Android Authority feed

[PhilC813]

Feed URL

https://www.androidauthority.com/feed/

Add any details, links, or screenshots about the article layout that's missing or wrong

In the following article, the name of the different sections, which are headers ("h2" elements), are stripped out in full content mode.

Article:
https://www.androidauthority.com/new-android-apps-658839/

Here's the stripped out HTML, which is rather simple:
<h2 id="[number]">[Header]</h2>

I would assume that the same would occur for other articles in the feed.

Unless I'm mistaken, header elements shouldn't be stripped out regardless of the feed; they are usually relevant to the content.

Thank you!!

[PhilC813]

I also noticed that the embedded YouTube videos are stripped out, but unlike the headers, they seem oddly integrated in the HTML as they only appear in the JSON data. I understand if this can't be improved in the parser.

"TED Tumblewords" app section:
{"resource":"nc-embed-youtube","video":{"youtubeId":"1Z9fVc6v2aY"}}

"Carrion" app section:
{"resource":"nc-embed-youtube","video":{"youtubeId":"M6NOM2-UZdw"}}

[jocmp]

Thanks for all the details. The missing headers should be simple. I'll see what I can do for that YouTube video JSON.

[jocmp]

I have a custom parser I'll test a little more before adding to the next release (jocmp/mercury-parser#27).

I haven't found a way to grab the YouTube videos - it looks like they're doing something weird with JavaScript for non-browser clients (like Capy). That said, headers are working. Stay tuned!

Before	After

[PhilC813]

Sounds promising! Thank you!

If you made a custom parser for this feed specifically, is it because you don't consider header elements to be relevant to the content for most feeds?

[jocmp]

I think headers are important. I want to avoid breaking the core parser since I don't fully understand all the different parts yet. So adding a custom fix is easiest now.

In general sites can misuse headers which is why the previous maintainers built the parser that way. Luckily they left a note on why the built it that way.

Remove any headers that appear before all other p tags in the
document. This probably means that it was part of the title, a
subtitle or something else extraneous like a datestamp or byline,
all of which should be handled by other metadata handling.

[jocmp]

Updated as of 2024.12.1085-dev

[PhilC813]

So far, articles seem rendered properly with the update to the parser! Thank you so much! It's a feed that I only check occasionally however so I'll let you know if I notice other things down the line.

---

One small thing, albeit not important;
For polls like in this article, the poll and voting choices get rendered in Capy but naturally, can't be interacted with because it must lack the JavaScript.

Website	Capy

Here's the HTML for that section:
<div class="e_e"><div class="e_Oh e_Kp"><h3 class="e_Np">Which Android XR form factor are you most excited about?</h3><div class="e_Op">1343 votes</div><div class="e_Pp e_Qh"><button type="button" class="e_Qp"><div class="e_Rp">The headset! Give me that immersion!</div><div class="e_Mp">10<!-- -->%</div><div class="e_Sp"><div class="e_Lp" style="width:10%"></div></div></button><button type="button" class="e_Qp"><div class="e_Rp">The glasses! I want that lightweight design I can use everywhere.</div><div class="e_Mp">46<!-- -->%</div><div class="e_Sp"><div class="e_Lp" style="width:46%"></div></div></button><button type="button" class="e_Qp"><div class="e_Rp">Both.</div><div class="e_Mp">14<!-- -->%</div><div class="e_Sp"><div class="e_Lp" style="width:14%"></div></div></button><button type="button" class="e_Qp"><div class="e_Rp">None. I&#x27;m already invested in another XR/VR/AR platform.</div><div class="e_Mp">3<!-- -->%</div><div class="e_Sp"><div class="e_Lp" style="width:3%"></div></div></button><button type="button" class="e_Qp"><div class="e_Rp">I don&#x27;t care about XR, AR, or VR.</div><div class="e_Mp">28<!-- -->%</div><div class="e_Sp"><div class="e_Lp" style="width:28%"></div></div></button></div></div></div>

No special element or meaningful class names that would allow you to tatget these polls in the custom parser, but after comparing with some of their other articles that contain polls, it appears that the classes e_Oh and e_Kp are the ones used to define these polls.

Should these be stripped out? Or maybe it could also be useful to the reader to know that there's actually a poll there, I'm not sure.

[jocmp]

@PhilC813 good find. Let me see if they're using JS for these. It would be nice to allow this functionality. If nothing else, I'll remove it to avoid the jarring markup.

• Remove polls in Android Authority parser #699

[PhilC813]

In this article:
https://www.androidauthority.com/best-android-apps-2024-3506812/

The title of the apps in display are stripped out in Capy. Could this be improved?

There's also another example of a poll when you'll need to confirm your results with #699!

[jocmp]

yep! I'll roll that in with the polls since the also use h3 headings. (jocmp/mercury-parser#41)

Here's a preview from my markup tester. "Google Gemini" and "Mozilla Thunderbird" were previously hidden.

[jocmp]

@PhilC813 the heading and poll updates are available as of 2025.01.1096-dev!