()[1].relation.name)
        elif self.relators().count() > 2:
            auths = self.relators()
            if auths[0].relation.code == 'aut':
                return "%s et al." % auths[0].author.name
            else:
                return "%s et al. (%ss)" % (auths[0].author.name, auths[0].relation.name)
        return ''

    def kindle_safe_title(self):
        """
        Removes accents, keeps letters and numbers, replaces non-Latin characters with "#", and replaces punctuation with "_"
        """
        safe = u''
        nkfd_form = unicodedata.normalize('NFKD', self.title) #unaccent accented letters
        for c in nkfd_form:
            ccat = unicodedata.category(c)
            #print ccat
            if ccat.startswith('L') or  ccat.startswith('N'): # only letters and numbers
                if ord(c) > 127:
                    safe = safe + '#' #a non latin script letter or number
                else:
                    safe = safe + c
            elif not unicodedata.combining(c): #not accents (combining forms)
                safe = safe + '_' #punctuation
        return safe

    def last_campaign(self):
        # stash away the last campaign to prevent repeated lookups
        if hasattr(self, '_last_campaign_'):
            return self._last_campaign_
        try:
            self._last_campaign_ = self.campaigns.order_by('-created')[0]
        except IndexError:
            self._last_campaign_ = None
        return self._last_campaign_

    @property
    def preferred_edition(self):
        if self.selected_edition:
            return self.selected_edition
        if self.last_campaign():
            if self.last_campaign().edition:
                self.selected_edition = self.last_campaign().edition
                self.save()
                return self.last_campaign().edition
        try:
            self.selected_edition = self.editions.all().order_by('-cover_image', '-created')[0] # prefer editions with covers
            self.save()
            return self.selected_edition
        except IndexError:
            #should only happen if there are no editions for the work,
            #which can happen when works are being merged
            try:
                return WasWork.objects.get(was=self.id).work.preferred_edition
            except WasWork.DoesNotExist:
                #should not happen
                logger.warning('work {} has no edition'.format(self.id))
                return None

    def last_campaign_status(self):
        campaign = self.last_campaign()
        if campaign:
            status = campaign.status
        else:
            if self.first_ebook():
                status = "Available"
            else:
                status = "No campaign yet"
        return status

    def percent_unglued(self):
        status = 0
        campaign = self.last_campaign()
        if campaign is not None:
            if campaign.status == 'SUCCESSFUL':
                status = 6
            elif campaign.status == 'ACTIVE':
                if campaign.target is not None:
                    target = float(campaign.target)
                else:
                    #shouldn't happen, but did once because of a failed pdf conversion
                    return 0
                if target <= 0:
                    status = 6
                else:
                    if campaign.type == BUY2UNGLUE:
                        status = int(6 - 6*campaign.left/campaign.target)
                    else:
                        status = int(float(campaign.current_total)*6/target)
                    if status >= 6:
                        status = 6
        return status

    def percent_of_goal(self):
        campaign = self.last_campaign()
        return 0 if campaign is None else campaign.percent_of_goal()

    def ebooks_all(self):
        return self.ebooks(all=True)

    def ebooks(self, all=False):
        if all:
            return Ebook.objects.filter(edition__work=self).order_by('-created')
        else:
            return Ebook.objects.filter(edition__work=self, active=True).order_by('-created')

    def ebookfiles(self):
        return EbookFile.objects.filter(edition__work=self).exclude(file='').order_by('-created')

    def epubfiles(self):
        # filter out non-epub because that's what booxtream accepts
        return EbookFile.objects.filter(edition__work=self, format='epub').exclude(file='').order_by('-created')

    def mobifiles(self):
        return EbookFile.objects.filter(edition__work=self, format='mobi').exclude(file='').order_by('-created')

    def pdffiles(self):
        return EbookFile.objects.filter(edition__work=self, format='pdf').exclude(file='').order_by('-created')
    
    def versions(self):
        version_labels = []
        for ebook in self.ebooks_all():
            if ebook.version_label and not ebook.version_label in version_labels:
                version_labels.append(ebook.version_label)
        return version_labels
    
    def formats(self):
        fmts = []
        for fmt in ['pdf', 'epub', 'mobi', 'html']:
            for ebook in self.ebooks().filter(format=fmt):
                fmts.append(fmt)
                break
        return fmts

    def remove_old_ebooks(self):
        # this method is triggered after an file upload or new ebook saved
        old = Ebook.objects.filter(edition__work=self, active=True).order_by('-version_iter', '-created')
        
