def _load_solrdata(self):
if self.type == "edition":
return {
'ebook_count': int(bool(self.document.ocaid)),
'edition_count': 1,
'work_count': 1,
'last_update': self.document.last_modified
}
else:
q = self.get_solr_query_term()
if q:
solr = get_solr()
result = solr.select(q, fields=["edition_count", "ebook_count_i"])
last_update_i = [doc['last_update_i'] for doc in result.docs if 'last_update_i' in doc]
if last_update_i:
last_update = self._inttime_to_datetime(last_update_i)
else:
# if last_update is not present in solr, consider last_modfied of
# that document as last_update
if self.type in ['work', 'author']:
last_update = self.document.last_modified
else:
last_update = None
return {
'ebook_count': sum(doc.get('ebook_count_i', 0) for doc in result.docs),
'edition_count': sum(doc.get('edition_count', 0) for doc in result.docs),
'work_count': result.num_found,
'last_update': last_update
}
return {}
def get_solr_works(work_key: Iterable[str]) -> dict[str, dict]:
from openlibrary.plugins.worksearch.search import get_solr
return {
doc['key']: doc
for doc in get_solr().get_many(set(work_key), fields=DEFAULT_SEARCH_FIELDS)
}
def GET(self):
i = web.input(q="", limit=5)
i.limit = safeint(i.limit, 5)
solr = get_solr()
q = solr.escape(i.q).strip()
if is_work_olid(q.upper()):
# ensure uppercase; key is case sensitive in solr
solr_q = 'key:"/works/%s"' % q.upper()
else:
solr_q = 'title:"%s"^2 OR title:(%s*)' % (q, q)
params = {
'q_op': 'AND',
'sort': 'edition_count desc',
'rows': i.limit,
'fq': 'type:work',
# limit the fields returned for better performance
'fl': 'key,title,subtitle,cover_i,first_publish_year,author_name,edition_count'
}
data = solr.select(solr_q, **params)
# exclude fake works that actually have an edition key
docs = [d for d in data['docs'] if d['key'][-1] == 'W']
for d in docs:
# Required by the frontend
d['name'] = d['key'].split('/')[-1]
d['full_title'] = d['title']
if 'subtitle' in d:
d['full_title'] += ": " + d['subtitle']
return to_json(docs)
def _get_solr_data(self):
fields = [
"cover_edition_key", "cover_id", "edition_key",
"first_publish_year", "has_fulltext", "lending_edition_s",
"checked_out", "public_scan_b", "ia"
]
solr = get_solr()
stats.begin("solr", query={"key": self.key}, fields=fields)
try:
d = solr.select({"key": self.key}, fields=fields)
except Exception as e:
logging.getLogger("openlibrary").exception(
"Failed to get solr data")
return None
finally:
stats.end()
if d.num_found > 0:
w = d.docs[0]
else:
w = None
# Replace _solr_data property with the attribute
self.__dict__['_solr_data'] = w
return w
def GET(self):
i = web.input(q="", limit=5)
i.limit = safeint(i.limit, 5)
solr = get_solr()
q = solr.escape(i.q).strip()
solr_q = ''
if is_author_olid(q.upper()):
# ensure uppercase; key is case sensitive in solr
solr_q = 'key:"/authors/%s"' % q.upper()
else:
prefix_q = q + "*"
solr_q = 'name:(%s) OR alternate_names:(%s)' % (prefix_q, prefix_q)
params = {
'q_op': 'AND',
'sort': 'work_count desc',
'rows': i.limit,
'fq': 'type:author'
}
data = solr.select(solr_q, **params)
docs = data['docs']
for d in docs:
if 'top_work' in d:
d['works'] = [d.pop('top_work')]
else:
d['works'] = []
d['subjects'] = d.pop('top_subjects', [])
return to_json(docs)
def find_matches(self, i):
"""
Tries to find an edition, or work, or multiple work candidates that match the given input data.
Case#1: No match. None is returned.
Case#2: Work match but not edition. Work is returned.
Case#3: Work match and edition match. Edition is returned
Case#4: Multiple work match. List of works is returned.
:param web.utils.Storage i: addbook user supplied formdata
:rtype: None or list or Work or Edition
:return: None or Work or Edition or list of Works that are likely matches.
