def push_data_keywords(pub_ids=None, index=None):
""" Go through all the publications and their datatables and move data
keywords from tables to their parent publications. """
if not pub_ids:
body = {"query": {"match_all": {}}}
results = es.search(index=index, doc_type=CFG_PUB_TYPE, body=body, _source=False)
pub_ids = [i["_id"] for i in results["hits"]["hits"]]
for pub_id in pub_ids:
query_builder = QueryBuilder()
query_builder.add_child_parent_relation(
"publication", relation="parent", must=True, related_query={"match": {"recid": pub_id}}
)
tables = es.search(index=index, doc_type=CFG_DATA_TYPE, body=query_builder.query, _source_include="keywords")
keywords = [d["_source"].get("keywords", None) for d in tables["hits"]["hits"]]
# Flatten the list
keywords = [i for inner in keywords for i in inner]
# Aggregate
agg_keywords = defaultdict(list)
for kw in keywords:
agg_keywords[kw["name"]].append(kw["value"])
# Remove duplicates
for k, v in agg_keywords.items():
agg_keywords[k] = list(set(v))
body = {"doc": {"data_keywords": dict(agg_keywords)}}
try:
es.update(index=index, doc_type=CFG_PUB_TYPE, id=pub_id, body=body)
except Exception as e:
log.error(e.message)
def push_data_keywords(pub_ids=None, index=None):
""" Go through all the publications and their datatables and move data
keywords from tables to their parent publications. """
if not pub_ids:
body = {'query': {'match_all': {}}}
results = es.search(index=index,
doc_type=CFG_PUB_TYPE,
body=body,
_source=False)
pub_ids = [i['_id'] for i in results['hits']['hits']]
for pub_id in pub_ids:
query_builder = QueryBuilder()
query_builder.add_child_parent_relation(
'publication',
relation='parent',
must=True,
related_query={'match': {
'recid': pub_id
}})
tables = es.search(index=index,
doc_type=CFG_DATA_TYPE,
body=query_builder.query,
_source_include='keywords')
keywords = [
d['_source'].get('keywords', None) for d in tables['hits']['hits']
]
# Flatten the list
keywords = [i for inner in keywords for i in inner]
# Aggregate
agg_keywords = defaultdict(list)
for kw in keywords:
agg_keywords[kw['name']].append(kw['value'])
# Remove duplicates
for k, v in agg_keywords.items():
agg_keywords[k] = list(set(v))
body = {"doc": {'data_keywords': dict(agg_keywords)}}
try:
es.update(index=index, doc_type=CFG_PUB_TYPE, id=pub_id, body=body)
except Exception as e:
log.error(e.message)
def search(query,
index=None,
filters=list(),
size=10,
include="*",
exclude="authors",
offset=0,
sort_field=None,
sort_order='',
post_filter=None):
""" Perform a search query.
:param query: [string] query string e.g. 'higgs boson'
:param index: [string] name of the index. If None a default is used
:param filters: [list of tuples] list of filters for the query.
Currently supported: ('author', author_fullname),
('collaboration', collaboration_name), ('date', date)
:param size: [int] max number of hits that should be returned
:param offset: [int] offset for the results (used for pagination)
:param sort_by: [string] sorting field. Currently supported fields:
"title", "collaboration", "date", "relevance"
:param sort_order: [string] order of the sorting either original
(for a particular field) or reversed. Supported:
'' or 'rev'
:return: [dict] dictionary with processed results and facets
"""
# If empty query then sort by date
if query == '' and not sort_field:
sort_field = 'date'
query = HEPDataQueryParser.parse_query(query)
# Build core query
data_query = get_query_by_type(CFG_DATA_TYPE, query)
pub_query = get_query_by_type(CFG_PUB_TYPE, query)
authors_query = get_authors_query(query)
query_builder = QueryBuilder()
query_builder.add_child_parent_relation(
CFG_DATA_TYPE,
relation="child",
related_query=data_query,
other_queries=[pub_query, authors_query])
# Add additional options
query_builder.add_pagination(size=size, offset=offset)
query_builder.add_sorting(sort_field=sort_field, sort_order=sort_order)
query_builder.add_filters(filters)
query_builder.add_post_filter(post_filter)
query_builder.add_aggregations()
query_builder.add_source_filter(include, exclude)
pub_result = es.search(index=index,
body=query_builder.query,
doc_type=CFG_PUB_TYPE)
parent_filter = {
"filtered": {
"filter": {
"terms": {
"_id": [hit["_id"] for hit in pub_result['hits']['hits']]
}
}
}
}
query_builder = QueryBuilder()
query_builder.add_child_parent_relation(CFG_PUB_TYPE,
relation="parent",
related_query=parent_filter,
must=True,
other_queries=[data_query])
query_builder.add_pagination(size=size * 50)
data_result = es.search(index=index,
body=query_builder.query,
doc_type=CFG_DATA_TYPE)
merged_results = merge_results(pub_result, data_result)
return map_result(merged_results)
def search(
query,
index=None,
filters=list(),
size=10,
include="*",
exclude="",
offset=0,
sort_field=None,
sort_order="",
post_filter=None,
):
""" Perform a search query.
:param query: [string] query string e.g. 'higgs boson'
:param index: [string] name of the index. If None a default is used
:param filters: [list of tuples] list of filters for the query.
Currently supported: ('author', author_fullname),
('collaboration', collaboration_name), ('date', date)
:param size: [int] max number of hits that should be returned
:param offset: [int] offset for the results (used for pagination)
:param sort_by: [string] sorting field. Currently supported fields:
"title", "collaboration", "date", "relevance"
:param sort_order: [string] order of the sorting either original
(for a particular field) or reversed. Supported:
'' or 'rev'
:return: [dict] dictionary with processed results and facets
"""
# If empty query then sort by date
if query == "" and not sort_field:
sort_field = "date"
query = HEPDataQueryParser.parse_query(query)
# Build core query
data_query = get_query_by_type(CFG_DATA_TYPE, query)
pub_query = get_query_by_type(CFG_PUB_TYPE, query)
authors_query = get_authors_query(query)
query_builder = QueryBuilder()
query_builder.add_child_parent_relation(
CFG_DATA_TYPE, relation="child", related_query=data_query, other_queries=[pub_query, authors_query]
)
# Add additional options
query_builder.add_pagination(size=size, offset=offset)
query_builder.add_sorting(sort_field=sort_field, sort_order=sort_order)
query_builder.add_filters(filters)
query_builder.add_post_filter(post_filter)
query_builder.add_aggregations()
query_builder.add_source_filter(include, exclude)
pub_result = es.search(index=index, body=query_builder.query, doc_type=CFG_PUB_TYPE)
parent_filter = {"filtered": {"filter": {"terms": {"_id": [hit["_id"] for hit in pub_result["hits"]["hits"]]}}}}
query_builder = QueryBuilder()
query_builder.add_child_parent_relation(
CFG_PUB_TYPE, relation="parent", related_query=parent_filter, must=True, other_queries=[data_query]
)
query_builder.add_pagination(size=size * 50)
data_result = es.search(index=index, body=query_builder.query, doc_type=CFG_DATA_TYPE)
merged_results = merge_results(pub_result, data_result)
return map_result(merged_results)