        # keep highest version ebook for each format and version label
        done_format_versions = []
        for eb in old:
            format_version = '{}_{}'.format(eb.format, eb.version_label)
            if format_version in done_format_versions:
                eb.deactivate()
            else:
                done_format_versions.append(format_version)
        
        # check for failed uploads.
        null_files = EbookFile.objects.filter(edition__work=self, file='')
        for ebf in null_files:
            ebf.file.delete()
            ebf.delete()

    @property
    def download_count(self):
        dlc = 0
        for ebook in self.ebooks(all=True):
            dlc += ebook.download_count
        return dlc

    def first_pdf(self):
        return self.first_ebook('pdf')

    def first_epub(self):
        return self.first_ebook('epub')

    def first_pdf_url(self):
        try:
            url = self.first_ebook('pdf').url
            return url
        except:
            return None

    def first_epub_url(self):
        try:
            url = self.first_ebook('epub').url
            return url
        except:
            return None

    def first_ebook(self, ebook_format=None):
        if ebook_format:
            for ebook in self.ebooks().filter(format=ebook_format):
                return ebook
        else:
            for ebook in self.ebooks():
                return ebook

    def wished_by(self, excluding=None):
        if excluding:
            return User.objects.filter(wishlist__wishes__work=self).exclude(wishlist__wishes__created__range=excluding)
        else:
            return User.objects.filter(wishlist__wishes__work=self)

    def update_num_wishes(self):
        self.num_wishes = self.wishes.count()
        self.save()

    def priority(self):
        if self.last_campaign():
            return 5
        freedom = 1 if self.is_free else 0
        wishing = int(math.log(self.num_wishes)) + 1 if self.num_wishes else 0
        return min(freedom + wishing, 5)

    def first_oclc(self):
        if self.preferred_edition is None:
            return ''
        preferred_id = self.preferred_edition.oclc
        if preferred_id:
            return preferred_id
        try:
            return self.identifiers.filter(type='oclc')[0].value
        except IndexError:
            return ''

    def first_isbn_13(self):
        if self.preferred_edition is None:
            return ''
        preferred_id = self.preferred_edition.isbn_13
        if preferred_id:
            return preferred_id
        try:
            return self.identifiers.filter(type='isbn')[0].value
        except IndexError:
            return ''

    @property
    def earliest_publication_date(self):
        for edition in Edition.objects.filter(work=self, publication_date__isnull=False).order_by('publication_date'):
            if edition.publication_date and len(edition.publication_date) >= 4:
                return edition.publication_date

    @property
    def publication_date(self):
        if self.publication_range:
            return  self.publication_range
        for edition in Edition.objects.filter(work=self, publication_date__isnull=False).order_by('publication_date'):
            if edition.publication_date:
                try:
                    earliest_publication = edition.publication_date[:4]
                except IndexError:
                    continue
                latest_publication = None
                for edition in Edition.objects.filter(work=self, publication_date__isnull=False).order_by('-publication_date'):
                    if edition.publication_date:
                        try:
                            latest_publication = edition.publication_date[:4]
                        except IndexError:
                            continue
                        break
                if earliest_publication == latest_publication:
                    publication_range = earliest_publication
                else:
                    publication_range = earliest_publication + "-" + latest_publication
                self.publication_range = publication_range
                self.save()
                return publication_range
        return ''

    @property
    def has_unglued_edition(self):
        """
        allows us to distinguish successful campaigns with ebooks still in progress from successful campaigns with ebooks available
        """
        if self.ebooks().filter(edition__unglued=True):
            return True
        return False

