def POST(self):
i = web.input(key=[], master=None, merge_key=[])
keys = uniq(i.key)
selected = uniq(i.merge_key)
# filter bad keys
keys = self.filter_authors(keys)
# doesn't make sense to merge master with it self.
if i.master in selected:
selected.remove(i.master)
formdata = web.storage(master=i.master, selected=selected)
if not i.master or len(selected) == 0:
return render_template(
"merge/authors",
keys,
top_books_from_author=top_books_from_author,
formdata=formdata,
)
else:
# redirect to the master. The master will display a progressbar and call the merge_authors_json to trigger the merge.
raise web.seeother("/authors/" + i.master + "/-/" +
"?merge=true&duplicates=" + ",".join(selected))
def merge_property(self, a, b):
if isinstance(a, list) and isinstance(b, list):
return uniq(a + b, key=dicthash)
elif not a:
return b
else:
return a
def GET(self):
i = web.input(key=[])
keys = uniq(i.key)
# filter bad keys
keys = self.filter_authors(keys)
return render_template('merge/authors', keys, top_books_from_author=top_books_from_author)
def get_ocaid(item):
# Circular import otherwise
from ..book_providers import is_non_ia_ocaid
possible_fields = [
'ocaid', # In editions
'identifier', # In ?? not editions/works/solr
'ia', # In solr work records and worksearch get_docs
'lending_identifier', # In solr works records + worksearch get_doc
]
# SOLR WORK RECORDS ONLY:
# Open Library only has access to a list of archive.org IDs
# and solr isn't currently equipped with the information
# necessary to determine which editions may be openly
# available. Using public domain date as a heuristic
# Long term solution is a full reindex, but this hack will work in the
# vast majority of cases for now.
# NOTE: there is still a risk pre-1923 books will get a print-diabled-only
# or lendable edition.
# Note: guaranteed to be int-able if none None
US_PD_YEAR = 1923
if float(item.get('first_publish_year') or '-inf') > US_PD_YEAR:
# Prefer `lending_identifier` over `ia` (push `ia` to bottom)
possible_fields.remove('ia')
possible_fields.append('ia')
ocaids = []
for field in possible_fields:
if item.get(field):
ocaids += item[field] if isinstance(item[field],
list) else [item[field]]
ocaids = uniq(ocaids)
return next((ocaid for ocaid in ocaids if not is_non_ia_ocaid(ocaid)),
None)
def merge_property(self, a, b):
if isinstance(a, list) and isinstance(b, list):
return uniq(a + b, key=dicthash)
elif not a:
return b
else:
return a
def GET(self):
i = web.input(key=[])
keys = uniq(i.key)
# filter bad keys
keys = self.filter_authors(keys)
return render_template('merge/authors', keys, top_books_from_author=top_books_from_author)
def merge_docs(self, master, dup):
# avoid merging other types.
if dup['type']['key'] == '/type/author':
master = BasicMergeEngine.merge_docs(self, master, dup)
if dup.get('name') and not name_eq(dup['name'], master.get('name') or ''):
master.setdefault('alternate_names', []).append(dup['name'])
if 'alternate_names' in master:
master['alternate_names'] = uniq(master['alternate_names'], key=space_squash_and_strip)
return master
def merge_docs(self, master, dup):
# avoid merging other types.
if dup['type']['key'] == '/type/author':
master = BasicMergeEngine.merge_docs(self, master, dup)
if dup.get('name') and not name_eq(dup['name'], master.get('name') or ''):
master.setdefault('alternate_names', []).append(dup['name'])
if 'alternate_names' in master:
master['alternate_names'] = uniq(master['alternate_names'], key=space_squash_and_strip)
return master
def POST(self):
i = web.input(key=[], master=None, merge_key=[])
keys = uniq(i.key)
selected = uniq(i.merge_key)
# filter bad keys
keys = self.filter_authors(keys)
# doesn't make sense to merge master with it self.
if i.master in selected:
selected.remove(i.master)
formdata = web.storage(
master=i.master,
selected=selected
)
if not i.master or len(selected) == 0:
return render_template("merge/authors", keys, top_books_from_author=top_books_from_author, formdata=formdata)
else:
# redirect to the master. The master will display a progressbar and call the merge_authors_json to trigger the merge.
master = web.ctx.site.get("/authors/" + i.master)
raise web.seeother(master.url() + "?merge=true&duplicates=" + ",".join(selected))
def convert_doc(self, doc, master, duplicates):
"""Converts references to any of the duplicates in the given doc to the master.
"""
if isinstance(doc, dict):
if len(doc) == 1 and doc.keys() == ['key']:
key = doc['key']
if key in duplicates:
return {"key": master}
else:
return doc
else:
return dict((k, self.convert_doc(v, master, duplicates)) for k, v in doc.iteritems())
elif isinstance(doc, list):
values = [self.convert_doc(v, master, duplicates) for v in doc]
return uniq(values, key=dicthash)
else:
return doc
def convert_doc(self, doc, master, duplicates):
"""Converts references to any of the duplicates in the given doc to the master.
