def batch_search(self, queries: List[str], qids: List[str], k: int = 10,
threads: int = 1) -> Dict[str, List[JSimpleSearcherResult]]:
"""Search the collection concurrently for multiple queries, using multiple threads.
Parameters
----------
queries : List[str]
List of query strings.
qids : List[str]
List of corresponding query ids.
k : int
Number of hits to return.
threads : int
Maximum number of threads to use.
Returns
-------
Dict[str, List[JSimpleSearcherResult]]
Dictionary holding the search results, with the query ids as keys and the corresponding lists of search
results as the values.
"""
query_strings = JArrayList()
qid_strings = JArrayList()
for query in queries:
jq = JString(query.encode('utf8'))
query_strings.add(jq)
for qid in qids:
jqid = JString(qid)
qid_strings.add(jqid)
results = self.object.batchSearch(query_strings, qid_strings, int(k), int(threads)).entrySet().toArray()
return {r.getKey(): r.getValue() for r in results}
def batch_search(
self,
queries: List[str],
qids: List[str],
k: int = 10,
threads: int = 1,
fields=dict()) -> Dict[str, List[JImpactSearcherResult]]:
"""Search the collection concurrently for multiple queries, using multiple threads.
Parameters
----------
queries : List[str]
List of query string.
qids : List[str]
List of corresponding query ids.
k : int
Number of hits to return.
threads : int
Maximum number of threads to use.
min_idf : int
Minimum idf for query tokens
fields : dict
Optional map of fields to search with associated boosts.
Returns
-------
Dict[str, List[JImpactSearcherResult]]
Dictionary holding the search results, with the query ids as keys and the corresponding lists of search
results as the values.
"""
query_lst = JArrayList()
qid_lst = JArrayList()
for q in queries:
encoded_query = self.query_encoder.encode(q)
jquery = JHashMap()
for (token, weight) in encoded_query.items():
if token in self.idf and self.idf[token] > self.min_idf:
jquery.put(token, JFloat(weight))
query_lst.add(jquery)
for qid in qids:
jqid = qid
qid_lst.add(jqid)
jfields = JHashMap()
for (field, boost) in fields.items():
jfields.put(field, JFloat(boost))
if not fields:
results = self.object.batchSearch(query_lst, qid_lst, int(k),
int(threads))
else:
results = self.object.batchSearchFields(query_lst, qid_lst, int(k),
int(threads), jfields)
return {r.getKey(): r.getValue() for r in results.entrySet().toArray()}
def batch_search(self, queries: List[str], qids: List[str], k: int = 10, threads: int = 1,
query_generator: JQueryGenerator = None, fields = dict()) -> Dict[str, List[JSimpleSearcherResult]]:
"""Search the collection concurrently for multiple queries, using multiple threads.
Parameters
----------
queries : List[str]
List of query strings.
qids : List[str]
List of corresponding query ids.
k : int
Number of hits to return.
threads : int
Maximum number of threads to use.
query_generator : JQueryGenerator
Generator to build queries. Set to ``None`` by default to use Anserini default.
fields : dict
Optional map of fields to search with associated boosts.
Returns
-------
Dict[str, List[JSimpleSearcherResult]]
Dictionary holding the search results, with the query ids as keys and the corresponding lists of search
results as the values.
"""
query_strings = JArrayList()
qid_strings = JArrayList()
for query in queries:
jq = JString(query.encode('utf8'))
query_strings.add(jq)
for qid in qids:
jqid = JString(qid)
qid_strings.add(jqid)
jfields = JHashMap()
for (field, boost) in fields.items():
jfields.put(JString(field), JFloat(boost))
if query_generator:
if not fields:
results = self.object.batchSearch(query_generator, query_strings, qid_strings, int(k), int(threads))
else:
results = self.object.batchSearchFields(query_generator, query_strings, qid_strings, int(k), int(threads), jfields)
else:
if not fields:
results = self.object.batchSearch(query_strings, qid_strings, int(k), int(threads))
else:
results = self.object.batchSearchFields(query_strings, qid_strings, int(k), int(threads), jfields)
return {r.getKey(): r.getValue() for r in results.entrySet().toArray()}