    @property
    def user_with_rights(self):
        """
        return queryset of users (should be at most one) who act for rights holders with active claims to the work
        """
        claims = self.claim.filter(status='active')
        assert claims.count() < 2, "There is more than one active claim on %r" % self.title
        try:
            return claims[0].user
        except:
            return False

    def get_absolute_url(self):
        return reverse('work', args=[str(self.id)])

    def publishers(self):
        # returns a set of publishers associated with this Work
        return Publisher.objects.filter(name__editions__work=self).distinct()

    def create_offers(self):
        for choice in OFFER_CHOICES:
            if not self.offers.filter(license=choice[0]):
                self.offers.create(license=choice[0], active=True, price=Decimal(10))
        return self.offers.all()

    def get_lib_license(self, user):
        lib_user = (lib.user for lib in user.profile.libraries)
        return self.get_user_license(lib_user)

    def borrowable(self, user):
        if user.is_anonymous:
            return False
        lib_license = self.get_lib_license(user)
        if lib_license and lib_license.borrowable:
            return True
        return False

    def lib_thanked(self, user):
        if user.is_anonymous:
            return False
        lib_license = self.get_lib_license(user)
        if lib_license and lib_license.thanked:
            return True
        return False

    def in_library(self, user):
        if user.is_anonymous:
            return False
        lib_license = self.get_lib_license(user)
        if lib_license and lib_license.acqs.count():
            return True
        return False

    @property
    def lib_acqs(self):
        return  self.acqs.filter(license=LIBRARY)

    @property
    def test_acqs(self):
        return  self.acqs.filter(license=TESTING).order_by('-created')

    class user_license:
        acqs = Identifier.objects.none() # Identifier is just convenient.
        def __init__(self, acqs):
            self.acqs = acqs

        @property
        def is_active(self):
            return  self.acqs.filter(expires__isnull=True).count() > 0 or self.acqs.filter(expires__gt=now()).count() > 0

        @property
        def borrowed(self):
            loans = self.acqs.filter(license=BORROWED, expires__gt=now())
            if loans.count() == 0:
                return None
            else:
                return loans[0]

        @property
        def purchased(self):
            purchases = self.acqs.filter(license=INDIVIDUAL, expires__isnull=True)
            if purchases.count() == 0:
                return None
            else:
                return purchases[0]

        @property
        def lib_acqs(self):
            return  self.acqs.filter(license=LIBRARY)

        @property
        def next_acq(self):
            """ This is the next available copy in the user's libraries"""
            loans = self.acqs.filter(license=LIBRARY, refreshes__gt=now()).order_by('refreshes')
            if loans.count() == 0:
                return None
            else:
                return loans[0]

        @property
        def borrowable(self):
            return  self.acqs.filter(license=LIBRARY, refreshes__lt=now()).count() > 0

        @property
        def thanked(self):
            return  self.acqs.filter(license=THANKED).count() > 0

        @property
        def borrowable_acq(self):
            for acq in self.acqs.filter(license=LIBRARY, refreshes__lt=now()):
                return acq

        @property
        def is_duplicate(self):
            # does user have two individual licenses?
            pending = self.acqs.filter(license=INDIVIDUAL, expires__isnull=True, gifts__used__isnull=True).count()
            return self.acqs.filter(license=INDIVIDUAL, expires__isnull=True).count() > pending


    def get_user_license(self, user):
        """ This is all the acqs, wrapped in user_license object for the work, user(s) """
        if user is None:
            return None
        if hasattr(user, 'is_anonymous'):
            if user.is_anonymous:
                return None
            return self.user_license(self.acqs.filter(user=user))
        else:
            # assume it's several users
            return self.user_license(self.acqs.filter(user__in=user))

    @property
    def has_marc(self):
        for record in  NewMARC.objects.filter(edition__work=self):
            return True
        return False

