Datasets:
Tasks:
Text Generation
Formats:
parquet
Sub-tasks:
language-modeling
Languages:
Danish
Size:
10M - 100M
ArXiv:
DOI:
License:
Kenneth Enevoldsen
commited on
clean up readme
Browse files
README.md
CHANGED
|
@@ -130,7 +130,7 @@ language_bcp47:
|
|
| 130 |
|
| 131 |
## Dataset Description
|
| 132 |
|
| 133 |
-
This is
|
| 134 |
|
| 135 |
### Dataset Summary
|
| 136 |
|
|
@@ -160,15 +160,14 @@ Each entry in the dataset consists of a single text with associated metadata
|
|
| 160 |
|
| 161 |
```py
|
| 162 |
{
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
}
|
| 172 |
}
|
| 173 |
```
|
| 174 |
|
|
@@ -181,9 +180,10 @@ An entry in the dataset consists of the following fields:
|
|
| 181 |
- `id` (`str`): An unique identifier for each document.
|
| 182 |
- `added` (`str`): An date for when the document was added to this collection.
|
| 183 |
- `created` (`str`): An date range for when the document was originally created.
|
| 184 |
-
- `
|
| 185 |
-
- `
|
| 186 |
- `metadata/source-pretty` (`str`): The long form version of the short-form source name
|
|
|
|
| 187 |
|
| 188 |
|
| 189 |
### Data Splits
|
|
@@ -196,90 +196,33 @@ The entire corpus is provided in the `train` split.
|
|
| 196 |
|
| 197 |
Below follows a brief overview of the sources in the corpus along with their individual license.
|
| 198 |
|
| 199 |
-
| Source | License
|
| 200 |
-
| ----------------- |
|
| 201 |
-
| adl | Creative Commons Legal Code 1.0 Universal
|
| 202 |
-
| botxt | Creative Commons Legal Code 1.0 Universal
|
| 203 |
-
| dannet | [dannet license]
|
| 204 |
-
| depbank | Attribution-ShareAlike 4.0 International
|
| 205 |
-
| ep | Creative Commons Legal Code 1.0 Universal
|
| 206 |
-
| ft | Creative Commons Legal Code 1.0 Universal
|
| 207 |
-
| gutenberg | [gutenberg license]
|
| 208 |
-
| hest | Creative Commons Legal Code 1.0 Universal
|
| 209 |
-
| jvj | Attribution-ShareAlike 4.0 International
|
| 210 |
-
| naat | Creative Commons Legal Code 1.0 Universal
|
| 211 |
-
| relig | Creative Commons Legal Code 1.0 Universal
|
| 212 |
-
| retsinformationdk | Danish
|
| 213 |
-
| retspraksis | Creative Commons Legal Code 1.0 Universal
|
| 214 |
-
| skat | Creative Commons Legal Code 1.0 Universal
|
| 215 |
-
| spont | Creative Commons Legal Code 1.0 Universal
|
| 216 |
-
| synne | Creative Commons Legal Code 1.0 Universal
|
| 217 |
-
| tv2r |
|
| 218 |
-
| wiki | Creative Commons Legal Code 1.0 Universal
|
| 219 |
-
| wikibooks | Creative Commons Legal Code 1.0 Universal
|
| 220 |
-
| wikisource | Creative Commons Legal Code 1.0 Universal
|
| 221 |
-
|
| 222 |
-
|
| 223 |
-
|
| 224 |
-
|
| 225 |
-
|
| 226 |
-
"retsinformationdk": "Legal",
|
| 227 |
-
"skat": "Legal",
|
| 228 |
-
"retspraksis": "Legal",
|
| 229 |
-
"hest": "Social Media",
|
| 230 |
-
"cc": "Web",
|
| 231 |
-
"adl": "Wiki & Books",
|
| 232 |
-
"botxt": "Other",
|
| 233 |
-
"danavis": "News",
|
| 234 |
-
"dannet": "dannet",
|
| 235 |
-
"depbank": "Other",
|
| 236 |
-
