meryyllebr543 commited on
Commit
da22a35
·
verified ·
1 Parent(s): 7d34ac0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +402 -25
README.md CHANGED
@@ -1,10 +1,404 @@
1
  ---
2
  license: other
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3
  ---
4
 
5
- # 💻 Processed Stack-Edu
6
 
7
- This repository contains a pre-downloaded and processed version of the **[Stack-Edu](https://huggingface.co/datasets/HuggingFaceTB/stack-edu)** dataset. The original dataset consists of SWHIDs pointing to files on Software Heritage. This version contains the **full text content**, saved in Parquet format and organized by programming language, making it immediately usable for training and analysis without the slow download process from S3.
8
 
9
  Stack-Edu is a 125B token dataset of high-quality, educational code filtered from [The Stack v2](https://huggingface.co/datasets/bigcode/the-stack-v2). It was curated by the Hugging Face team for training the StarCoder2 models.
10
 
@@ -14,45 +408,28 @@ The data is organized into one Parquet file per programming language. Each langu
14
 
15
  ### Available Languages (Configurations)
16
 
17
- You can load data for any of the following languages:
18
  `c`, `cpp`, `csharp`, `go`, `java`, `javascript`, `markdown`, `php`, `python`, `ruby`, `rust`, `shell`, `sql`, `swift`, `typescript`.
19
 
20
  ## How to Use
21
 
22
- You can easily load any language-specific subset using the `datasets` library.
23
 
24
  ```python
25
  from datasets import load_dataset
26
 
27
  # Load a specific language, for example, Python
28
- # This will automatically use the loading script in the repository
29
- python_data = load_dataset("meryyllebr543/stack-edu-huggingface", name="python", split="train")
30
 
31
  # Load the C++ data
32
  cpp_data = load_dataset("meryyllebr543/stack-edu-huggingface", name="cpp", split="train")
33
 
34
  # You can then iterate through the dataset
35
- print(python_data[0])
36
- # Expected output:
37
- # {
38
- # 'language': 'python',
39
- # 'text': '# Your python code here...',
40
- # ... (other columns)
41
- # }
42
  ```
43
 
44
- ### Data Schema
45
-
46
- The dataset contains the following columns:
47
- * `language`: The programming language of the file.
48
- * `src_encoding`: The detected source encoding.
49
- * `length_bytes`: The size of the file in bytes.
50
- * `score`: The educational quality score (float).
51
- * `int_score`: The integer version of the quality score.
52
- * `detected_licenses`: A list of licenses detected in the file.
53
- * `license_type`: The general type of license.
54
- * `text`: The full text content of the code file.
55
-
56
  ## Original Dataset Information
57
 
58
  For more details on the curation process, please refer to the original **[Stack-Edu dataset card](https://huggingface.co/datasets/HuggingFaceTB/stack-edu)**.
 