    ### for compatibility with MARC output
    def marc_records(self):
        record_list = []
        record_list.extend(NewMARC.objects.filter(edition__work=self))
        for obj in record_list:
            break
        else:
            for ebook in self.ebooks():
                record_list.append(ebook.edition)
                break
        return record_list

class WorkRelation(models.Model):
    to_work = models.ForeignKey('Work', on_delete=models.CASCADE, related_name='works_related_to')
    from_work= models.ForeignKey('Work', on_delete=models.CASCADE, related_name='works_related_from')
    relation = models.CharField(max_length=15, choices=TEXT_RELATION_CHOICES)


class Author(models.Model):
    created = models.DateTimeField(auto_now_add=True)
    name = models.CharField(max_length=255, unique=True)
    editions = models.ManyToManyField("Edition", related_name="authors", through="Relator")

    def __unicode__(self):
        return self.name

    @property
    def last_name_first(self):
        names = self.name.rsplit()
        if len(names) == 0:
            return ''
        elif len(names) == 1:
            return names[0]
        elif len(names) == 2:
            return names[1] + ", " + names[0]
        else:
            reversed_name = names[-1]+","
            for name in names[0:-1]:
                reversed_name += " "
                reversed_name += name
            return reversed_name

class Relation(models.Model):
    code = models.CharField(max_length=3, blank=False, db_index=True, unique=True)
    name = models.CharField(max_length=30, blank=True,)

class Relator(models.Model):
    relation = models.ForeignKey('Relation', on_delete=models.CASCADE, default=1) #first relation should have code='aut'
    author = models.ForeignKey('Author', on_delete=models.CASCADE)
    edition = models.ForeignKey('Edition', on_delete=models.CASCADE, related_name='relators')
    class Meta:
        db_table = 'core_author_editions'

    @property
    def name(self):
        if self.relation.code == 'aut':
            return self.author.name
        else:
            return "%s (%s)" % (self.author.name, self.relation.name)

    def set(self, relation_code):
        if self.relation.code != relation_code:
            try:
                self.relation = Relation.objects.get(code=relation_code)
                self.save()
            except Relation.DoesNotExist:
                logger.warning("relation not found: code = %s" % relation_code)

class Subject(models.Model):
    created = models.DateTimeField(auto_now_add=True)
    name = models.CharField(max_length=200, unique=True)
    works = models.ManyToManyField("Work", related_name="subjects")
    is_visible = models.BooleanField(default=True)
    authority = models.CharField(max_length=10, blank=False, default="")

    class Meta:
        ordering = ['name']
    
    @classmethod
    def set_by_name(cls, subject, work=None, authority=None):
        ''' use this method whenever you would be creating a new subject!'''
        subject = subject.strip()
        
        # make sure it's not a ; delineated list
        subjects = subject.split(';')
        for additional_subject in subjects[1:]:
            cls.set_by_name(additional_subject, work, authority)
        subject = subjects[0]
        # make sure there's no heading
        headingmatch = re.match(r'^!(.+):(.+)', subject)
        if headingmatch:
            subject = headingmatch.group(2).strip()
            authority = headingmatch.group(1).strip()
        elif subject.startswith('nyt:'):
            subject = subject[4:].split('=')[0].replace('_', ' ').strip().capitalize()
            subject = 'NYT Bestseller - {}'.format(subject)
            authority = 'nyt'
        elif subject.startswith('award:'):
            subject = subject[6:].split('=')[0].replace('_', ' ').strip().capitalize()
            subject = 'Award Winner - {}'.format(subject)
            authority = 'award'

        if valid_subject(subject):
            (subject_obj, created) = cls.objects.get_or_create(name=subject)
            if not subject_obj.authority and authority:
                subject_obj.authority = authority
                subject_obj.save()
        
            subject_obj.works.add(work)
            return subject_obj   
        else:
            return None
    
    def __unicode__(self):
        return self.name


    @property
    def kw(self):
        return 'kw.%s' % self.name

    def free_works(self):
        return self.works.filter(is_free=True)