"ep": "Conversation",
|
| 237 |
-
"ft": "Conversation",
|
| 238 |
-
"gutenberg": "Wiki & Books",
|
| 239 |
-
"jvj": "Wiki & Books",
|
| 240 |
-
"naat": "Conversation",
|
| 241 |
-
"opensub": "Conversation",
|
| 242 |
-
"relig": "Wiki & Books",
|
| 243 |
-
"spont": "Conversation",
|
| 244 |
-
"synne": "Other",
|
| 245 |
-
"tv2r": "News",
|
| 246 |
-
"wiki": "Wiki & Books",
|
| 247 |
-
"wikibooks": "Wiki & Books",
|
| 248 |
-
"wikisource": "Wiki & Books",
|
| 249 |
-
"twfv19": "Social Media", # not present in this version of the dataset
|
| 250 |
-
}
|
| 251 |
-
```
|
| 252 |
-
|
| 253 |
-
And the following mapping translates between the short form and the long form of the source name
|
| 254 |
-
```python
|
| 255 |
-
# mapping from domain to its long name format
|
| 256 |
-
longname_mapping_dict = {
|
| 257 |
-
"retsinformationdk": "retsinformation.dk (Danish legal information)",
|
| 258 |
-
"skat": "Skat (Danish tax authority)",
|
| 259 |
-
"retspraksis": "retspraksis (Danish legal information)",
|
| 260 |
-
"hest": "Hestenettet (Danish debate forum)",
|
| 261 |
-
"cc": "Common Crawl",
|
| 262 |
-
"adl": " Archive for Danish Literature",
|
| 263 |
-
"botxt": "Bornholmsk (Danish dialect)",
|
| 264 |
-
"danavis": "Danish daily newspapers",
|
| 265 |
-
"dannet": "DanNet (Danish WordNet)",
|
| 266 |
-
"depbank": "Danish Dependency Treebank",
|
| 267 |
-
"ep": "European Parliament",
|
| 268 |
-
"ft": "Folketinget (Danish Parliament)",
|
| 269 |
-
"gutenberg": "Gutenberg",
|
| 270 |
-
"jvj": "Johannes V. Jensen (Danish author/poet)",
|
| 271 |
-
"naat": "NAAT",
|
| 272 |
-
"opensub": "Open Subtitles",
|
| 273 |
-
"relig": "Religious texts",
|
| 274 |
-
"spont": "Spontaneous speech",
|
| 275 |
-
"synne": "Synderjysk (Danish dialect)",
|
| 276 |
-
"tv2r": "TV 2 Radio (Danish news)",
|
| 277 |
-
"wiki": "Wikipedia",
|
| 278 |
-
"wikibooks": "Wikibooks",
|
| 279 |
-
"wikisource": "Wikisource",
|
| 280 |
-
"twfv19": "Twitter Folketingsvalget 2019 (Danish election tweets)", # not present in this version of the dataset
|
| 281 |
-
}
|
| 282 |
-
```
|
| 283 |
|
| 284 |
|
| 285 |
## Additional Information
|
|
@@ -302,22 +245,4 @@ The original version of Danish Gigawords was created as a part of the following
|
|
| 302 |
booktitle = {Proceedings of the 23rd Nordic Conference on Computational Linguistics},
|
| 303 |
publisher = {NEALT}
|
| 304 |
}
|
| 305 |
-
```
|
| 306 |
-
|
| 307 |
-
|
| 308 |
-
<!--
|
| 309 |
-
Todo:
|
| 310 |
-
|
| 311 |
-
add tests
|
| 312 |
-
- unique ids
|
| 313 |
-
- valid metadata
|
| 314 |
-
|
| 315 |
-
add ci:
|
| 316 |
-
- summary statistics
|
| 317 |
-
- tables
|
| 318 |
-
|
| 319 |
-
prettify:
|
| 320 |
-
- license as independent column
|
| 321 |
-
- ensure pretty_name is standard
|
| 322 |
-
- potentially remove some columns
|
| 323 |
-
-->
|
|
|
|
| 130 |
|
| 131 |
## Dataset Description
|
| 132 |
|
| 133 |
+
This is iteration on the Danish Gigaword corpus. It is intended to be continually updated with new data sources.
