Datasets:

Modalities:
Text
Languages:
code
ArXiv:
Libraries:
Datasets
License:
loubnabnl HF staff commited on
Commit
f1e6e86
1 Parent(s): 1d577d2

add dataset script

Browse files
Files changed (1) hide show
  1. xlcost-text-to-code.py +221 -0
xlcost-text-to-code.py ADDED
@@ -0,0 +1,221 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
3
+ #
4
+ # Licensed under the Apache License, Version 2.0 (the "License");
5
+ # you may not use this file except in compliance with the License.
6
+ # You may obtain a copy of the License at
7
+ #
8
+ # http://www.apache.org/licenses/LICENSE-2.0
9
+ #
10
+ # Unless required by applicable law or agreed to in writing, software
11
+ # distributed under the License is distributed on an "AS IS" BASIS,
12
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
+ # See the License for the specific language governing permissions and
14
+ # limitations under the License.
15
+ """APPS dataset."""
16
+
17
+
18
+ import json
19
+ import datasets
20
+
21
+
22
+ _CITATION = """\
23
+ @misc{zhu2022xlcost,
24
+ title = {XLCoST: A Benchmark Dataset for Cross-lingual Code Intelligence},
25
+ url = {https://arxiv.org/abs/2206.08474},
26
+ author = {Zhu, Ming and Jain, Aneesh and Suresh, Karthik and Ravindran, Roshan and Tipirneni, Sindhu and Reddy, Chandan K.},
27
+ year = {2022},
28
+ eprint={2206.08474},
29
+ archivePrefix={arXiv}
30
+ }
31
+ """
32
+
33
+ _DESCRIPTION = """\
34
+ XLCoST is a machine learning benchmark dataset that contains fine-grained parallel data in 7 commonly used programming languages (C++, Java, Python, C#, Javascript, PHP, C), and natural language (English)."""
35
+
36
+ _HOMEPAGE = "https://github.com/reddy-lab-code-research/XLCoST"
37
+ _URLS = {
38
+ "train": "train.json",
39
+ "test": "test.json",
40
+ "valid": "valid.json",
41
+ }
42
+
43
+ def new_url(name):
44
+ urls = {"train": f"data/{name}/{_URLS['train']}",
45
+ "test": f"data/{name}/{_URLS['test']}",
46
+ "valid": f"data/{name}/{_URLS['valid']}"}
47
+ return urls
48
+
49
+ class XlcostConfig(datasets.BuilderConfig):
50
+ """BuilderConfig for SuperGLUE."""
51
+
52
+ def __init__(self, name, description, features, **kwargs):
53
+ #super().__init__(name, description, features, version=datasets.Version("2.1.0"), **kwargs,) #=datasets.Version("2.1.0")
54
+ super(XlcostConfig, self).__init__(version=datasets.Version("2.1.0", ""), **kwargs)
55
+ self.name = name
56
+ self.description = description
57
+ self.features = features
58
+
59
+ def split_generator(dl_manager, name):
60
+ downloaded_files = new_url(name)
61
+ downloaded_files = dl_manager.download(new_url(name))
62
+ return [
63
+ datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": downloaded_files["train"]}),
64
+ datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": downloaded_files["test"]}),
65
+ datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files["valid"]}),
66
+ ]
67
+
68
+ class Xlcost(datasets.GeneratorBasedBuilder):
69
+ """The SuperGLUE benchmark."""
70
+ VERSION = datasets.Version("2.1.0")
71
+ BUILDER_CONFIGS = [
72
+ XlcostConfig(
73
+ name="Python-snippet-level",
74
+ #version=VERSION,
75
+ description="snippet level text and code pairs for Python",
76
+ features=["text", "code"]
77
+ ),
78
+ XlcostConfig(
79
+ name="Python-program-level",
80
+ #version=VERSION,
81
+ description="program level text and code pairs for Python",
82
+ features=["text", "code"]
83
+ ),
84
+
85
+ XlcostConfig(
86
+ name="C-snippet-level",
87
+ #version=VERSION,
88
+ description="snippet level text and code pairs for C",
89
+ features=["text", "code"]
90
+ ),
91
+ XlcostConfig(
92
+ name="C-program-level",
93
+ #version=VERSION,
94
+ description="program level text and code pairs for C",
95
+ features=["text", "code"]
96
+ ),
97
+
98
+ XlcostConfig(
99
+ name="Java-snippet-level",
100