class Edition(models.Model):
    created = models.DateTimeField(auto_now_add=True)
    title = models.CharField(max_length=1000)
    publisher_name = models.ForeignKey("PublisherName", on_delete=models.CASCADE, related_name="editions", null=True, blank=True)
    publication_date = models.CharField(max_length=50, null=True, blank=True, db_index=True)
    work = models.ForeignKey("Work", on_delete=models.CASCADE, related_name="editions", null=True)
    cover_image = models.URLField(null=True, blank=True)
    unglued = models.BooleanField(default=False)
    note = models.ForeignKey("EditionNote", on_delete=models.CASCADE, null=True, blank=True)

    def __unicode__(self):
        if self.isbn_13:
            return "%s (ISBN %s) %s" % (self.title, self.isbn_13, self.publisher)
        if self.oclc:
            return "%s (OCLC %s) %s" % (self.title, self.oclc, self.publisher)
        if self.googlebooks_id:
            return "%s (GOOG %s) %s" % (self.title, self.googlebooks_id, self.publisher)
        else:
            return "%s (GLUE %s) %s" % (self.title, self.id, self.publisher)

    def cover_image_large(self):
        #550 pixel high image
        if self.cover_image:
            try:
                im = get_thumbnail(self.cover_image, 'x550', crop='noop', quality=95)
                if im.exists():
                    return im.url
            except IOError:
                pass
        elif self.googlebooks_id:
            url = "https://encrypted.google.com/books?id=%s&printsec=frontcover&img=1&zoom=0" % self.googlebooks_id
            try:
                im = get_thumbnail(url, 'x550', crop='noop', quality=95)
                if not im.exists() or im.storage.size(im.name) == 16392: # check for "image not available" image
                    url = "https://encrypted.google.com/books?id=%s&printsec=frontcover&img=1&zoom=1" % self.googlebooks_id
                    im = get_thumbnail(url, 'x550', crop='noop', quality=95)
                if im.exists():
                    return im.url
            except IOError:
                pass
        return ''

    def cover_image_small(self):
        #80 pixel high image
        if self.cover_image:
            try:
                im = get_thumbnail(self.cover_image, 'x80', crop='noop', quality=95)
                if im.exists():
                    return im.url
            except IOError:
                pass
        if self.googlebooks_id:
            return "https://encrypted.google.com/books?id=%s&printsec=frontcover&img=1&zoom=5" % self.googlebooks_id
        return ''

    def cover_image_thumbnail(self):
        #128 pixel wide image
        if self.cover_image:
            try:
                im = get_thumbnail(self.cover_image, '128', crop='noop', quality=95)
                if im.exists():
                    return im.url
            except IOError:
                pass
        if self.googlebooks_id:
            return "https://encrypted.google.com/books?id=%s&printsec=frontcover&img=1&zoom=1" % self.googlebooks_id
        else:
            return ''

    def has_cover_image(self):
        if self.cover_image:
            return self.cover_image
        elif self.googlebooks_id:
            return True
        else:
            return False

    @property
    def publisher(self):
        if self.publisher_name:
            return self.publisher_name.name
        return ''

    @property
    def isbn_10(self):
        return regluit.core.isbn.convert_13_to_10(self.isbn_13)

    def id_for(self, type):
        if type in WORK_IDENTIFIERS:
            return self.work.id_for(type)
        return id_for(self, type)

    @property
    def isbn_13(self):
        return self.id_for('isbn')

    @property
    def googlebooks_id(self):
        return self.id_for('goog')

    @property
    def librarything_id(self):
        return self.id_for('thng')

    @property
    def oclc(self):
        return self.id_for('oclc')

    @property
    def goodreads_id(self):
        return self.id_for('gdrd')

