sps commited on
Commit
94a6d2b
1 Parent(s): 1caa6cf

Update README

Browse files
Files changed (1) hide show
  1. README.md +115 -45
README.md CHANGED
@@ -25,7 +25,7 @@ task_ids:
25
  - extractive-qa
26
  ---
27
 
28
- # Dataset Card for [Dataset Name]
29
 
30
  ## Table of Contents
31
  - [Table of Contents](#table-of-contents)
@@ -54,43 +54,60 @@ task_ids:
54
 
55
  ## Dataset Description
56
 
57
- - **Homepage:**
58
  - **Repository:** [Code repo](https://github.com/adityakanade/natural-cubert/)
 
59
  - **Paper:**
60
 
61
  ### Dataset Summary
62
 
63
  CodeQueries allows to explore extractive question-answering methodology over code
64
- by providing semantic queries as question and answer pairs over code context involving
65
- complex concepts and long chains of reasoning.
66
 
67
  ### Supported Tasks and Leaderboards
68
 
69
- [More Information Needed]
70
 
71
  ### Languages
72
 
73
- The code section have taken from `python` files.
74
 
75
  ## Dataset Structure
76
 
77
- ### Data Instances
78
-
79
- All splits of all settings have same format. An example looks as follows -
80
- ```
 
 
 
 
 
 
 
 
 
 
 
 
81
  ```
82
 
83
- ### Data Fields
 
84
 
85
- - examples
 
86
  - query_name (query name to uniquely identify the query)
87
- - context_blocks (code blocks supplied as input to the model for prediction)
88
- - answer_spans (code in answer spans)
89
- - supporting_fact_spans (code in supporting-fact spans)
90
  - code_file_path (relative source file path w.r.t. ETH Py150 corpus)
91
- - example_type (positive(1) or negative(0) example type)
92
- - subtokenized_input_sequence (example subtokens)
 
 
 
 
93
  - label_sequence (example subtoken labels)
 
 
94
 
95
  ### Data Splits
96
 
@@ -104,62 +121,115 @@ All splits of all settings have same format. An example looks as follows -
104
 
105
  ## Dataset Creation
106
 
107
- ### Curation Rationale
108
 
109
- [More Information Needed]
110
 
111
- ### Source Data
112
 
113
- #### Initial Data Collection and Normalization
 
 
114
 
115
  [More Information Needed]
116
 
117
- #### Who are the source language producers?
118
 
119
- [More Information Needed]
120
 
121
- ### Annotations
 
 
 
 
 
 
 
 
 
 
 
 
122
 
123
- #### Annotation process
124
 
125
- [More Information Needed]
 
 
126
 
127
- #### Who are the annotators?
128
 
129
- [More Information Needed]
 
130
 
131
- ### Personal and Sensitive Information
132
 
133
- [More Information Needed]
134
 
135
- ## Considerations for Using the Data
136
 
137
- ### Social Impact of Dataset
138
 
139
- [More Information Needed]
140
 
141
- ### Discussion of Biases
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
142
 
143
- [More Information Needed]
 
144
 
145
- ### Other Known Limitations
 
 
 
 
 
 
 
 
 
 
 
 
146
 
147
- [More Information Needed]
148
 
149
- ## Additional Information
150
 
151
- ### Dataset Curators
152
 
153
- [More Information Needed]
154
 
 
155
  ### Licensing Information
156
 
157
- [More Information Needed]
158
 
159
  ### Citation Information
160
 
161
  [More Information Needed]
162
-
163
- ### Contributions
164
-
165
- Thanks to [@github-username](https://github.com/<github-username>) for adding this dataset.
 
25
  - extractive-qa
26
  ---
27
 
28
+ # Dataset Card for Codequeries
29
 
30
  ## Table of Contents
31
  - [Table of Contents](#table-of-contents)
 
54
 
55
  ## Dataset Description
56
 
57
+ - **Homepage:** [Codequeires](https://huggingface.co/datasets/thepurpleowl/codequeries)
58
  - **Repository:** [Code repo](https://github.com/adityakanade/natural-cubert/)
59
+ - **Leaderboard:** [Code repo](https://github.com/adityakanade/natural-cubert/)
60
  - **Paper:**
61
 
62
  ### Dataset Summary
63
 
64
  CodeQueries allows to explore extractive question-answering methodology over code
65
+ by providing semantic natural language queries as question and code spans as answer or supporting fact. Given a query, finding the answer/supporting fact spans in code context involves analysis complex concepts and long chains of reasoning. The dataset is provided with five separate settings; details on the setting can be found in the [paper]().
 
