To reproduce the bug:
```bibtex
@inproceedings{Vaswani2017AttentionIA,
title = {Attention is All you Need},
author = {Ashish Vaswani and Noam M. Shazeer and Niki Parmar and Jakob Uszkoreit and Llion Jones and Aidan N. Gomez and Lukasz Kaiser and Illia Polosukhin},
booktitle = {Neural Information Processing Systems},
year = {2017},
doi = {10.48550/arXiv.1706.03762},
altmetric = {21021191}
}
```
The bug is
1. It seems to be some weird property of the liquid template that [line
252-254](8d82670ff1/_layouts/bib.liquid (L252-L254))
doesn't work at all. According to [this
post](https://stackoverflow.com/questions/59887447/liquid-how-to-assign-the-output-of-an-operator-to-a-variable)
and [this issue](https://github.com/Shopify/liquid/issues/236), liquid
doesn't support assign the output of operator to a variable nor a
ternary operator. So based on my console log, the value of
`entry_has_altmetric_badge` is always a string value of
`entry.altmetric` when altmetric is provided in bibtex.
```liquid
{% assign entry_has_altmetric_badge = entry.altmetric or entry.doi or entry.eprint or entry.pmid or entry.isbn %}
{% assign entry_has_dimensions_badge = entry.dimensions or entry.doi or entry.pmid %}
{% assign entry_has_google_scholar_badge = entry.google_scholar_id %}
{% if entry_has_altmetric_badge or entry_has_dimensions_badge or entry_has_google_scholar_badge %}
<div class="badges">
{% if site.enable_publication_badges.altmetric and entry_has_altmetric_badge %}
<span
...
```
Note that this could be problematic that a string in liquid is always
evaluated as true as long as it is defined regardless if it is "" or
"false".
[reference](https://shopify.github.io/liquid/basics/truthy-and-falsy/)
2. when altmetric is defined in bibtex, now the order of set attribute
to badge is eprint > doi > altmetric id > pmid > ISBN, and the badge
doesn't work when an arxiv doi is provided.
I think the expected behavior should be
1. as documented in CUSTOMIZE.md, only render the badge when the entry
is set to either "true" or the altmetric id. (It could also implement to
always render the badge whenever doi or other related attribute is set,
and set altmetric to "false" to disable it)
```md
- `altmetric`: Adds an [Altmetric](https://www.altmetric.com/) badge (Note: if DOI is provided just use `true`, otherwise only add the altmetric identifier here - the link is generated automatically)
```
2. if the almetric id is set, use it first.
This PR fixes a very niche bug.
If there is a co-author that is not *me* (=`<em>` tagged) and does not
have a `coauthor_url` (=`<a>` tagged), there will be a leading
whitespace before the separating comma. If the author list is split into
multiple lines (due to long author lists or a small view), the
whitespace can lead to a line beginning with a comma.
As I said, very niche, but interestingly, the other cases already did
the same whitespace handling. Just the case not-me/no-url was mising.
Currently, we do not have this case on vanilla al-folio, but you can see
it for example on [my website](https://christianmainka.de/publications/)
if the view is small (e.g., 430px).
Related to #1502.
As discussed in #1181, I suggest to make embedding videos an optional
feature.
This behavior aligns well with recently merged PR #2312.
Open questions:
1. I added a youtube link to `papers.bib`. Is this link okay?
2. I set `enable_video_embedding: false` as the default. I argue that
privacy settings should be the default. Also, the current implementation
of `video.liquid` only works for some very specific video URLs. For
example, to embed youtube, specialized links must be used to avoid
`X-Frame-Option` issues. This behavior can lead to a broken embedding,
which would not look very nice.
Feedback welcome.
I added further filds to `bib.liquid`.
A common field is `location`, which is now printed.
I added fields for `acceptance` and `acceptance_rate`. These fields
might be usefull for external people to better determine a vanue's
quality.
`acceptance` is used for the accepted vs submitted papers.
`acceptance_rate` is a percent-value.
<img width="785" alt="Bildschirmfoto 2024-04-07 um 18 11 49"
src="https://github.com/alshedivat/al-folio/assets/1998723/d5eaeb08-7f67-4fb7-b9b9-fd5f4a3c331c">
The existence of two distinct variables stems from the fact that the
'acceptance_rate' is represented using the venue color. On my website, I
employ venue colors correlated with computer science conference
rankings, which naturally leads to associating the acceptance rate with
the same color scheme for coherence and better understanding.
<img width="785" alt="Bildschirmfoto 2024-04-07 um 18 14 54"
src="https://github.com/alshedivat/al-folio/assets/1998723/33ef0419-00ee-4fea-93d9-18c0269188de">
You can find more examples on my
[website](https://christianmainka.de/publications/peerreviewed).
Also, I added a CVE Score, which is mainly usefull for IT-Security
researchers, see this [example](https://christianmainka.de/cve/).
***
Let me know what you think about this proposal.
This PR adds an "award" button to publications.
It takes the `award` value from the bibtex entry and displays(incl.
Markdown rendering) the text in a box similarly to abstract and bibtex.
User can set the entry `award_name` to configure the value. See example
config with `award_name: Nobel Prize`.
The color of the award box can be configured in `_base.scss`.
Note, there is a similar PR #2175, it I saw to issues with:
1. There was no progress
2. The award button just prints the text directly in the button,
similarly to `award_name`. Long award names could clutter the webpage.
3. IMHO, it brokes the current al-folio design, since butons do have a
fixed size/text. However, variable prize names are also possible with
this PR.
***
Pictures:
Default. Text are hidden:
<img width="708" alt="grafik"
src="https://github.com/alshedivat/al-folio/assets/1998723/1221c82c-c384-4297-807e-39385e2ce4fd">
Additional info is shown when the button is clicked. Markdown supported.
<img width="684" alt="grafik"
src="https://github.com/alshedivat/al-folio/assets/1998723/2354aeee-12b0-4d32-b194-5d2ea80d8363">
Only one text box shown at the same time, like it is with "ABS" and
"BIB":
<img width="691" alt="grafik"
src="https://github.com/alshedivat/al-folio/assets/1998723/d3937bb9-d9c2-47ac-b819-b92aec3d916a">
***
Feedback welcome.
You can also check [my
website](https://christianmainka.de/publications/awarded), which was the
base for this PR.
Our score went down a little bit after the last few changes. Avoiding
lazy loading some images (as introduced in #2183), since [it is strongly
recommended to omit hero images and other images or iframes that are
likely to appear above the fold from being
lazy-loaded](https://web.dev/articles/browser-level-lazy-loading-for-cmss#avoid_lazy_loading_above-the-fold_elements).
Also added missing `alt` to google scholar field.
---------
Signed-off-by: George Araujo <george.gcac@gmail.com>
Added GitHub action to check for broken links in repo before commit,
ignoring files with liquid tags since these were not yet processed and
also README since some sites created from this template doesn't have
valid urls anymore. Fixed some broken links in repo.
---------
Signed-off-by: George Araújo <george.gcac@gmail.com>
summary:
- adds prettier formatter configuration
- formats the entire repo using prettier, ignoring minified files
(`*.min.css`) and heavy generated html
- changes extensions of all `.html` files to `.liquid`, which is more
correct and necessary for prettier to work correctly
- replaces "%-" and "-%" with just "%" — manual liquid formatting using
minus signs is superfluous since we are compressing and minifying the
code anyway
- adds CI action for running prettier check on PR and pushes to master
ariseus
CheariX
Jake Nabasny, PhD
George
Ethan Olesinski
Maruan