    @staticmethod
    def get_by_isbn(isbn):
        if len(isbn) == 10:
            isbn = regluit.core.isbn.convert_10_to_13(isbn)
        try:
            return Identifier.objects.get(type='isbn', value=isbn).edition
        except Identifier.DoesNotExist:
            return None

    def add_author(self, author_name, relation='aut'):
        if author_name:
            (author, created) = Author.objects.get_or_create(name=author_name)
            (relation, created) = Relation.objects.get_or_create(code=relation)
            (new_relator, created) = Relator.objects.get_or_create(author=author, edition=self)
            if new_relator.relation != relation:
                new_relator.relation = relation
                new_relator.save()

    def remove_author(self, author):
        if author:
            try:
                relator = Relator.objects.get(author=author, edition=self)
                relator.delete()
            except Relator.DoesNotExist:
                pass

    def set_publisher(self, publisher_name):
        if publisher_name and publisher_name != '':
            try:
                pub_name = PublisherName.objects.get(name=publisher_name)
                if pub_name.publisher:
                    pub_name = pub_name.publisher.name
            except PublisherName.DoesNotExist:
                pub_name = PublisherName.objects.create(name=publisher_name)
                pub_name.save()

            self.publisher_name = pub_name
            self.save()

    #### following methods for compatibility with marc outputter
    def downloads(self):
        return self.ebooks.filter(active=True)

    def download_via_url(self):
        return settings.BASE_URL_SECURE + reverse('download', args=[self.work_id])

    def authnames(self):
        return [auth.last_name_first for auth in self.authors.all()]

    @property
    def license(self):
        try:
            return self.ebooks.all()[0].rights
        except:
            return None

    @property
    def funding_info(self):
        if self.ebooks.all().count() == 0:
            return ''
        if self.unglued:
            return 'The book is available as a free download thanks to the generous support of interested readers and organizations, who made donations using the crowd-funding website Unglue.it.'
        else:
            if self.ebooks.all()[0].rights in cc.LICENSE_LIST:
                return 'The book is available as a free download thanks to a Creative Commons license.'
            else:
                return 'The book is available as a free download because it is in the Public Domain.'

    @property
    def description(self):
        return self.work.description

class EditionNote(models.Model):
    note = models.CharField(max_length=64, null=True, blank=True, unique=True)
    def __unicode__(self):
        return self.note

class Publisher(models.Model):
    created = models.DateTimeField(auto_now_add=True)
    name = models.ForeignKey('PublisherName', on_delete=models.CASCADE, related_name='key_publisher')
    url = models.URLField(max_length=1024, null=True, blank=True)
    logo_url = models.URLField(max_length=1024, null=True, blank=True)
    description = models.TextField(default='', null=True, blank=True)

    def __unicode__(self):
        return self.name.name

class PublisherName(models.Model):
    name = models.CharField(max_length=255, blank=False, unique=True)

    publisher = models.ForeignKey('Publisher', on_delete=models.CASCADE, related_name='alternate_names', null=True)

    def __unicode__(self):
        return self.name

    def save(self, *args, **kwargs):
        super(PublisherName, self).save(*args, **kwargs) # Call the "real" save() method.
        if self.publisher and self != self.publisher.name:
            #this name is an alias, repoint all editions with this name to the other.
            for edition in Edition.objects.filter(publisher_name=self):
                edition.publisher_name = self.publisher.name
                edition.save()


class WasWork(models.Model):
    work = models.ForeignKey('Work', on_delete=models.CASCADE)
    was = models.IntegerField(unique=True)
    moved = models.DateTimeField(auto_now_add=True)
    user = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, null=True)

def safe_get_work(work_id):
    """
    use this rather than querying the db directly for a work by id
    """
    try:
        work = Work.objects.get(id=work_id)
    except Work.DoesNotExist:
        try:
            work = WasWork.objects.get(was=work_id).work
        except WasWork.DoesNotExist:
            raise Work.DoesNotExist()
    except ValueError:
        #work_id is not a number
        raise Work.DoesNotExist()
    return work

def path_for_file(instance, filename):
    return "ebf/{}.{}".format(uuid.uuid4().get_hex(), instance.format)

class EbookFile(models.Model):
    file = models.FileField(upload_to=path_for_file)
    format = models.CharField(max_length=25, choices=settings.FORMATS)
    edition = models.ForeignKey('Edition', on_delete=models.CASCADE, related_name='ebook_files')
    created = models.DateTimeField(auto_now_add=True)
    asking = models.BooleanField(default=False)
    ebook = models.ForeignKey('Ebook', on_delete=models.CASCADE, related_name='ebook_files', null=True)
    source = models.URLField(null=True, blank=True)
    mobied = models.IntegerField(default=0) #-1 indicates a failed conversion attempt
    version = None
    def check_file(self):
        if self.format == 'epub':
            return test_epub(self.file)
        return None