1
  ---
2
  license: other
3
+ dataset_info:
4
+ - config_name: c
5
+ features:
6
+ - name: language
7
+ dtype: string
8
+ - name: src_encoding
9
+ dtype: string
10
+ - name: length_bytes
11
+ dtype: int64
12
+ - name: score
13
+ dtype: float64
14
+ - name: int_score
15
+ dtype: int64
16
+ - name: detected_licenses
17
+ sequence: string
18
+ - name: license_type
19
+ dtype: string
20
+ - name: text
21
+ dtype: string
22
+ splits:
23
+ - name: train
24
+ num_bytes: ???
25
+ num_examples: ???
26
+ - config_name: cpp
27
+ features:
28
+ - name: language
29
+ dtype: string
30
+ - name: src_encoding
31
+ dtype: string
32
+ - name: length_bytes
33
+ dtype: int64
34
+ - name: score
35
+ dtype: float64
36
+ - name: int_score
37
+ dtype: int64
38
+ - name: detected_licenses
39
+ sequence: string
40
+ - name: license_type
41
+ dtype: string
42
+ - name: text
43
+ dtype: string
44
+ splits:
45
+ - name: train
46
+ num_bytes: ???
47
+ num_examples: ???
48
+ - config_name: csharp
49
+ features:
50
+ - name: language
51
+ dtype: string
52
+ - name: src_encoding
53
+ dtype: string
54
+ - name: length_bytes
55
+ dtype: int64
56
+ - name: score
57
+ dtype: float64
58
+ - name: int_score
59
+ dtype: int64
60
+ - name: detected_licenses
61
+ sequence: string
62
+ - name: license_type
63
+ dtype: string
64
+ - name: text
65
+ dtype: string
66
+ splits:
67
+ - name: train
68
+ num_bytes: ???
69
+ num_examples: ???
70
+ - config_name: go
71
+ features:
72
+ - name: language
73
+ dtype: string
74
+ - name: src_encoding
75
+ dtype: string
76
+ - name: length_bytes
77
+ dtype: int64
78
+ - name: score
79
+ dtype: float64
80
+ - name: int_score
81
+ dtype: int64
82
+ - name: detected_licenses
83
+ sequence: string
84
+ - name: license_type
85
+ dtype: string
86
+ - name: text
87
+ dtype: string
88
+ splits:
89
+ - name: train
90
+ num_bytes: ???
91
+ num_examples: ???
92
+ - config_name: java
93
+ features:
94
+ - name: language
95
+ dtype: string
96
+ - name: src_encoding
97
+ dtype: string
98
+ - name: length_bytes
99
+ dtype: int64
100
+ - name: score
101
+ dtype: float64
102
+ - name: int_score
103
+ dtype: int64
104
+ - name: detected_licenses
105
+ sequence: string
106
+ - name: license_type
107
+ dtype: string
108
+ - name: text
109
+ dtype: string
110
+ splits:
111
+ - name: train
112
+ num_bytes: ???
113
+ num_examples: ???
114
+ - config_name: javascript
115
+ features:
116
+ - name: language
117
+ dtype: string
118
+ - name: src_encoding
119
+ dtype: string
120
+ - name: length_bytes
121
+ dtype: int64
122
+ - name: score
123
+ dtype: float64
124
+ - name: int_score
125
+ dtype: int64
126
+ - name: detected_licenses
127
+ sequence: string
128
+ - name: license_type
129
+ dtype: string
130
+ - name: text
131
+ dtype: string
132
+ splits:
133
+ - name: train
134
+ num_bytes: ???
135
+ num_examples: ???
136
+ - config_name: markdown
137
+ features:
138
+ - name: language
139
+ dtype: string
140
+ - name: src_encoding
141
+ dtype: string
142
+ - name: length_bytes
143
+ dtype: int64
144
+ - name: score
145
+ dtype: float64
146
+ - name: int_score
147
+ dtype: int64
148
+ - name: detected_licenses
149
+ sequence: string
150
+ - name: license_type
151
+ dtype: string
152
+ - name: text
153
+ dtype: string
154
+ splits:
155
+ - name: train
156
+ num_bytes: ???
157
+ num_examples: ???
158
+ - config_name: php
159
+ features:
160
+ - name: language
161
+ dtype: string
162
+ - name: src_encoding
163
+ dtype: string
164
+ - name: length_bytes
165
+ dtype: int64
166
+ - name: score
167
+ dtype: float64
168
+ - name: int_score
169
+ dtype: int64
170
+ - name: detected_licenses
171
+ sequence: string
172
+ - name: license_type
173
+ dtype: string
174
+ - name: text
175
+ dtype: string
176
+ splits:
177
+ - name: train
178
+ num_bytes: ???
179
+ num_examples: ???
180
+ - config_name: python
181
+ features:
182
+ - name: language
183
+ dtype: string
184
+ - name: src_encoding
185
+ dtype: string
186
+ - name: length_bytes
187
+ dtype: int64
188
+ - name: score
189
+ dtype: float64
190
+ - name: int_score
191
+ dtype: int64
192
+ - name: detected_licenses
193
+ sequence: string
194
+ - name: license_type
195
+ dtype: string
196
+ - name: text
197
+ dtype: string
198
+ splits:
199
+ - name: train
200
+ num_bytes: ???
201
+ num_examples: ???
202
+ - config_name: ruby
203
+ features:
204
