Datasets:
Tasks:
Text Retrieval
Modalities:
Text
Formats:
parquet
Sub-tasks:
document-retrieval
Languages:
Polish
Size:
100K - 1M
ArXiv:
Tags:
wikipedia
License:
Delete loading script
Browse files- PUGG_IR.py +0 -93
PUGG_IR.py
DELETED
@@ -1,93 +0,0 @@
|
|
1 |
-
import json
|
2 |
-
import os
|
3 |
-
|
4 |
-
import datasets
|
5 |
-
|
6 |
-
logger = datasets.logging.get_logger(__name__)
|
7 |
-
|
8 |
-
_CORPUS = "corpus"
|
9 |
-
_QUERIES = "queries"
|
10 |
-
_QRELS = "qrels"
|
11 |
-
|
12 |
-
URL = ""
|
13 |
-
_URLs = {
|
14 |
-
_CORPUS: f"corpus.jsonl",
|
15 |
-
_QUERIES: f"queries.jsonl",
|
16 |
-
_QRELS: f"qrels/test.tsv",
|
17 |
-
}
|
18 |
-
|
19 |
-
|
20 |
-
class PuggIr(datasets.GeneratorBasedBuilder):
|
21 |
-
|
22 |
-
BUILDER_CONFIGS = [
|
23 |
-
datasets.BuilderConfig(
|
24 |
-
name=_CORPUS,
|
25 |
-
),
|
26 |
-
datasets.BuilderConfig(
|
27 |
-
name=_QUERIES,
|
28 |
-
),
|
29 |
-
datasets.BuilderConfig(
|
30 |
-
name=_QRELS,
|
31 |
-
),
|
32 |
-
]
|
33 |
-
|
34 |
-
def _info(self):
|
35 |
-
if self.config.name == _CORPUS:
|
36 |
-
features = datasets.Features(
|
37 |
-
{
|
38 |
-
"_id": datasets.Value("string"),
|
39 |
-
"title": datasets.Value("string"),
|
40 |
-
"text": datasets.Value("string"),
|
41 |
-
}
|
42 |
-
)
|
43 |
-
elif self.config.name == _QUERIES:
|
44 |
-
features = datasets.Features(
|
45 |
-
{
|
46 |
-
"_id": datasets.Value("string"),
|
47 |
-
"text": datasets.Value("string"),
|
48 |
-
}
|
49 |
-
)
|
50 |
-
elif self.config.name == _QRELS:
|
51 |
-
features = datasets.Features(
|
52 |
-
{
|
53 |
-
"query-id": datasets.Value("string"),
|
54 |
-
"corpus-id": datasets.Value("string"),
|
55 |
-
"score": datasets.Value("int32"),
|
56 |
-
}
|
57 |
-
)
|
58 |
-
|
59 |
-
return datasets.DatasetInfo(
|
60 |
-
features=features,
|
61 |
-
)
|
62 |
-
|
63 |
-
def _split_generators(self, dl_manager):
|
64 |
-
"""Returns SplitGenerators."""
|
65 |
-
my_urls = _URLs[self.config.name]
|
66 |
-
data_dir = dl_manager.download_and_extract(my_urls)
|
67 |
-
|
68 |
-
return [
|
69 |
-
datasets.SplitGenerator(
|
70 |
-
name=datasets.Split.TEST,
|
71 |
-
gen_kwargs={"filepath": data_dir},
|
72 |
-
),
|
73 |
-
]
|
74 |
-
|
75 |
-
def _generate_examples(self, filepath):
|
76 |
-
"""Yields examples."""
|
77 |
-
if self.config.name in [_CORPUS, _QUERIES]:
|
78 |
-
with open(filepath, encoding="utf-8") as f:
|
79 |
-
for i, line in enumerate(f):
|
80 |
-
data = json.loads(line)
|
81 |
-
yield i, data
|
82 |
-
|
83 |
-
elif self.config.name == _QRELS:
|
84 |
-
with open(filepath, encoding="utf-8") as f:
|
85 |
-
for i, line in enumerate(f):
|
86 |
-
if i == 0:
|
87 |
-
continue # Skip header
|
88 |
-
query_id, corpus_id, score = line.strip().split("\t")
|
89 |
-
yield i, {
|
90 |
-
"query-id": query_id,
|
91 |
-
"corpus-id": corpus_id,
|
92 |
-
"score": int(score),
|
93 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|