Repository navigation
Possibility of collaboration between wallabag and Nextcloud #408
Description
@nicosomb, @jogaulupeau, we from Nextcloud (http://github.com/nextcloud) are preparing to enter Google Summer of Code this year. Among the ideas is a read-it-later/Pocket/Readability app: https://github.com/owncloud/core/wiki/GSoC,-GCI-and-OPW#read-later-app-similar-to-pocket--instapaper--readability
There I wrote that it might be best for Nextcloud to work on an own solution, but screw that. I think collaboration is very important here, especially in this field where there is a lack of open source solutions while the data you give away (a lot of stuff you’re interested in) is very personal and sensitive.
So how could we collaborate? (Edited after discussion)
- Poche is working to separate their app into a server which exposes an API and a fully JS client app which could be offline-capable for use as Firefox OS app for instance.
- Poche also has an Android app and browser extensions working on that same API. Now we should discuss together to decide how the API should look like.
- For Nextcloud, we will also implement a server which works with this API, hence we could make use of the existing Android app, browser extensions and Firefox OS app.
- We can also have a front-end for the Read-later items inside Nextcloud. Probably best to combine it with the Bookmarks app.
I think the more important point is that we can share the extensions and app ecosystem, since that requires most work and polish especially when it comes to the little details (hence the bunch of design enhancement issues I filed ;)
cc @cosenal (GSoC coordinator for Nextcloud, wants a Pocket alternative as well ;)
Activity
jancborchardt commentedon Jan 21, 2014
Also cc @tcitworld for the Poche Chrome extension.
And @eMerzh @nalt for the ownCloud Bookmarks app, which could be a start for the ownCloud read-later web interface.
What do you all think?
nicosomb commentedon Jan 21, 2014
And @vjousse for poche v2.
tcitworld commentedon Jan 21, 2014
Not very related, but I had written some time ago http://mail.kde.org/pipermail/owncloud/2013-September/010726.html (I didn't find the time :( )
Maybe it can be made easier if poche is already included.
I leave @nicosomb answer all.
cosenal commentedon Jan 21, 2014
I am not familiar with the code of Poche, but here is my experience with porting apps that already exists on their own to ownCloud. When I started to work on the ownCloud News app (https://github.com/owncloud/news), the first idea was to port Lilina (http://getlilina.org/). The premises were the same as here: the use case of the app is exactly the same, it's written in PHP, it uses the same library to parse feed. Yet, it turned out that it would have been much easier to write a new owncloud app from scratch. Thanks to appframework it's going to be even easier now.
That being said, I totally agree that collaboration is useful and points 1. and 2. that @jancborchardt makes are must-do in my opinion (about 3, I am not confident there will be much to share).
nicosomb commentedon Jan 21, 2014
Wow, your idea is great.
I'm very interested in it.
What do you expect of poche team?
@vjousse and I are working on poche v2 (https://github.com/inthepoche/poche/tree/v2-silex), still in PHP, we are using Silex framework.
This version will have an API.
Since our discussion on IRC (##poche on freenode), I understand you don't want that we will create a read-it-later app for owncloud. You expect a common API. OK, I think it's doable.
Where can I find the ownCloud API documentation?
jancborchardt commentedon Jan 21, 2014
I updated the 1–4 points in the original post with what we talked about in IRC. :)
nicosomb commentedon Jan 22, 2014
@jancborchardt In your first message, you wrote "Now we should discuss together to decide how the API should look like". Here is the current API for poche v2: https://github.com/inthepoche/poche/blob/v2-silex/app/controllers/api.php Do you think ownCloud needs something special? Do you think that we need to change something?
I ask you these questions because I don't want that poche becomes dependant of something. I don't know if I'm clear.
For me, each software has its own API, each API can talk with each other. Maybe I'm wrong.
If you can explain it to me :)
cosenal commentedon Jan 22, 2014
@nicosomb The idea is the following: we want to make easier the job of those who write external software (such as browser extensions, or mobile apps) that interacts with our apps. If we have a common API, then the software they write will interface in one shot with both Poche and the ownCloud app (perhaps with only very small changes).