"""
if isinstance(doc, dict):
if len(doc) == 1 and doc.keys() == ['key']:
key = doc['key']
if key in duplicates:
return {"key": master}
else:
return doc
else:
return dict((k, self.convert_doc(v, master, duplicates))
for k, v in doc.iteritems())
elif isinstance(doc, list):
values = [self.convert_doc(v, master, duplicates) for v in doc]
return uniq(values, key=dicthash)
else:
return doc
def update_references(self, doc, master, duplicates):
"""
Converts references to any of the duplicates in the given doc to the master.
:param doc:
:param str master:
:param list of str duplicates:
:rtype: Any
"""
if isinstance(doc, dict):
if list(doc) == ['key']:
return {"key": master} if doc['key'] in duplicates else doc
else:
return dict(
(k, self.update_references(v, master, duplicates))
for k, v in doc.items())
elif isinstance(doc, list):
values = [self.update_references(v, master, duplicates) for v in doc]
return uniq(values, key=dicthash)
else:
return doc
def GET(self):
i = web.input(key=[],merge_key=[])
keys = uniq(i.key)
merge_keys = uniq(i.merge_key)
assert all(k is not None for k in merge_keys)
if not merge_keys:
return render_template('merge/editions', keys)
full_keys = ['/books/' + k for k in merge_keys]
editions = [web.ctx.site.get('/books/' + k) for k in merge_keys]
master = None
for e in editions:
if e.key == '/books/' + i.master:
master = e
break
all_keys = set()
for e in editions:
for k in e.keys():
if e[k] is not None and e[k] != {}:
all_keys.add(k)
merged = {}
possible_values = defaultdict(lambda: defaultdict(int))
k = 'publish_date'
publish_dates = set(e[k] for e in editions if k in e and len(e[k]) != 4)
k = 'pagination'
all_pagination = set(e[k] for e in editions if e.get(k))
one_item_lists = {}
for k in 'lc_classifications', 'publishers', 'contributions', 'series':
one_item_lists[k] = set(e[k][0].strip('.') for e in editions if e.get(k) and len(set(e[k])) == 1)
for k in ['other_titles', 'isbn_10', 'series']:
if k not in all_keys:
continue
merged[k] = []
for e in editions:
for v in e.get(k, []):
if v not in merged[k]:
possible_values[k][v] += 1
merged[k].append(v)
k = 'ocaid'
for e in editions:
v = e.get(k)
if not v:
continue
possible_values[k][v] += 1
if 'ia:' + v not in merged.get('source_records', []):
merged.setdefault('source_records', []).append(v)
k = 'identifiers'
if k in all_keys:
merged[k] = {}
for e in editions:
if k not in e:
continue
for a, b in e[k].items():
for c in b:
if c in merged[k].setdefault(a, []):
continue
merged[k][a].append(c)
any_publish_country = False
k = 'publish_country'
if k in all_keys:
for e in editions:
if e.get(k) and not e[k].strip().startswith('xx'):
any_publish_country = True
for k in 'source_records', 'ia_box_id':
merged[k] = []
for e in editions:
if e.get(k) and isinstance(e[k], six.string_types):
e[k] = [e[k]]
if e.get(k):
assert isinstance(e[k], list)
for sr in e.get(k, []):
if sr not in merged[k]:
merged[k].append(sr)
for k in all_keys:
if k in ('source_records', 'ia_box_id', 'identifiers', 'ocaid', 'other_titles', 'series'):
continue
uniq_values = defaultdict(list)
for num, e in enumerate(editions):
v = e.get(k)
if v:
if isinstance(v, list):
for lv in v:
possible_values[k][lv] += 1
elif not isinstance(v, dict):
possible_values[k][v] += 1
if k == 'publish_date' and len(v) == 4 and v.isdigit and any(v in pd for pd in publish_dates):
continue
if k == 'pagination' and any(len(i) > len(v) and v in i for i in all_pagination):
continue
if k in one_item_lists and len(set(e.get(k, []))) == 1 and any(len(i) > len(v[0].strip('.')) and v[0].strip('.') in i for i in one_item_lists[k]):
continue
if k == 'publish_country' and any_publish_country and e.get(k, '').strip().startswith('xx'):
continue
if k == 'edition_name' and v.endswith(' ed edition'):
v = v[:-len(' edition')]
uniq_values[re_nonword.sub('', repr(v).lower())].append(num)
if len(uniq_values) == 1:
merged[k] = editions[uniq_values.values()[0][0]][k]
continue
if k == 'covers':
assert all(isinstance(e[k], list) for e in editions if k in e)
covers = set()
for e in editions:
if k in e:
covers.update(c for c in e[k] if c != -1)
merged['covers'] = sorted(covers)
continue
if k == 'notes':
merged['notes'] = ''
for e in editions:
if e.get('notes'):
merged['notes'] += e['notes'] + '\n'
continue
if k == 'ocaid':
for e in editions:
if e.get('ocaid'):
#assert not e['ocaid'].endswith('goog')
merged['ocaid'] = e['ocaid']
break
assert merged['ocaid']
continue
return render_template('merge/editions2', master, editions, all_keys, merged, possible_values)
def load(rec, account_key=None):
"""Given a record, tries to add/match that edition in the system.