+ #version=VERSION,
101
+ description="snippet level text and code pairs for Java",
102
+ features=["text", "code"]
103
+ ),
104
+ XlcostConfig(
105
+ name="Java-program-level",
106
+ #version=VERSION,
107
+ description="program level text and code pairs for Java",
108
+ features=["text", "code"]
109
+ ),
110
+
111
+ XlcostConfig(
112
+ name="Javascript-snippet-level",
113
+ #version=VERSION,
114
+ description="snippet level text and code pairs for PHP",
115
+ features=["text", "code"]
116
+ ),
117
+ XlcostConfig(
118
+ name="Javascript-program-level",
119
+ #version=VERSION,
120
+ description="program level text and code pairs for PHP",
121
+ features=["text", "code"]
122
+ ),
123
+
124
+ XlcostConfig(
125
+ name="Csharp-snippet-level",
126
+ #version=VERSION,
127
+ description="snippet level text and code pairs for C#",
128
+ features=["text", "code"]
129
+ ),
130
+ XlcostConfig(
131
+ name="Csharp-program-level",
132
+ #version=VERSION,
133
+ description="program level text and code pairs for C#",
134
+ features=["text", "code"]
135
+ ),
136
+
137
+ XlcostConfig(
138
+ name="C++-snippet-level",
139
+ #version=VERSION,
140
+ description="snippet level text and code pairs for C#",
141
+ features=["text", "code"]
142
+ ),
143
+ XlcostConfig(
144
+ name="C++-program-level",
145
+ #version=VERSION,
146
+ description="program level text and code pairs for C#",
147
+ features=["text", "code"]
148
+ ),
149
+
150
+ XlcostConfig(
151
+ name="PHP-snippet-level",
152
+ #version=VERSION,
153
+ description="snippet level text and code pairs for PHP",
154
+ features=["text", "code"]
155
+ ),
156
+ XlcostConfig(
157
+ name="PHP-program-level",
158
+ #version=VERSION,
159
+ description="program level text and code pairs for PHP",
160
+ features=["text", "code"]
161
+ ),
162
+ ]
163
+ DEFAULT_CONFIG_NAME = "C-snippet-level"
164
+
165
+ def _info(self):
166
+ return datasets.DatasetInfo(
167
+ description=_DESCRIPTION,
168
+ features=datasets.Features({"text": datasets.Value("string"),
169
+ "code": datasets.Value("string")}),
170
+ homepage=_HOMEPAGE,
171
+ citation=_CITATION,
172
+ )
173
+
174
+ def _split_generators(self, dl_manager):
175
+ if self.config.name == "Python-snippet-level":
176
+ return split_generator(dl_manager, self.config.name)
177
+
178
+ elif self.config.name == "Python-program-level":
179
+ return split_generator(dl_manager, self.config.name)
180
+
181
+ elif self.config.name == "C-snippet-level":
182
+ return split_generator(dl_manager, self.config.name)
183
+
184
+ elif self.config.name == "C-program-level":
185
+ return split_generator(dl_manager, self.config.name)
186
+
187
+ elif self.config.name == "Java-snippet-level":
188
+ return split_generator(dl_manager, self.config.name)
189
+
190
+ elif self.config.name == "Java-program-level":
191
+ return split_generator(dl_manager, self.config.name)
192
+
193
+ elif self.config.name == "Javascript-snippet-level":
194
+ return split_generator(dl_manager, self.config.name)
195
+
196
+ elif self.config.name == "Javascript-program-level":
197
+ return split_generator(dl_manager, self.config.name)
198
+
199
+ elif self.config.name == "Csharp-snippet-level":
200
+ return split_generator(dl_manager, self.config.name)
201
+
202
+ elif self.config.name == "Csharp-program-level":
203
+ return split_generator(dl_manager, self.config.name)
204
+
205
+ elif self.config.name == "PHP-snippet-level":
206
+ return split_generator(dl_manager, self.config.name)
207
+
208
+ elif self.config.name == "PHP-program-level":
209
+ return split_generator(dl_manager, self.config.name)
210
+
211
+ def _generate_examples(self, filepath):
212
+ key = 0
213
+ with open(filepath) as f:
214
+ for line in f:
215
+ row = json.loads(line)
216
+ key += 1
217
+ yield key, {
218
+ "text": row["text"],
219
+ "code": row["code"],
220
+ }
221
+ key += 1