    @property
    def active(self):
        try:
            return Ebook.objects.filter(url=self.file.url)[0].active
        except:
            return False

    def make_mobi(self):
        if not self.format == 'epub' or not settings.MOBIGEN_URL:
            return False
        if self.mobied < 0:
            return False
        try:
            mobi_cf = ContentFile(mobi.convert_to_mobi(self.file.url))
        except:
            self.mobied = -1
            self.save()
            return False
        new_mobi_ebf = EbookFile.objects.create(
            edition=self.edition,
            format='mobi',
            asking=self.asking,
            source=self.file.url
        )
            
        new_mobi_ebf.file.save(path_for_file(new_mobi_ebf, None), mobi_cf)
        new_mobi_ebf.save()
        if self.ebook:
            new_ebook = Ebook.objects.create(
                edition=self.edition,
                format='mobi',
                provider='Unglue.it',
                url=new_mobi_ebf.file.url,
                rights=self.ebook.rights,
                version_label=self.ebook.version_label,
                version_iter=self.ebook.version_iter,
            )
            new_mobi_ebf.ebook = new_ebook
        new_mobi_ebf.save()
        self.mobied = 1
        self.save()
        return True

send_to_kindle_limit = 7492232

class Ebook(models.Model):
    url = models.URLField(max_length=1024) #change to unique?
    created = models.DateTimeField(auto_now_add=True, db_index=True,)
    format = models.CharField(max_length=25, choices=settings.FORMATS, blank=False)
    provider = models.CharField(max_length=255)
    download_count = models.IntegerField(default=0)
    active = models.BooleanField(default=True)
    filesize = models.PositiveIntegerField(null=True)
    version_label = models.CharField(max_length=255, default="", blank=True)
    version_iter = models.PositiveIntegerField(default=0)

    # use 'PD-US', 'CC BY', 'CC BY-NC-SA', 'CC BY-NC-ND', 'CC BY-NC', 'CC BY-ND', 'CC BY-SA', 'CC0'
    rights = models.CharField(max_length=255, null=True, choices=cc.CHOICES, db_index=True)
    edition = models.ForeignKey('Edition', on_delete=models.CASCADE, related_name='ebooks')
    user = models.ForeignKey(settings.AUTH_USER_MODEL, on_delete=models.CASCADE, null=True)

    def kindle_sendable(self):
        if not self.filesize or self.filesize < send_to_kindle_limit:
            return True
        else:
            return False

    def get_archive(self):  # returns an open file
        ebf = self.get_archive_ebf()
        if not ebf:
            return None
        try:
            ebf.file.open()
        except ValueError:
            logger.error(u'couldn\'t open EbookFile {}'.format(ebf.id))
            return None
        except IOError:
            logger.error(u'EbookFile {} does not exist'.format(ebf.id))
            return None
        return ebf.file

    def get_archive_ebf(self): # returns an ebf
        if not self.ebook_files.filter(asking=False):
            if not self.provider in good_providers:
                return None
            try:
                r = urllib2.urlopen(self.url)
                try:
                    self.filesize = int(r.info().getheaders("Content-Length")[0])
                    if self.save:
                        self.filesize = self.filesize if self.filesize < 2147483647 else 2147483647  # largest safe positive integer
                        self.save()
                    ebf = EbookFile.objects.create(
                        edition=self.edition,
                        ebook=self,
                        format=self.format,
                        source=self.url
                    )
                    ebf.file.save(path_for_file(ebf, None), ContentFile(r.read()))
                    ebf.file.close()
                    ebf.save()
                    return ebf
                except IndexError:
                    # response has no Content-Length header probably a bad link
                    logging.error('Bad link error: {}'.format(self.url))
            except IOError:
                logger.error(u'could not open {}'.format(self.url))
        else:
            ebf = self.ebook_files.filter(asking=False).order_by('-created')[0]
            if not self.filesize:
                self.filesize = ebf.file.size
                self.save()
            return ebf

    def set_provider(self):
        self.provider = Ebook.infer_provider(self.url)
        return self.provider