+ - name: language
205
+ dtype: string
206
+ - name: src_encoding
207
+ dtype: string
208
+ - name: length_bytes
209
+ dtype: int64
210
+ - name: score
211
+ dtype: float64
212
+ - name: int_score
213
+ dtype: int64
214
+ - name: detected_licenses
215
+ sequence: string
216
+ - name: license_type
217
+ dtype: string
218
+ - name: text
219
+ dtype: string
220
+ splits:
221
+ - name: train
222
+ num_bytes: ???
223
+ num_examples: ???
224
+ - config_name: rust
225
+ features:
226
+ - name: language
227
+ dtype: string
228
+ - name: src_encoding
229
+ dtype: string
230
+ - name: length_bytes
231
+ dtype: int64
232
+ - name: score
233
+ dtype: float64
234
+ - name: int_score
235
+ dtype: int64
236
+ - name: detected_licenses
237
+ sequence: string
238
+ - name: license_type
239
+ dtype: string
240
+ - name: text
241
+ dtype: string
242
+ splits:
243
+ - name: train
244
+ num_bytes: ???
245
+ num_examples: ???
246
+ - config_name: shell
247
+ features:
248
+ - name: language
249
+ dtype: string
250
+ - name: src_encoding
251
+ dtype: string
252
+ - name: length_bytes
253
+ dtype: int64
254
+ - name: score
255
+ dtype: float64
256
+ - name: int_score
257
+ dtype: int64
258
+ - name: detected_licenses
259
+ sequence: string
260
+ - name: license_type
261
+ dtype: string
262
+ - name: text
263
+ dtype: string
264
+ splits:
265
+ - name: train
266
+ num_bytes: ???
267
+ num_examples: ???
268
+ - config_name: sql
269
+ features:
270
+ - name: language
271
+ dtype: string
272
+ - name: src_encoding
273
+ dtype: string
274
+ - name: length_bytes
275
+ dtype: int64
276
+ - name: score
277
+ dtype: float64
278
+ - name: int_score
279
+ dtype: int64
280
+ - name: detected_licenses
281
+ sequence: string
282
+ - name: license_type
283
+ dtype: string
284
+ - name: text
285
+ dtype: string
286
+ splits:
287
+ - name: train
288
+ num_bytes: ???
289
+ num_examples: ???
290
+ - config_name: swift
291
+ features:
292
+ - name: language
293
+ dtype: string
294
+ - name: src_encoding
295
+ dtype: string
296
+ - name: length_bytes
297
+ dtype: int64
298
+ - name: score
299
+ dtype: float64
300
+ - name: int_score
301
+ dtype: int64
302
+ - name: detected_licenses
303
+ sequence: string
304
+ - name: license_type
305
+ dtype: string
306
+ - name: text
307
+ dtype: string
308
+ splits:
309
+ - name: train
310
+ num_bytes: ???
311
+ num_examples: ???
312
+ - config_name: typescript
313
+ features:
314
+ - name: language
315
+ dtype: string
316
+ - name: src_encoding
317
+ dtype: string
318
+ - name: length_bytes
319
+ dtype: int64
320
+ - name: score
321
+ dtype: float64
322
+ - name: int_score
323
+ dtype: int64
324
+ - name: detected_licenses
325
+ sequence: string
326
+ - name: license_type
327
+ dtype: string
328
+ - name: text
329
+ dtype: string
330
+ splits:
331
+ - name: train
332
+ num_bytes: ???
333
+ num_examples: ???
334
+
335
+ # The `configs` section maps the config names to the actual data files
336
+ configs:
337
+ - config_name: c
338
+ data_files:
339
+ - split: train
340
+ path: c.parquet
341
+ - config_name: cpp
342
+ data_files:
343
+ - split: train
344
+ path: cpp.parquet
345
+ - config_name: csharp
346
+ data_files:
347
+ - split: train
348
+ path: csharp.parquet
349
+ - config_name: go
350
+ data_files:
351
+ - split: train
352
+ path: go.parquet
353
+ - config_name: java
354
+ data_files:
355
+ - split: train
356
+ path: java.parquet
357
+ - config_name: javascript
358
+ data_files:
359
+ - split: train
360
+ path: javascript.parquet
361
+ - config_name: markdown
362
+ data_files:
363
+ - split: train
364
+ path: markdown.parquet
365
+ - config_name: php
366
+ data_files:
367
+ - split: train
368
+ path: php.parquet
369
+ - config_name: python
370
+ data_files:
371
+ - split: train
372
+ path: python.parquet
373
+ - config_name: ruby
374
+ data_files:
375
+ - split: train
376
+ path: ruby.parquet
377
+ - config_name: rust
378
+ data_files:
379
+ - split: train
380
+ path: rust.parquet
381
+ - config_name: shell
382
+ data_files:
383
+ - split: train
384
+ path: shell.parquet
385
+ - config_name: sql
386
+ data_files:
387
+ - split: train
388
+ path: sql.parquet
389
+ - config_name: swift
390
+ data_files:
391
+ - split: train
392
+ path: swift.parquet
393
+ - config_name: typescript
394
+ data_files:
395
+ - split: train
396
+ path: typescript.parquet
397
  ---
398
 
