fernando-peres
commited on
Commit
•
fcc0e32
1
Parent(s):
f11f1f0
Builder created
Browse files- test_builder.py +94 -0
test_builder.py
ADDED
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
import os
|
3 |
+
import textwrap
|
4 |
+
from textwrap import TextWrapper
|
5 |
+
import datasets
|
6 |
+
import pyarrow.parquet as pq
|
7 |
+
|
8 |
+
_URLS = {
|
9 |
+
"original_text": "./original_text",
|
10 |
+
"unlabeled_sentences": "./unlabeled_sentences"
|
11 |
+
}
|
12 |
+
|
13 |
+
_metadata = {
|
14 |
+
"citation": """\
|
15 |
+
@InProceedings{
|
16 |
+
huggingface:dataset,
|
17 |
+
title = {Paraguay Legislation Dataset},
|
18 |
+
author={Peres, Fernando; Costa, Victor},
|
19 |
+
year={2023}
|
20 |
+
}
|
21 |
+
""",
|
22 |
+
|
23 |
+
"description": "Dataset for researching.",
|
24 |
+
|
25 |
+
"homepage": "https://www.leyes.com.py/",
|
26 |
+
|
27 |
+
"license": "apache-2.0",
|
28 |
+
}
|
29 |
+
|
30 |
+
|
31 |
+
class TestBuilder(datasets.GeneratorBasedBuilder):
|
32 |
+
VERSION = datasets.Version("1.0.0")
|
33 |
+
|
34 |
+
BUILDER_CONFIGS = [
|
35 |
+
datasets.BuilderConfig(
|
36 |
+
name="raw_text",
|
37 |
+
version=VERSION,
|
38 |
+
description="desc raw text",
|
39 |
+
),
|
40 |
+
|
41 |
+
datasets.BuilderConfig(
|
42 |
+
name="unlabeled_sentences",
|
43 |
+
version=VERSION,
|
44 |
+
description="desc unlabeled",
|
45 |
+
),
|
46 |
+
]
|
47 |
+
|
48 |
+
def _info(self):
|
49 |
+
|
50 |
+
features = None
|
51 |
+
if self.config.name == "raw_text":
|
52 |
+
|
53 |
+
features = datasets.Features(
|
54 |
+
{
|
55 |
+
"id": datasets.Value(dtype="int64"),
|
56 |
+
"text": datasets.Value(dtype="string"),
|
57 |
+
}
|
58 |
+
)
|
59 |
+
|
60 |
+
if self.config.name == "unlabeled_sentences":
|
61 |
+
features = features = datasets.Features(
|
62 |
+
{
|
63 |
+
"id": datasets.Value(dtype="int64"),
|
64 |
+
"text_2": datasets.Value(dtype="string"),
|
65 |
+
"note": datasets.Value(dtype="string"),
|
66 |
+
}
|
67 |
+
)
|
68 |
+
|
69 |
+
return datasets.DatasetInfo(
|
70 |
+
builder_name=self.config.name,
|
71 |
+
description="description xxxxxxxxxxxxxxx",
|
72 |
+
features=features,
|
73 |
+
homepage=_metadata["homepage"],
|
74 |
+
license=_metadata["license"],
|
75 |
+
citation=_metadata["citation"],
|
76 |
+
)
|
77 |
+
|
78 |
+
def _split_generators(self, dl_manager):
|
79 |
+
urls_to_download = _URLS
|
80 |
+
|
81 |
+
filepaths = dl_manager.download_and_extract(urls_to_download)
|
82 |
+
|
83 |
+
return datasets.SplitGenerator(
|
84 |
+
name=datasets.Split.TRAIN,
|
85 |
+
gen_kwargs={"filepath": filepaths},
|
86 |
+
)
|
87 |
+
|
88 |
+
def _generate_examples(self, filepath):
|
89 |
+
pq_table = pq.read_table(filepath)
|
90 |
+
for i in range(len(pq_table)):
|
91 |
+
yield i, {
|
92 |
+
col_name: pq_table[col_name][i].as_py()
|
93 |
+
for col_name in pq_table.column_names
|
94 |
+
}
|