def guess_collection_of_a_record(recID, referer=None, recreate_cache_if_needed=True):
"""Return collection name a record recid belongs to, by first testing
the referer URL if provided and otherwise returning the
primary collection."""
if referer:
dummy, hostname, path, dummy, query, dummy = urlparse.urlparse(referer)
# requests can come from different invenio installations, with
# different collections
if CFG_SITE_URL.find(hostname) < 0:
return guess_primary_collection_of_a_record(recID)
g = _re_collection_url.match(path)
if g:
name = urllib.unquote_plus(g.group(1))
# check if this collection actually exist (also normalize the name
# if case-insensitive)
name = Collection.query.filter_by(name=name).value('name')
if name and recID in get_collection_reclist(name):
return name
elif path.startswith('/search'):
if recreate_cache_if_needed:
collection_reclist_cache.recreate_cache_if_needed()
query = cgi.parse_qs(query)
for name in query.get('cc', []) + query.get('c', []):
name = Collection.query.filter_by(name=name).value('name')
if name and recID in get_collection_reclist(name, recreate_cache_if_needed=False):
return name
return guess_primary_collection_of_a_record(recID)
def guess_collection_of_a_record(recID, referer=None, recreate_cache_if_needed=True):
"""Return collection name a record recid belongs to, by first testing
the referer URL if provided and otherwise returning the
primary collection."""
if referer:
dummy, hostname, path, dummy, query, dummy = urlparse.urlparse(referer)
# requests can come from different invenio installations, with
# different collections
if CFG_SITE_URL.find(hostname) < 0:
return guess_primary_collection_of_a_record(recID)
g = _re_collection_url.match(path)
if g:
name = urllib.unquote_plus(g.group(1))
# check if this collection actually exist (also normalize the name
# if case-insensitive)
name = Collection.query.filter_by(name=name).value('name')
if name and recID in get_collection_reclist(name):
return name
elif path.startswith('/search'):
if recreate_cache_if_needed:
collection_reclist_cache.recreate_cache_if_needed()
query = cgi.parse_qs(query)
for name in query.get('cc', []) + query.get('c', []):
name = Collection.query.filter_by(name=name).value('name')
if name and recID in get_collection_reclist(name, recreate_cache_if_needed=False):
return name
return guess_primary_collection_of_a_record(recID)
def get_records_that_can_be_displayed(permitted_restricted_collections,
hitset_in_any_collection,
current_coll=None,
colls=None):
"""Return records that can be displayed."""
current_coll = current_coll or cfg['CFG_SITE_NAME']
records_that_can_be_displayed = intbitset()
if colls is None:
colls = [current_coll]
policy = cfg['CFG_WEBSEARCH_VIEWRESTRCOLL_POLICY'].strip().upper()
# real & virtual
current_coll_children = get_collection_allchildren(current_coll)
# Add all restricted collections, that the user has access to, and are
# under the current collection do not use set here, in order to maintain a
# specific order: children of 'cc' (real, virtual, restricted), rest of 'c'
# that are not cc's children
colls_to_be_displayed = set([
coll for coll in current_coll_children
if coll in colls or coll in permitted_restricted_collections
])
colls_to_be_displayed |= set(
[coll for coll in colls if coll not in colls_to_be_displayed])
# Get all records in applicable collections
records_that_can_be_displayed = intbitset()
for coll in colls_to_be_displayed:
records_that_can_be_displayed |= get_collection_reclist(coll)
if policy == 'ANY':
# User needs to have access to at least one collection that restricts
# the records. We need this to be able to remove records that are both
# in a public and restricted collection.
permitted_recids = intbitset()
notpermitted_recids = intbitset()
for collection in restricted_collection_cache.cache:
if collection in permitted_restricted_collections:
permitted_recids |= get_collection_reclist(collection)
else:
notpermitted_recids |= get_collection_reclist(collection)
notpermitted_recids -= permitted_recids
else:
# User needs to have access to all collections that restrict a records.
notpermitted_recids = intbitset()
for collection in restricted_collection_cache.cache:
if collection not in permitted_restricted_collections:
notpermitted_recids |= get_collection_reclist(collection)
# Remove records that can not be seen by user
records_that_can_be_displayed -= notpermitted_recids
# Intersect only if there are some matched records
if not hitset_in_any_collection.is_infinite():
records_that_can_be_displayed &= hitset_in_any_collection
return records_that_can_be_displayed
def get_records_that_can_be_displayed(permitted_restricted_collections,
hitset_in_any_collection,
current_coll=None, colls=None):
"""Return records that can be displayed."""