66
 
67
  ### Supported Tasks and Leaderboards
68
 
69
+ Query comprehension for code, Extractive question answering for code. Refer the [paper]().
70
 
71
  ### Languages
72
 
73
+ The dataset contains code context from `python` files.
74
 
75
  ## Dataset Structure
76
 
77
+ ### How to use
78
+ The dataset can directly used with huggingface datasets. You can load and iterate through the dataset for the proposed five settings with the following two lines of code:
79
+ ```python
80
+ from datasets import load_dataset
81
+
82
+ ds = load_dataset("thepurpleowl/codequeries", "<ideal/file_ideal/prefix/twostep>", split="train")
83
+ print(next(iter(ds)))
84
+ #OUTPUT:
85
+ {
86
+ 'code': "import mod189 from './mod189';\nvar value=mod189+1;\nexport default value;\n",
87
+ 'repo_name': 'MirekSz/webpack-es6-ts',
88
+ 'path': 'app/mods/mod190.js',
89
+ 'language': 'JavaScript',
90
+ 'license': 'isc',
91
+ 'size': 73
92
+ }
93
  ```
94
 
95
+ ### Data Splits and Data Fields
96
+ Detailed information on the data splits for proposed settings can be found in the paper.
97
 
98
+ In general, data splits in all prpoposed settings have examples in following fields -
99
+ ```
100
  - query_name (query name to uniquely identify the query)
 
 
 
101
  - code_file_path (relative source file path w.r.t. ETH Py150 corpus)
102
+ - context_blocks (code blocks as context with metadata) [`prefix` setting doesn't have this field]
103
+ - answer_spans (answer spans with metadata)
104
+ - supporting_fact_spans (supporting-fact spans with metadata)
105
+ - example_type (1(positive)) or 0(negative)) example type)
106
+ - single_hop (True or False - for query type)
107
+ - subtokenized_input_sequence (example subtokens) [`prefix` setting has the corresponding token ids]
108
  - label_sequence (example subtoken labels)
109
+ - relevance_label (0 (not relevant) or 1 (relevant) - relevance label of a block)
110
+ ```
111
 
112
  ### Data Splits
113
 
 
121
 
122
  ## Dataset Creation
123
 
124
+ The dataset is created by using [ETH Py150 Open corpus](https://github.com/google-research-datasets/eth_py150_open) as source for code contexts. To get natural language queries and corresponding answer/supporting spans in ETH Py150 Open corpus files, CodeQL was used.
125
 
 
126
 
127
+ ### Licensing Information
128
 
129
+ Codequeries dataset is licensed under the [Apache-2.0](https://opensource.org/licenses/Apache-2.0) License.
130
+
131
+ ### Citation Information
132
 
133
  [More Information Needed]
134
 
135
+ ### Contributions
136
 
137
+ Thanks to [@github-username](https://github.com/<github-username>) for adding this dataset.# Dataset Card for Codequeries
138
 
139
+ ## Table of Contents
140
+ - [Table of Contents](#table-of-contents)
141
+ - [Dataset Description](#dataset-description)
142
+ - [Dataset Summary](#dataset-summary)
143
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
144
+ - [Languages](#languages)
145
+ - [Dataset Structure](#dataset-structure)
146
+ - [How to use](#how-to-use)
147
+ - [Data Splits and Data Fields](#data-splits-and-data-fields)
148
+ - [Dataset Creation](#dataset-creation)
149
+ - [Additional Information](#additional-information)
150
+ - [Licensing Information](#licensing-information)
151
+ - [Citation Information](#citation-information)
152
 
153
+ ## Dataset Description
154
 
155
+ - **Homepage:** [Codequeires](https://huggingface.co/datasets/thepurpleowl/codequeries)
156
+ - **Repository:** [Code repo](https://github.com/adityakanade/natural-cubert/)
157
+ - **Paper:**
158
 