399
+ # 💻 Processed Stack-Edu (Direct Access)
400
 
401
+ This repository contains a pre-downloaded and processed version of the **[Stack-Edu](https://huggingface.co/datasets/HuggingFaceTB/stack-edu)** dataset. The original dataset consists of SWHIDs pointing to files on Software Heritage. This version contains the **full text content**, saved in Parquet format and organized by programming language, making it immediately usable for training and analysis without the slow download process.
402
 
403
  Stack-Edu is a 125B token dataset of high-quality, educational code filtered from [The Stack v2](https://huggingface.co/datasets/bigcode/the-stack-v2). It was curated by the Hugging Face team for training the StarCoder2 models.
404
 
 
408
 
409
  ### Available Languages (Configurations)
410
 
411
+ You can load data for any of the following languages by specifying its name in the `load_dataset` function:
412
  `c`, `cpp`, `csharp`, `go`, `java`, `javascript`, `markdown`, `php`, `python`, `ruby`, `rust`, `shell`, `sql`, `swift`, `typescript`.
413
 
414
  ## How to Use
415
 
416
+ You can easily load any language-specific subset using the `datasets` library. The dataset card metadata in this `README.md` allows for direct loading of each configuration.
417
 
418
  ```python
419
  from datasets import load_dataset
420
 
421
  # Load a specific language, for example, Python
422
+ # Note: It is recommended to use streaming=True for large datasets to avoid downloading everything at once.
423
+ python_data = load_dataset("meryyllebr543/stack-edu-huggingface", name="python", split="train", streaming=True)
424
 
425
  # Load the C++ data
426
  cpp_data = load_dataset("meryyllebr543/stack-edu-huggingface", name="cpp", split="train")
427
 
428
  # You can then iterate through the dataset
429
+ for example in python_data.take(5):
430
+ print(example['text'][:200]) # Print the first 200 characters of the text
 
 
 
 
 
431
  ```
432
 
 
 
 
 
 
 
 
 
 
 
 
 
433
  ## Original Dataset Information
434
 
435
  For more details on the curation process, please refer to the original **[Stack-Edu dataset card](https://huggingface.co/datasets/HuggingFaceTB/stack-edu)**.