current_coll = current_coll or cfg['CFG_SITE_NAME']
records_that_can_be_displayed = intbitset()
if colls is None:
colls = [current_coll]
policy = cfg['CFG_WEBSEARCH_VIEWRESTRCOLL_POLICY'].strip().upper()
# real & virtual
current_coll_children = get_collection_allchildren(current_coll)
# Add all restricted collections, that the user has access to, and are
# under the current collection do not use set here, in order to maintain a
# specific order: children of 'cc' (real, virtual, restricted), rest of 'c'
# that are not cc's children
colls_to_be_displayed = set([
coll for coll in current_coll_children
if coll in colls or coll in permitted_restricted_collections
])
colls_to_be_displayed |= set([coll for coll in colls
if coll not in colls_to_be_displayed])
# Get all records in applicable collections
records_that_can_be_displayed = intbitset()
for coll in colls_to_be_displayed:
records_that_can_be_displayed |= get_collection_reclist(coll)
if policy == 'ANY':
# User needs to have access to at least one collection that restricts
# the records. We need this to be able to remove records that are both
# in a public and restricted collection.
permitted_recids = intbitset()
notpermitted_recids = intbitset()
for collection in restricted_collection_cache.cache:
if collection in permitted_restricted_collections:
permitted_recids |= get_collection_reclist(collection)
else:
notpermitted_recids |= get_collection_reclist(collection)
notpermitted_recids -= permitted_recids
else:
# User needs to have access to all collections that restrict a records.
notpermitted_recids = intbitset()
for collection in restricted_collection_cache.cache:
if collection not in permitted_restricted_collections:
notpermitted_recids |= get_collection_reclist(collection)
# Remove records that can not be seen by user
records_that_can_be_displayed -= notpermitted_recids
# Intersect only if there are some matched records
if not hitset_in_any_collection.is_infinite():
records_that_can_be_displayed &= hitset_in_any_collection
return records_that_can_be_displayed
def browse_pattern_phrases(req, colls, p, f, rg, ln=CFG_SITE_LANG):
"""Returns either biliographic phrases or words indexes."""
## is p enclosed in quotes? (coming from exact search)
if p.startswith('"') and p.endswith('"'):
p = p[1:-1]
## okay, "real browse" follows:
## FIXME: the maths in the get_nearest_terms_in_bibxxx is just a test
if not f and p.find(":") > 0: # does 'p' contain ':'?
f, p = p.split(":", 1)
coll_hitset = intbitset()
for coll_name in colls:
coll_hitset |= get_collection_reclist(coll_name)
index_id = get_index_id_from_field(f)
if index_id != 0:
browsed_phrases_in_colls = get_nearest_terms_in_idxphrase_with_collection(
p, index_id, rg / 2, rg / 2, coll_hitset)
else:
browsed_phrases = get_nearest_terms_in_bibxxx(p, f, (rg + 1) / 2 + 1,
(rg - 1) / 2 + 1)
while not browsed_phrases:
# try again and again with shorter and shorter pattern:
try:
p = p[:-1]
browsed_phrases = get_nearest_terms_in_bibxxx(
p, f, (rg + 1) / 2 + 1, (rg - 1) / 2 + 1)
except:
register_exception(req=req, alert_admin=True)
# probably there are no hits at all:
return []
## try to check hits in these particular collection selection:
browsed_phrases_in_colls = []
if 0:
for phrase in browsed_phrases:
phrase_hitset = intbitset()
phrase_hitsets = search_pattern("", phrase, f, 'e')
for coll in colls:
phrase_hitset.union_update(phrase_hitsets[coll])
if len(phrase_hitset) > 0:
# okay, this phrase has some hits in colls, so add it:
browsed_phrases_in_colls.append(
[phrase, len(phrase_hitset)])
## were there hits in collections?
if browsed_phrases_in_colls == []:
if browsed_phrases != []:
#write_warning(req, """<p>No match close to <em>%s</em> found in given collections.
#Please try different term.<p>Displaying matches in any collection...""" % p_orig)
## try to get nbhits for these phrases in any collection:
for phrase in browsed_phrases:
nbhits = get_nbhits_in_bibxxx(phrase, f, coll_hitset)
if nbhits > 0:
browsed_phrases_in_colls.append([phrase, nbhits])
return browsed_phrases_in_colls
def guess_primary_collection_of_a_record(recID):
"""Return primary collection name a record recid belongs to, by
testing 980 identifier.
