acl-org / acl-org/acl-anthology
Missing paragraph breaks in rendered abstracts
- Dominant language
- Python
- Stars
- 796
- Forks
- 408
- Avg merge
- 3d 13h
- Merged PRs (30d)
- 34
Description
### Confirm that this is a bug report
- [x] I want to report an issue that does _not_ concern paper or author metadata.
- [x] I have searched for similar existing issues first.
### Problem Description
Once #8250 was deployed with updated metadata for a paper I co-authored, I noticed that "[Remember what you did so you know what to do next](https://aclanthology.org/2023.findings-emnlp.104/)" is missing the paragraph break that should be present. Instead of displaying a new paragraph, two sentences weren't even separated by whitespace! Here are the PDF and the HTML, respectively:
The simplest solution to this is to replace each newline with a space, instead of the empty string, as is currently being done at [`python/acl_anthology/utils/text.py:115`](https://github.com/acl-org/acl-anthology/blob/0bc9cabfd987c66629e0d76e96aec2c8fef988a4/python/acl_anthology/utils/text.py#L115). The effect of this was unclear to me, so I created #8551 to make metadata processing deterministic and generated diffs with the simple change applied. Unfortunately, the changes touched some very large files, so I was only able to diff a sample of the files:
- Most files I looked at are like [diff_newline_handling_P16.html](https://github.com/user-attachments/files/28573220/diff_newline_handling_P16.html) and [diff_newline_handling_E17.html](https://github.com/user-attachments/files/28573244/diff_newline_handling_E17.html), where sentences are now properly separated.
- I noticed one weird case in [diff_newline_handling_2025.jeptalnrecital.html](https://github.com/user-attachments/files/28573223/diff_newline_handling_2025.jeptalnrecital.html), which has newlines in a URL that are now incorrectly converted into spaces. It's probably better to just fix metadata with bad newlines when people notice the problem.
If you think this is a reasonable change, I'll create a PR with my patch.
I personally feel that a better solution is needed to preserve the original semantics. I'm not sure how to do this properly, but some sort of "paragraph break" element can be added to the existing XML formatting syntax. This might be difficult to insert automatically, but it can at least be documented for new abstracts and fixed by people who notice inaccurate abstract formatting.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.