    @property
    def version(self):
        if self.version_label is None:
            return '.{}'.format(self.version_iter)
        else:
            return '().{}'.format(self.version_label, self.version_iter)
    
    def set_version(self, version):
        #set both version_label and version_iter with one string with format "version.iter"
        version_pattern = r'(.*)\.(\d+)$'
        match = re.match(version_pattern,version)
        if match:
            (self.version_label, self.version_iter) = (match.group(1), match.group(2))
        else:
            self.version_label = version
        self.save()
        
    def set_next_iter(self):
        # set the version iter to the next unused iter for that version
        for ebook in Ebook.objects.filter(
                    edition=self.edition, 
                    version_label=self.version_label,
                    format=self.format,
                    provider=self.provider
                ).order_by('-version_iter'):
             iter = ebook.version_iter
             break
        self.version_iter = iter + 1
        self.save()
          
    @property
    def rights_badge(self):
        if self.rights is None:
            return cc.CCLicense.badge('PD-US')
        return cc.CCLicense.badge(self.rights)

    @staticmethod
    def infer_provider(url):
        if not url:
            return None
        # provider derived from url. returns provider value. remember to call save() afterward
        if re.match(r'https?://books.google.com/', url):
            provider = 'Google Books'
        elif re.match(r'https?://www.gutenberg.org/', url):
            provider = 'Project Gutenberg'
        elif re.match(r'https?://(www\.|)archive.org/', url):
            provider = 'Internet Archive'
        elif url.startswith('http://hdl.handle.net/2027/') or url.startswith('http://babel.hathitrust.org/'):
            provider = 'Hathitrust'
        elif re.match(r'https?://\w\w\.wikisource\.org/', url):
            provider = 'Wikisource'
        elif re.match(r'https?://\w\w\.wikibooks\.org/', url):
            provider = 'Wikibooks'
        elif re.match(r'https://github\.com/[^/ ]+/[^/ ]+/raw/[^ ]+', url):
            provider = 'Github'
        elif re.match(r'https?://www\.oapen\.org/download', url):
            provider = 'OAPEN Library'
        else:
            provider = None
        return provider

    def increment(self):
        Ebook.objects.filter(id=self.id).update(download_count=F('download_count') +1)

    @property
    def download_url(self):
        return settings.BASE_URL_SECURE + reverse('download_ebook', args=[self.id])

    def is_direct(self):
        return self.provider not in ('Google Books', 'Project Gutenberg')

    def __unicode__(self):
        return "%s (%s from %s)" % (self.edition.title, self.format, self.provider)

    def deactivate(self):
        self.active = False
        self.save()

    def activate(self):
        self.active = True
        self.save()

def set_free_flag(sender, instance, created, **kwargs):
    if created:
        if not instance.edition.work.is_free and instance.active:
            instance.edition.work.is_free = True
            instance.edition.work.save()
    elif not instance.active and instance.edition.work.is_free and instance.edition.work.ebooks().count() == 0:
        instance.edition.work.is_free = False
        instance.edition.work.save()
    elif instance.active and not instance.edition.work.is_free and instance.edition.work.ebooks().count() > 0:
        instance.edition.work.is_free = True
        instance.edition.work.save()

post_save.connect(set_free_flag, sender=Ebook)

def reset_free_flag(sender, instance, **kwargs):
    # if the Work associated with the instance Ebook currenly has only 1 Ebook, then it's no longer a free Work
    # once the instance Ebook is deleted.
    if instance.edition.work.ebooks().count() == 1:
        instance.edition.work.is_free = False
        instance.edition.work.save()

pre_delete.connect(reset_free_flag, sender=Ebook)