|
| 134 |
|
| 135 |
### Dataset Summary
|
| 136 |
|
|
|
|
| 160 |
|
| 161 |
```py
|
| 162 |
{
|
| 163 |
+
"text": "SAMLEDE VÆRKER\n\nJEPPE AAKJÆR GYLDENDALSKE BOGHANDEL...",
|
| 164 |
+
"source": "adl",
|
| 165 |
+
"id": "adl_aakjaer06val",
|
| 166 |
+
"added": "2020-09-14",
|
| 167 |
+
"created": "1700-01-01, 2022-01-01",
|
| 168 |
+
"license": "Creative Commons Legal Code\n\nCC0 1.0 Universal",
|
| 169 |
+
"domain": "Wiki & Books",
|
| 170 |
+
"metadata": {"source-pretty": "Archive for Danish Literature"},
|
|
|
|
| 171 |
}
|
| 172 |
```
|
| 173 |
|
|
|
|
| 180 |
- `id` (`str`): An unique identifier for each document.
|
| 181 |
- `added` (`str`): An date for when the document was added to this collection.
|
| 182 |
- `created` (`str`): An date range for when the document was originally created.
|
| 183 |
+
- `license` (`str`): The license of the document. The licenses vary according to the source.
|
| 184 |
+
- `domain` (`str`): The domain of the source
|
| 185 |
- `metadata/source-pretty` (`str`): The long form version of the short-form source name
|
| 186 |
+
- `metadata/*`: Potentially additional metadata
|
| 187 |
|
| 188 |
|
| 189 |
### Data Splits
|
|
|
|
| 196 |
|
| 197 |
Below follows a brief overview of the sources in the corpus along with their individual license.
|
| 198 |
|
| 199 |
+
| Source | License |
|
| 200 |
+
| ----------------- | -------------------------------------------------------- |
|
| 201 |
+
| adl | Creative Commons Legal Code 1.0 Universal |
|
| 202 |
+
| botxt | Creative Commons Legal Code 1.0 Universal |
|
| 203 |
+
| dannet | [dannet license] |
|
| 204 |
+
| depbank | Attribution-ShareAlike 4.0 International |
|
| 205 |
+
| ep | Creative Commons Legal Code 1.0 Universal |
|
| 206 |
+
| ft | Creative Commons Legal Code 1.0 Universal |
|
| 207 |
+
| gutenberg | [gutenberg license] |
|
| 208 |
+
| hest | Creative Commons Legal Code 1.0 Universal |
|
| 209 |
+
| jvj | Attribution-ShareAlike 4.0 International |
|
| 210 |
+
| naat | Creative Commons Legal Code 1.0 Universal |
|
| 211 |
+
| relig | Creative Commons Legal Code 1.0 Universal |
|
| 212 |
+
| retsinformationdk | [Other (Danish Law)] |
|
| 213 |
+
| retspraksis | Creative Commons Legal Code 1.0 Universal |
|
| 214 |
+
| skat | Creative Commons Legal Code 1.0 Universal |
|
| 215 |
+
| spont | Creative Commons Legal Code 1.0 Universal |
|
| 216 |
+
| synne | Creative Commons Legal Code 1.0 Universal |
|
| 217 |
+
| tv2r | [Custom, Creative Commons Attribution 4.0 International] |
|
| 218 |
+
| wiki | Creative Commons Legal Code 1.0 Universal |
|
| 219 |
+
| wikibooks | Creative Commons Legal Code 1.0 Universal |
|
| 220 |
+
| wikisource | Creative Commons Legal Code 1.0 Universal |
|
| 221 |
+
|
| 222 |
+
[Custom, Creative Commons Attribution 4.0 International]: https://huggingface.co/datasets/danish-foundation-models/danish-gigaword-2/blob/main/data/tv2r/tv2r.md#license-information
|
| 223 |
+
[gutenberg license]: https://www.gutenberg.org/policy/license.html
|
| 224 |
+
[dannet license]: https://cst.ku.dk/projekter/dannet/license.txt
|
| 225 |
+
[Other (Danish Law)]: https://huggingface.co/datasets/danish-foundation-models/danish-gigaword-2/blob/main/data/retsinformationdk/retsinformationdk.md#license-information
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 226 |
|
| 227 |
|
| 228 |
## Additional Information
|
|
|
|
| 245 |
booktitle = {Proceedings of the 23rd Nordic Conference on Computational Linguistics},
|
| 246 |
publisher = {NEALT}
|
| 247 |
}
|
| 248 |
+
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|