"""
i.publish_year = i.publish_date and self.extract_year(i.publish_date)
author_key = i.authors and i.authors[0].author.key
# work is set from the templates/books/check.html page.
work_key = i.get('work')
# work_key is set to none-of-these when user selects none-of-these link.
if work_key == 'none-of-these':
return None # Case 1, from check page
work = work_key and web.ctx.site.get(work_key)
if work:
edition = self.try_edition_match(work=work,
publisher=i.publisher,
publish_year=i.publish_year,
id_name=i.id_name,
id_value=i.id_value)
return edition or work # Case 3 or 2, from check page
edition = self.try_edition_match(title=i.title,
author_key=author_key,
publisher=i.publisher,
publish_year=i.publish_year,
id_name=i.id_name,
id_value=i.id_value)
if edition:
return edition # Case 2 or 3 or 4, from add page
solr = get_solr()
# Less exact solr search than try_edition_match(), search by supplied title and author only.
result = solr.select(
{
'title': i.title,
'author_key': author_key.split("/")[-1]
},
doc_wrapper=make_work,
q_op="AND")
if result.num_found == 0:
return None # Case 1, from add page
elif result.num_found == 1:
return result.docs[0] # Case 2
else:
return result.docs # Case 4
def _get_edition_keys_from_solr(self, query_terms):
if not query_terms:
return
q = " OR ".join(query_terms)
solr = get_solr()
result = solr.select(q, fields=["edition_key"], rows=10000)
for doc in result['docs']:
if 'edition_key' not in doc:
continue
for k in doc['edition_key']:
yield "/books/" + k
def _get_edition_keys_from_solr(self, query_terms):
if not query_terms:
return
q = " OR ".join(query_terms)
solr = get_solr()
result = solr.select(q, fields=["edition_key"], rows=10000)
for doc in result['docs']:
if 'edition_key' not in doc:
continue
for k in doc['edition_key']:
yield "/books/" + k
def get_solr_query_term(self):
if self.type == 'edition':
return "edition_key:" + self._get_document_basekey()
elif self.type == 'work':
return 'key:/works/' + self._get_document_basekey()
elif self.type == 'author':
return "author_key:" + self._get_document_basekey()
elif self.type == 'subject':
type, value = self.key.split(":", 1)
# escaping value as it can have special chars like : etc.
value = get_solr().escape(value)
return "%s_key:%s" % (type, value)
def get_solr_query_term(self):
if self.type == 'edition':
return "edition_key:" + self._get_document_basekey()
elif self.type == 'work':
return 'key:/works/' + self._get_document_basekey()
elif self.type == 'author':
return "author_key:" + self._get_document_basekey()
elif self.type == 'subject':
type, value = self.key.split(":", 1)
# escaping value as it can have special chars like : etc.
value = get_solr().escape(value)
return "%s_key:%s" % (type, value)
def _get_all_subjects(self):
solr = get_solr()
q = self._get_solr_query_for_subjects()
# Solr has a maxBooleanClauses constraint there too many seeds, the
if len(self.seeds) > 500:
logger.warn(
"More than 500 seeds. skipping solr query for finding subjects."
)
return []
facet_names = [
'subject_facet', 'place_facet', 'person_facet', 'time_facet'
]
try:
result = solr.select(q,
fields=[],
facets=facet_names,
facet_limit=20,
facet_mincount=1)
except OSError:
logger.error("Error in finding subjects of list %s",
self.key,
exc_info=True)
return []
def get_subject_prefix(facet_name):
name = facet_name.replace("_facet", "")
if name == 'subject':
return ''
else:
return name + ":"
def process_subject(facet_name, title, count):
prefix = get_subject_prefix(facet_name)
key = prefix + title.lower().replace(" ", "_")
url = "/subjects/" + key
return web.storage({
"title": title,
"name": title,
"count": count,
"key": key,
"url": url
})
def process_all():
facets = result['facets']
for k in facet_names:
for f in facets.get(k, []):
yield process_subject(f.name, f.value, f.count)
return sorted(process_all(), reverse=True, key=lambda s: s["count"])
def _solr_data(self):
fields = [
"cover_edition_key", "cover_id", "edition_key", "first_publish_year",
"has_fulltext", "lending_edition_s", "public_scan_b", "ia"]
solr = get_solr()
stats.begin("solr", get=self.key, fields=fields)
try:
return solr.get(self.key, fields=fields)
except Exception as e:
logging.getLogger("openlibrary").exception("Failed to get solr data")
return None
finally:
stats.end()
def GET(self):
i = web.input(q="", limit=5)
i.limit = safeint(i.limit, 5)
solr = get_solr()
q = solr.escape(i.q).strip()
query_is_key = is_work_olid(q.upper())
if query_is_key:
# ensure uppercase; key is case sensitive in solr
solr_q = 'key:"/works/%s"' % q.upper()
else:
solr_q = f'title:"{q}"^2 OR title:({q}*)'
params = {
'q_op':
'AND',
'sort':
'edition_count desc',
'rows':
i.limit,
'fq':
'type:work',
# limit the fields returned for better performance
'fl':
'key,title,subtitle,cover_i,first_publish_year,author_name,edition_count',
}
data = solr.select(solr_q, **params)
# exclude fake works that actually have an edition key
docs = [d for d in data['docs'] if d['key'][-1] == 'W']
if query_is_key and not docs:
# Grumble! Work not in solr yet. Create a dummy.
key = '/works/%s' % q.upper()
work = web.ctx.site.get(key)
if work:
docs = [work.as_fake_solr_record()]
for d in docs:
# Required by the frontend
d['name'] = d['key'].split('/')[-1]
d['full_title'] = d['title']
if 'subtitle' in d:
d['full_title'] += ": " + d['subtitle']
return to_json(docs)
def random_ebooks(limit=2000):
solr = search.get_solr()
sort = "edition_count desc"
result = solr.select(
query='has_fulltext:true -public_scan_b:false',
rows=limit,
sort=sort,
fields=[
'has_fulltext',
'key',
'ia',
"title",
"cover_edition_key",
"author_key", "author_name",
])
return [format_work_data(doc) for doc in result.get('docs', []) if doc.get('ia')]
def _get_all_subjects(self):
solr = get_solr()
q = self._get_solr_query_for_subjects()
# Solr has a maxBooleanClauses constraint there too many seeds, the
if len(self.seeds) > 500:
logger.warn("More than 500 seeds. skipping solr query for finding subjects.")
return []
facet_names = ['subject_facet', 'place_facet', 'person_facet', 'time_facet']
try:
result = solr.select(q,
fields=[],
facets=facet_names,
facet_limit=20,
facet_mincount=1)
except IOError:
logger.error("Error in finding subjects of list %s", self.key, exc_info=True)
return []
def get_subject_prefix(facet_name):
name = facet_name.replace("_facet", "")
if name == 'subject':
return ''
else:
return name + ":"
def process_subject(facet_name, title, count):
prefix = get_subject_prefix(facet_name)
key = prefix + title.lower().replace(" ", "_")
url = "/subjects/" + key
return web.storage({
"title": title,
"name": title,
"count": count,
"key": key,
"url": url
})
def process_all():
facets = result['facets']
for k in facet_names:
for f in facets.get(k, []):
yield process_subject(f.name, f.value, f.count)
return sorted(process_all(), reverse=True, key=lambda s: s["count"])
def random_ebooks(limit=2000):
solr = search.get_solr()
sort = "edition_count desc"
result = solr.select(
query='has_fulltext:true -public_scan_b:false',
rows=limit,
sort=sort,
fields=[
'has_fulltext',
'key',
'ia',
"title",
"cover_edition_key",
"author_key", "author_name",
])
return [format_work_data(doc) for doc in result.get('docs', []) if doc.get('ia')]
def GET(self):
from openlibrary.plugins.worksearch.search import get_solr
result = get_solr().select(query='borrowed_b:false', fields=['key', 'lending_edition_s'], limit=100)
def make_doc(d):
# Makes a store doc from solr doc
return {
"_key": "ebooks/books/" + d['lending_edition_s'],
"_rev": None, # Don't worry about consistancy
"type": "ebook",
"book_key": "/books/" + d['lending_edition_s'],
"borrowed": "false"
}
docs = [make_doc(d) for d in result['docs']]
docdict = dict((d['_key'], d) for d in docs)
web.ctx.site.store.update(docdict)
return delegate.RawText("ok\n")
def random_ebooks(limit=2000):
solr = search.get_solr()
sort = "edition_count desc"
result = solr.select(query='has_fulltext:true -public_scan_b:false',
rows=limit,
sort=sort,
fields=[
'has_fulltext',
'key',
'ia',
"title",
"cover_edition_key",
"author_key",
"author_name",
])
def process_doc(doc):
d = {}
key = doc.get('key', '')
# New solr stores the key as /works/OLxxxW
if not key.startswith("/works/"):
key = "/works/" + key
d['url'] = key
d['title'] = doc.get('title', '')
if 'author_key' in doc and 'author_name' in doc:
d['authors'] = [{
"key": key,
"name": name
} for key, name in zip(doc['author_key'], doc['author_name'])]
if 'cover_edition_key' in doc:
d['cover_url'] = h.get_coverstore_url(
) + "/b/olid/%s-M.jpg" % doc['cover_edition_key']
d['read_url'] = "//archive.org/stream/" + doc['ia'][0]
return d
return [
process_doc(doc) for doc in result.get('docs', []) if doc.get('ia')
]
def get_solr_query_term(self):
if self.type == 'subject':
typ, value = self.key.split(":", 1)