159
+ ### Dataset Summary
160
 
161
+ CodeQueries allows to explore extractive question-answering methodology over code
162
+ by providing semantic natural language queries as question and code spans as answer or supporting fact. Given a query, finding the answer/supporting fact spans in code context involves analysis complex concepts and long chains of reasoning. The dataset is provided with five separate settings; details on the setting can be found in the [paper]().
163
 
164
+ ### Supported Tasks and Leaderboards
165
 
166
+ Query comprehension for code, Extractive question answering for code.
167
 
168
+ ### Languages
169
 
170
+ The dataset contains code context from `python` files.
171
 
172
+ ## Dataset Structure
173
 
174
+ ### How to use
175
+ The dataset can directly used with huggingface datasets. You can load and iterate through the dataset for the proposed five settings with the following two lines of code:
176
+ ```python
177
+ import datasets
178
+
179
+ # instead `twostep`, other settings are <ideal/file_ideal/prefix>.
180
+ ds = datasets.load_dataset("thepurpleowl/codequeries", "twostep", split=datasets.Split.TEST)
181
+ print(next(iter(ds)))
182
+
183
+ #OUTPUT:
184
+ {'query_name': 'Unused import',
185
+ 'code_file_path': 'rcbops/glance-buildpackage/glance/tests/unit/test_db.py',
186
+ 'context_block': {'content': '# vim: tabstop=4 shiftwidth=4 softtabstop=4\n\n# Copyright 2010-2011 OpenStack, LLC\ ...',
187
+ 'metadata': 'root',
188
+ 'header': "['module', '___EOS___']",
189
+ 'index': 0},
190
+ 'answer_spans': [{'span': 'from glance.common import context',
191
+ 'start_line': 19,
192
+ 'start_column': 0,
193
+ 'end_line': 19,
194
+ 'end_column': 33}
195
+ ],
196
+ 'supporting_fact_spans': [],
197
+ 'example_type': 1,
198
+ 'single_hop': False,
199
+ 'subtokenized_input_sequence': ['[CLS]_', 'Un', 'used_', 'import_', '[SEP]_', 'module_', '\\u\\u\\uEOS\\u\\u\\u_', '#', ' ', 'vim', ':', ...],
200
+ 'label_sequence': [4, 4, 4, 4, 4, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, ...],
201
+ 'relevance_label': 1
202
+ }
203
+ ```
204
 
205
+ ### Data Splits and Data Fields
206
+ Detailed information on the data splits for proposed settings can be found in the paper.
207
 
208
+ In general, data splits in all prpoposed settings have examples in following fields -
209
+ ```
210
+ - query_name (query name to uniquely identify the query)
211
+ - code_file_path (relative source file path w.r.t. ETH Py150 corpus)
212
+ - context_blocks (code blocks as context with metadata) [`prefix` setting doesn't have this field and `twostep` has `context_block`]
213
+ - answer_spans (answer spans with metadata)
214
+ - supporting_fact_spans (supporting-fact spans with metadata)
215
+ - example_type (1(positive)) or 0(negative)) example type)
216
+ - single_hop (True or False - for query type)
217
+ - subtokenized_input_sequence (example subtokens) [`prefix` setting has the corresponding token ids]
218
+ - label_sequence (example subtoken labels)
219
+ - relevance_label (0 (not relevant) or 1 (relevant) - relevance label of a block) [only `twostep` setting has this field]
220
+ ```
221
 
 
222
 
223
+ ## Dataset Creation
224
 
225
+ The dataset is created by using [ETH Py150 Open corpus](https://github.com/google-research-datasets/eth_py150_open) as source for code contexts. To get natural language queries and corresponding answer/supporting spans in ETH Py150 Open corpus files, CodeQL was used.
226
 
 
227
 
228
+ ## Additional Information
229
  ### Licensing Information
230
 
231
+ Codequeries dataset is licensed under the [Apache-2.0](https://opensource.org/licenses/Apache-2.0) License.
232
 
233
  ### Citation Information
234
 
235
  [More Information Needed]