Record is a dictionary containing all the metadata of the edition.
The following fields are mandatory:
* title: str
* source_records: list
:param dict rec: Edition record to add
:rtype: dict
:return: a dict to be converted into a JSON HTTP response, same as load_data()
"""
required_fields = ['title', 'source_records'
] # ['authors', 'publishers', 'publish_date']
for field in required_fields:
if not rec.get(field):
raise RequiredField(field)
if not isinstance(rec['source_records'], list):
rec['source_records'] = [rec['source_records']]
# Split subtitle if required and not already present
if ':' in rec.get('title') and not rec.get('subtitle'):
title, subtitle = split_subtitle(rec.get('title'))
if subtitle:
rec['title'] = title
rec['subtitle'] = subtitle
rec = normalize_record_isbns(rec)
edition_pool = build_pool(rec)
# deduplicate authors
rec['authors'] = uniq(rec.get('authors', []), dicthash)
if not edition_pool:
# No match candidates found, add edition
return load_data(rec, account_key=account_key)
match = early_exit(rec)
if not match:
match = find_exact_match(rec, edition_pool)
if not match:
rec['full_title'] = rec['title']
if rec.get('subtitle'):
rec['full_title'] += ' ' + rec['subtitle']
e1 = build_marc(rec)
add_db_name(e1)
match = find_match(e1, edition_pool)
if not match:
# No match found, add edition
return load_data(rec, account_key=account_key)
# We have an edition match at this point
need_work_save = need_edition_save = False
w = None
e = web.ctx.site.get(match)
# check for, and resolve, author redirects
for a in e.authors:
while is_redirect(a):
if a in e.authors:
e.authors.remove(a)
a = web.ctx.site.get(a.location)
if not is_redirect(a):
e.authors.append(a)
if e.get('works'):
w = e.works[0].dict()
work_created = False
else:
# Found an edition without a work
work_created = need_work_save = need_edition_save = True
w = new_work(e.dict(), rec)
e.works = [{'key': w['key']}]
# Add subjects to work, if not already present
if 'subjects' in rec:
work_subjects = list(w.get('subjects', []))
for s in rec['subjects']:
if s not in work_subjects:
work_subjects.append(s)
need_work_save = True
if need_work_save and work_subjects:
w['subjects'] = work_subjects
# Add cover to edition
if 'cover' in rec and not e.get_covers():
cover_url = rec['cover']
cover_id = add_cover(cover_url, e.key, account_key=account_key)
if cover_id:
e['covers'] = [cover_id]
need_edition_save = True
# Add cover to work, if needed
if not w.get('covers') and e.get_covers():
w['covers'] = [e['covers'][0]]
need_work_save = True
# Add description to work, if needed
if not w.get('description') and e.get('description'):
w['description'] = e['description']
need_work_save = True
# Add authors to work, if needed
if not w.get('authors'):
authors = [import_author(a) for a in rec.get('authors', [])]
w['authors'] = [{
'type': {
'key': '/type/author_role'
},
'author': a.key
} for a in authors if a.get('key')]
if w.get('authors'):
need_work_save = True
# Add ocaid to edition (str), if needed
if 'ocaid' in rec and not e.ocaid:
e['ocaid'] = rec['ocaid']
need_edition_save = True
# Add list fields to edition as needed
edition_fields = [
'local_id',
'lccn',
'lc_classifications',
'source_records',
]
for f in edition_fields:
if f not in rec:
continue
# ensure values is a list
values = rec[f] if isinstance(rec[f], list) else [rec[f]]
if f in e:
# get values from rec that are not currently on the edition
to_add = [v for v in values if v not in e[f]]
e[f] += to_add
else:
e[f] = to_add = values
if to_add:
need_edition_save = True
edits = []
reply = {
'success': True,
'edition': {
'key': match,
'status': 'matched'
},
'work': {
'key': w['key'],
'status': 'matched'
},
}
if need_edition_save:
reply['edition']['status'] = 'modified'
edits.append(e.dict())
if need_work_save:
reply['work']['status'] = 'created' if work_created else 'modified'
edits.append(w)
if edits:
web.ctx.site.save_many(edits,
comment='import existing book',
action='edit-book')
if 'ocaid' in rec:
update_ia_metadata_for_ol_edition(match.split('/')[-1])
return reply
