Datasets:
File size: 1,114 Bytes
2f5bf7b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 |
from datasets import load_dataset
def hf_beir_queries(queries):
queries_beir = {}
for query in queries:
queries_beir[query['_id']] = query['text']
return(queries_beir)
def hf_beir_corpus(corpus):
corpus_beir = {}
for doc in corpus:
corpus_beir[doc['_id']] = doc
return(corpus_beir)
def hf_beir_qrels(qrels):
qrels_beir = {}
for el in qrels:
if str(el['query-id']) in qrels_beir:
qrels_beir[str(el['query-id'])][str(el['corpus-id'])] = el['score']
else:
qrels_beir[str(el['query-id'])] = {str(el['corpus-id']): el['score']}
return(qrels_beir)
def load_data(
path,
lang
):
queries = load_dataset(path, 'queries-' + str(lang), split='default')
queries = hf_beir_queries(queries)
corpus = load_dataset(path, 'corpus-' + str(lang), split='default')
corpus = hf_beir_corpus(corpus)
qrels = load_dataset(path, 'qrels-' + str(lang), split='default')
qrels = hf_beir_qrels(qrels)
images = load_dataset(path, 'images-' + str(lang), split='default')
return(queries, corpus, qrels, images)
|