# escaping value as it can have special chars like : etc.
value = get_solr().escape(value)
return f"{typ}_key:{value}"
else:
doc_basekey = self.document.key.split("/")[-1]
if self.type == 'edition':
return f"edition_key:{doc_basekey}"
elif self.type == 'work':
return f'key:/works/{doc_basekey}'
elif self.type == 'author':
return f"author_key:{doc_basekey}"
else:
logger.warning(
f"Cannot get solr query term for seed type {self.type}",
extra={'list': self._list.key, 'seed': self.key},
)
return None
def GET(self):
i = web.input(q="", limit=5)
i.limit = safeint(i.limit, 5)
solr = get_solr()
q = solr.escape(i.q).strip()
query_is_key = is_author_olid(q.upper())
if query_is_key:
# ensure uppercase; key is case sensitive in solr
solr_q = 'key:"/authors/%s"' % q.upper()
else:
prefix_q = q + "*"
solr_q = f'name:({prefix_q}) OR alternate_names:({prefix_q})'
params = {
'q_op': 'AND',
'sort': 'work_count desc',
'rows': i.limit,
'fq': 'type:author',
}
data = solr.select(solr_q, **params)
docs = data['docs']
if query_is_key and not docs:
# Grumble! Must be a new author. Fetch from db, and build a "fake" solr resp
key = '/authors/%s' % q.upper()
author = web.ctx.site.get(key)
if author:
docs = [author.as_fake_solr_record()]
for d in docs:
if 'top_work' in d:
d['works'] = [d.pop('top_work')]
else:
d['works'] = []
d['subjects'] = d.pop('top_subjects', [])
return to_json(docs)
def try_edition_match(
self,
work=None,
title=None,
author_key=None,
publisher=None,
publish_year=None,
id_name=None,
id_value=None,
):
"""
Searches solr for potential edition matches.
:param web.Storage work:
:param str title:
:param str author_key: e.g. /author/OL1234A
:param str publisher:
:param str publish_year: yyyy
:param str id_name: from list of values in mapping below
:param str id_value:
:rtype: None or Edition or list
:return: None, an Edition, or a list of Works
"""
# insufficient data
if not publisher and not publish_year and not id_value:
return
q = {}
work and q.setdefault('key', work.key.split("/")[-1])
title and q.setdefault('title', title)
author_key and q.setdefault('author_key', author_key.split('/')[-1])
publisher and q.setdefault('publisher', publisher)
# There are some errors indexing of publish_year. Use publish_date until it is fixed
publish_year and q.setdefault('publish_date', publish_year)
mapping = {
'isbn_10': 'isbn',
'isbn_13': 'isbn',
'lccn': 'lccn',
'oclc_numbers': 'oclc',
'ocaid': 'ia',
}
if id_value and id_name in mapping:
if id_name.startswith('isbn'):
id_value = id_value.replace('-', '')
q[mapping[id_name]] = id_value
solr = get_solr()
result = solr.select(q, doc_wrapper=make_work, q_op="AND")
if len(result.docs) > 1:
# found multiple work matches
return result.docs
elif len(result.docs) == 1:
# found one work match
work = result.docs[0]
publisher = publisher and fuzzy_find(
publisher,
work.publisher,
stopwords=("publisher", "publishers", "and"))
editions = web.ctx.site.get_many(
["/books/" + key for key in work.edition_key])
for e in editions:
d = {}
if publisher:
if not e.publishers or e.publishers[0] != publisher:
continue
if publish_year:
if not e.publish_date or publish_year != self.extract_year(
e.publish_date):
continue
if id_value and id_name in mapping:
if not id_name in e or id_value not in e[id_name]:
continue
# return the first good likely matching Edition
return e