May lead to bad guesses when a collection is defined dynamically
via dbquery.
In that case, return 'CFG_SITE_NAME'."""
out = CFG_SITE_NAME
dbcollids = get_fieldvalues(recID, "980__a")
for dbcollid in dbcollids:
variants = ("collection:" + dbcollid, 'collection:"' + dbcollid + '"',
"980__a:" + dbcollid, '980__a:"' + dbcollid + '"',
'980:' + dbcollid, '980:"' + dbcollid + '"')
res = run_sql(
"SELECT name FROM collection WHERE dbquery IN (%s,%s,%s,%s,%s,%s)",
variants)
if res:
out = res[0][0]
break
if CFG_CERN_SITE:
recID = int(recID)
# dirty hack for ATLAS collections at CERN:
if out in ('ATLAS Communications', 'ATLAS Internal Notes'):
for alternative_collection in (
'ATLAS Communications Physics',
'ATLAS Communications General',
'ATLAS Internal Notes Physics',
'ATLAS Internal Notes General',
):
if recID in get_collection_reclist(alternative_collection):
return alternative_collection
# dirty hack for FP
FP_collections = {
'DO': ['Current Price Enquiries', 'Archived Price Enquiries'],
'IT': [
'Current Invitation for Tenders',
'Archived Invitation for Tenders'
],
'MS': ['Current Market Surveys', 'Archived Market Surveys']
}
fp_coll_ids = [coll for coll in dbcollids if coll in FP_collections]
for coll in fp_coll_ids:
for coll_name in FP_collections[coll]:
if recID in get_collection_reclist(coll_name):
return coll_name
return out
def guess_primary_collection_of_a_record(recID):
"""Return primary collection name a record recid belongs to, by
testing 980 identifier.
May lead to bad guesses when a collection is defined dynamically
via dbquery.
In that case, return 'CFG_SITE_NAME'."""
out = CFG_SITE_NAME
dbcollids = get_fieldvalues(recID, "980__a")
for dbcollid in dbcollids:
variants = (
"collection:" + dbcollid,
'collection:"' + dbcollid + '"',
"980__a:" + dbcollid,
'980__a:"' + dbcollid + '"',
"980:" + dbcollid,
'980:"' + dbcollid + '"',
)
res = run_sql("SELECT name FROM collection WHERE dbquery IN (%s,%s,%s,%s,%s,%s)", variants)
if res:
out = res[0][0]
break
if CFG_CERN_SITE:
recID = int(recID)
# dirty hack for ATLAS collections at CERN:
if out in ("ATLAS Communications", "ATLAS Internal Notes"):
for alternative_collection in (
"ATLAS Communications Physics",
"ATLAS Communications General",
"ATLAS Internal Notes Physics",
"ATLAS Internal Notes General",
):
if recID in get_collection_reclist(alternative_collection):
return alternative_collection
# dirty hack for FP
FP_collections = {
"DO": ["Current Price Enquiries", "Archived Price Enquiries"],
"IT": ["Current Invitation for Tenders", "Archived Invitation for Tenders"],
"MS": ["Current Market Surveys", "Archived Market Surveys"],
}
fp_coll_ids = [coll for coll in dbcollids if coll in FP_collections]
for coll in fp_coll_ids:
for coll_name in FP_collections[coll]:
if recID in get_collection_reclist(coll_name):
return coll_name
return out
def get_restricted_collections_for_recid(recid, recreate_cache_if_needed=True):
"""Return the list of restricted collections to which recid belongs."""
if recreate_cache_if_needed:
restricted_collection_cache.recreate_cache_if_needed()
collection_reclist_cache.recreate_cache_if_needed()
return [collection for collection in restricted_collection_cache.cache
if recid in get_collection_reclist(
collection, recreate_cache_if_needed=False)]
def browse_pattern_phrases(req, colls, p, f, rg, ln=CFG_SITE_LANG):
"""Returns either biliographic phrases or words indexes."""
