Added preprocessing advices
Browse files
README.md
CHANGED
|
@@ -452,6 +452,15 @@ future version.
|
|
| 452 |
|
| 453 |
[Needs More Information]
|
| 454 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 455 |
## Additional Information
|
| 456 |
|
| 457 |
### Dataset Curators
|
|
|
|
| 452 |
|
| 453 |
[Needs More Information]
|
| 454 |
|
| 455 |
+
### Preprocessing advices
|
| 456 |
+
|
| 457 |
+
- Filter out empty contents.
|
| 458 |
+
- Remove unwanted references from the text, and replace either by the "references_text" or by the reference content itself.
|
| 459 |
+
- Unescape HTML special characters: `import html; html.unescape(my_text)`
|
| 460 |
+
- Remove superfluous line break in text.
|
| 461 |
+
- Remove XML tags / replace by special tokens (<italic>, <sup>, <sub>, ...).
|
| 462 |
+
- Join the items of the contents' lists.
|
| 463 |
+
|
| 464 |
## Additional Information
|
| 465 |
|
| 466 |
### Dataset Curators
|