internetarchive / openlibrary

One webpage for every book ever published!
https://openlibrary.org
GNU Affero General Public License v3.0
5.1k stars 1.33k forks source link

Create /browse view to render custom IA query results on OpenLibrary #602

Open mekarpeles opened 6 years ago

mekarpeles commented 6 years ago

Requires #2070:

Right now, many of the new carousels on our homepage pull their data from archive.org (because its elasticsearch allows us to query based on only available book). But the links we display (to show more results) take the user to archive.org (not a page on openlibrary.org) which is jarring.

We should use #2070 to make a /browse experience (powered by the /browse.json endpoint which is backed by Archive.org's ElasticSearch) which allows patrons to browse available books via archive.org ElasticSearch query results using openlibrary.

e.g. a url like https://openlibrary.org/browser?query=creator:(Dahl,Roald) OR creator:(Clancy,Tom)

Details

We could create a new infogami page (for managing "presets") which maps e.g. 50 top/featured/popular subject queries to their corresponding Archive.org ElasticSearch queries (like we do for our homepage carousels:

https://github.com/internetarchive/openlibrary/blob/master/openlibrary/plugins/openlibrary/home.py#L28-L33

mekarpeles commented 6 years ago

cc: @jeansaung, @bernatfortet :)

LeadSongDog commented 6 years ago

It's not just jarring, it is outright dysfunctional. Many network group policies block all access to IA as a potential attack vector/security threat/gaming site/etc. This means that apparently OL-local links fail. e.g. at https://openlibrary.org/works/OL15344038W/L'_economia_delle_citt%C3%A0 "You might also like" links to https://archive.org/search.php?query=collection:(inlibrary)%20AND%20loans__status__status:AVAILABLE%20AND%20openlibrary_work:(*)%20AND%20(subject:%22Industries%22%20OR%20subject:%22Urban%20economics%22%20OR%20subject:%22Social%20aspects%22%20OR%20subject:%22Economic%20development%22%20OR%20subject:%22Social%20aspects%20of%20Industries%22)%20AND%20!openlibrary_work:(OL15344038W)

Additionally, the IA image content is not suitable for low bandwidth users (such as third world mobiles) or low-vision/handsfree users of screenreaders. It might even violate the A.D.A.

The OL wiki functionality should stand on its own. Serving images from IA is purely a secondary function.

mekarpeles commented 6 years ago

603

We may make this a primary focus of the hackathon. We'd love designers to join us for the Hackthon, either remotely or in person in SF - Internet Archive (300 Funston) @ 11am.

Goals:

mekarpeles commented 6 years ago

cc @rchrd2, this is related to #630

jdlrobson commented 5 years ago

Is this still a problem? I'm not quite understanding the issue here. I'm not seeing an links to archive.org

xayhewalo commented 4 years ago

@mekarpeles can you elaborate on Jon's point? The home page doesn't link to the Internet Archive currently. I'm not sure a /browse feature would add much more value.

jimchamp commented 3 years ago

To clarify, most carousel header links ("Books We Love", "Recently Returned", etc) will send patrons to an Internet Archive search results page. From this search page there doesn't appear to be a path back to Open Library (aside from the browser's back button). To my mind this is unexpected behavior, and a browse page would improve user experience.

SaravgiYash commented 3 years ago

@jamesachamp If the link opens in a new tab, it might help in improving the user experience (for the time being)

tfmorris commented 3 years ago

To clarify, most carousel header links (...) will send patrons to an Internet Archive search results page. From this search page there doesn't appear to be a path back to Open Library (aside from the browser's back button). To my mind this is unexpected behavior, and a browse page would improve user experience.

Yup, that sucks, as @LeadSongDog noted over 3 years ago -- and even for the individual search results, the link back to OpenLibrary will be "below the fold" (ie collapsed).

Another aspect of IA based searches is that they don't use any OpenLibrary metadata, so, for example, authors are considered the same if they are lexical match rather than if they actually have the same OpenLibrary author record.

LeadSongDog commented 3 years ago

Well, to be fair, it has only been unblocked since Apr 2019 when #2070 closed and #2071 merged. Still it would be nice to get this cleared up.