## is p enclosed in quotes? (coming from exact search)
if p.startswith('"') and p.endswith('"'):
p = p[1:-1]
## okay, "real browse" follows:
## FIXME: the maths in the get_nearest_terms_in_bibxxx is just a test
if not f and p.find(":") > 0: # does 'p' contain ':'?
f, p = p.split(":", 1)
coll_hitset = intbitset()
for coll_name in colls:
coll_hitset |= get_collection_reclist(coll_name)
index_id = get_index_id_from_field(f)
if index_id != 0:
browsed_phrases_in_colls = get_nearest_terms_in_idxphrase_with_collection(
p, index_id, rg / 2, rg / 2, coll_hitset
)
else:
browsed_phrases = get_nearest_terms_in_bibxxx(p, f, (rg + 1) / 2 + 1, (rg - 1) / 2 + 1)
while not browsed_phrases:
# try again and again with shorter and shorter pattern:
try:
p = p[:-1]
browsed_phrases = get_nearest_terms_in_bibxxx(p, f, (rg + 1) / 2 + 1, (rg - 1) / 2 + 1)
except:
register_exception(req=req, alert_admin=True)
# probably there are no hits at all:
return []
## try to check hits in these particular collection selection:
browsed_phrases_in_colls = []
if 0:
for phrase in browsed_phrases:
phrase_hitset = intbitset()
phrase_hitsets = search_pattern("", phrase, f, "e")
for coll in colls:
phrase_hitset.union_update(phrase_hitsets[coll])
if len(phrase_hitset) > 0:
# okay, this phrase has some hits in colls, so add it:
browsed_phrases_in_colls.append([phrase, len(phrase_hitset)])
## were there hits in collections?
if browsed_phrases_in_colls == []:
if browsed_phrases != []:
# write_warning(req, """<p>No match close to <em>%s</em> found in given collections.
# Please try different term.<p>Displaying matches in any collection...""" % p_orig)
## try to get nbhits for these phrases in any collection:
for phrase in browsed_phrases:
nbhits = get_nbhits_in_bibxxx(phrase, f, coll_hitset)
if nbhits > 0:
browsed_phrases_in_colls.append([phrase, nbhits])
return browsed_phrases_in_colls
def get_restricted_collections_for_recid(recid, recreate_cache_if_needed=True):
"""Return the list of restricted collections to which recid belongs."""
if recreate_cache_if_needed:
restricted_collection_cache.recreate_cache_if_needed()
collection_reclist_cache.recreate_cache_if_needed()
return [
collection for collection in restricted_collection_cache.cache
if recid in get_collection_reclist(collection,
recreate_cache_if_needed=False)
]
def get_all_collections_of_a_record(recID, recreate_cache_if_needed=True):
"""Return all the collection names a record belongs to.
Note this function is O(n_collections)."""
ret = []
if recreate_cache_if_needed:
collection_reclist_cache.recreate_cache_if_needed()
for name in collection_reclist_cache.cache.keys():
if recID in get_collection_reclist(name, recreate_cache_if_needed=False):
ret.append(name)
return ret
def get_all_collections_of_a_record(recID, recreate_cache_if_needed=True):
"""Return all the collection names a record belongs to.
Note this function is O(n_collections)."""
ret = []
if recreate_cache_if_needed:
collection_reclist_cache.recreate_cache_if_needed()
for name in collection_reclist_cache.cache.keys():
if recID in get_collection_reclist(name, recreate_cache_if_needed=False):
ret.append(name)
return ret
def get_facets_for_query(self, qid, limit=20, parent=None):
"""Return record ids as intbitset."""
recIDsHitSet = self.get_recids_intbitset(qid)
parent = request.args.get('parent', None)
if parent is not None:
collection = Collection.query.filter(
Collection.name == parent).first_or_404()
else:
cc = search_results_cache.get(
get_search_results_cache_key_from_qid(qid) + '::cc')
if cc is not None:
collection = Collection.query.filter(
Collection.name == cc).first_or_404()
else:
collection = Collection.query.get(1)
facet = []
for c in collection.collection_children_r:
num_records = len(get_collection_reclist(
c.name, recreate_cache_if_needed=False
).intersection(recIDsHitSet))
if num_records:
facet.append((c.name, num_records, c.name_ln))
return sorted(facet, key=lambda x: x[1], reverse=True)[0:limit]
def __call__(self, req, form):
""" Perform a search. """
argd = wash_search_urlargd(form)
_ = gettext_set_language(argd['ln'])
if req.method == 'POST':
raise apache.SERVER_RETURN, apache.HTTP_METHOD_NOT_ALLOWED
uid = getUid(req)
user_info = collect_user_info(req)
if uid == -1:
return page_not_authorized(req, "../",
text=_("You are not authorized to view this area."),
navmenuid='search')
elif uid > 0:
pref = get_user_preferences(uid)
try:
if 'rg' not in form:
# fetch user rg preference only if not overridden via URL
argd['rg'] = int(pref['websearch_group_records'])
except (KeyError, ValueError):
pass
if argd['rg'] > CFG_WEBSEARCH_MAX_RECORDS_IN_GROUPS and acc_authorize_action(req, 'runbibedit')[0] != 0:
argd['rg'] = CFG_WEBSEARCH_MAX_RECORDS_IN_GROUPS
involved_collections = set()
involved_collections.update(argd['c'])
involved_collections.add(argd['cc'])
if argd['id'] > 0:
argd['recid'] = argd['id']
if argd['idb'] > 0:
argd['recidb'] = argd['idb']
if argd['sysno']:
tmp_recid = find_record_from_sysno(argd['sysno'])
if tmp_recid:
argd['recid'] = tmp_recid
if argd['sysnb']:
tmp_recid = find_record_from_sysno(argd['sysnb'])
if tmp_recid:
argd['recidb'] = tmp_recid
if argd['recid'] > 0:
if argd['recidb'] > argd['recid']:
# Hack to check if among the restricted collections
# at least a record of the range is there and
# then if the user is not authorized for that
# collection.
recids = intbitset(xrange(argd['recid'], argd['recidb']))
restricted_collection_cache.recreate_cache_if_needed()
for collname in restricted_collection_cache.cache:
(auth_code, auth_msg) = acc_authorize_action(user_info, VIEWRESTRCOLL, collection=collname)
if auth_code and user_info['email'] == 'guest':
coll_recids = get_collection_reclist(collname)
if coll_recids & recids:
cookie = mail_cookie_create_authorize_action(VIEWRESTRCOLL, {'collection' : collname})
target = CFG_SITE_SECURE_URL + '/youraccount/login' + \
make_canonical_urlargd({'action': cookie, 'ln' : argd['ln'], 'referer' : CFG_SITE_SECURE_URL + req.unparsed_uri}, {})
return redirect_to_url(req, target, norobot=True)
elif auth_code:
return page_not_authorized(req, "../", \
text=auth_msg, \
navmenuid='search')
else:
involved_collections.add(guess_primary_collection_of_a_record(argd['recid']))
# If any of the collection requires authentication, redirect
# to the authentication form.
for coll in involved_collections:
if collection_restricted_p(coll):
(auth_code, auth_msg) = acc_authorize_action(user_info, VIEWRESTRCOLL, collection=coll)
if auth_code and user_info['email'] == 'guest':
cookie = mail_cookie_create_authorize_action(VIEWRESTRCOLL, {'collection' : coll})
target = CFG_SITE_SECURE_URL + '/youraccount/login' + \
make_canonical_urlargd({'action': cookie, 'ln' : argd['ln'], 'referer' : CFG_SITE_SECURE_URL + req.unparsed_uri}, {})
return redirect_to_url(req, target, norobot=True)
elif auth_code:
return page_not_authorized(req, "../", \
text=auth_msg, \
navmenuid='search')
#check if the user has rights to set a high wildcard limit
#if not, reduce the limit set by user, with the default one
if CFG_WEBSEARCH_WILDCARD_LIMIT > 0 and (argd['wl'] > CFG_WEBSEARCH_WILDCARD_LIMIT or argd['wl'] == 0):
auth_code, auth_message = acc_authorize_action(req, 'runbibedit')
if auth_code != 0:
argd['wl'] = CFG_WEBSEARCH_WILDCARD_LIMIT
# only superadmins can use verbose parameter for obtaining debug information
if not isUserSuperAdmin(user_info):
argd['verbose'] = 0
# Keep all the arguments, they might be reused in the
# search_engine itself to derivate other queries
req.argd = argd
# mod_python does not like to return [] in case when of=id:
out = perform_request_search(req, **argd)
if isinstance(out, intbitset):
return out.fastdump()
elif out == []:
return str(out)
else:
return out
def record_public_p(recID, recreate_cache_if_needed=True):
"""Return 1 if the record is public, i.e. if it can be found in the Home collection.
Return 0 otherwise.
"""
return recID in get_collection_reclist(CFG_SITE_NAME, recreate_cache_if_needed=recreate_cache_if_needed)
def record_public_p(recID, recreate_cache_if_needed=True):
"""Return 1 if the record is public, i.e. if it can be found in the Home collection.
Return 0 otherwise.
"""
return recID in get_collection_reclist(CFG_SITE_NAME, recreate_cache_if_needed=recreate_cache_if_needed)