@jancborchardt 's point was not about ownCloud interacting with Poche through the Poche API, although that's something to consider too.
nicosomb commentedon Jan 22, 2014
@cosenal OK.
cosenal commentedon Jan 22, 2014
Here is some documentation (I don't how helpful it is): http://doc.owncloud.org/server/5.0/developer_manual/app/app/externalapi.html
As an example, here is the API of the News app: https://github.com/owncloud/news/wiki/API-1.2
Let's discuss it here: https://github.com/owncloud/bookmarks/wiki(as Jan pointed out, you guys don't have access to that)
A related thread: owncloud-archive/bookmarks#2
cc @Raydiation (creator of the ownCloud appframework)
BernhardPosselt commentedon Jan 22, 2014
You probably don't want to use the OCS api since it forces a protocol on you and there is some weirdness involved in getting put and delete parameters. Just creating a normal controller that returns json should be fine
nicosomb commentedon Jan 22, 2014
OK, thank you @Raydiation.
nalt commentedon Jan 23, 2014
I'm interested in this concept, and the application is really useful. Concerning the integration with the ownCloud bookmark app: It seems to me the requirements for the API are quite similar - so it makes sense to join forces to get a more stable API.
Yet, we should discuss how much distinction there should be between bookmarks and read-it-later from the user perspective. E.g. should there be two different frontends, different views or just a special read-it-later tag?
cosenal commentedon Jan 23, 2014
About the distinction between bookmarks and read-it-later, one big difference in the way I use the two apps has always been the following: I bookmark websites (shops from which I want to buy, nice coding tutorial websites, ...), but I read-it-later pages (a page in a newspaper, in a blog). With respect to that, the second use case is more similar to News than to bookmarks.
Btw, should we keep this discussion here or in https://github.com/owncloud/bookmarks ?
nicosomb commentedon Jan 23, 2014
When poche v2 will be released, ownCloud will create its own app. So we can have the discussion here or in github.com/owncloud/readitlater (or what you want), but not in bookmarks I think.
38 remaining items
jtojnar commentedon Apr 28, 2018
Please use GitHub’s thumbs up function. Nobody is interested in you being interested, unless your interest actually includes an implementation.
bepolymathe commentedon Apr 29, 2018
But everybody is not dev...
marcelklehr commentedon May 8, 2018
Hey everyone :)
I'm kind of late to the party, but I'll add my two cents anyway :)
I've been working on revamping the bookmarks app in nextcloud for some time. And we have of course been considering adding read-it-later functionality.
I don't see the benefit in making people install wallabag in addition to nextcloud while having a separate frontend for it in nextcloud, as long as there's no integration with nextcloud. Collabora for example integrates with nextcloud by allowing you to edit files. A wallabag nextcloud app would simply be a passthrough facade with no benefit, no?
I think the original idea of sharing a common API would be more beneficial, so as to establish a kind of standard that grows an ecosystem around it (see php-fig, etc.).
On the nextcloud side read-it-later functionality in nextcloud would fit into either the news app or the bookmarks app. I'm personally not particular about where it should be added. From a UX perspective, the concepts of the news app seem to be closer to read-it-later, as it already allows reading articles, and has collections of them. On the other hand, at its core, it's a feed reader, which fetches feeds of articles regularly and makes them available, whereas read-it-later would be on-demand fetching of arbitrary articles and assembling them in a separate collection, possibly with the need to search and organize them. This is closer to the usage concepts of the bookmarks app.
Here's an idea: We could add the wallabag API to the bookmarks app for adding and managing the pages/articles, but also add feeds for bookmark tags, which then can be integrated into the news app. How does that sound?
silverhook commentedon May 8, 2018
If the API would work in both directions – so changing a Nc bookmark would also change the Wallabag article (and vice versa) – that could make sense.
jancborchardt commentedon May 8, 2018
@marcelklehr that sounds good! So basically you could then use the Wallabag mobile apps to connect to your Nextcloud and have the content from the Bookmarks app available offline.
For Nextcloud News, there are already Android and iOS apps, so there integration is less needed. Bookmarks seems good since it's a more similar format to Wallabag. :)
@nicosomb @tcitworld what do you think?
oakbrad commentedon May 8, 2018
@marcelklehr I think simply adding RSS feeds to bookmark tags would open up a lot of options for users, I would greatly appreciate that feature. ie I use this script with my feed reader to automatically import favorited articles into Wallabag.
https://github.com/Findus23/rss2wallabag
I think the main difference between 'bookmarks' vs 'read-it-later' vs 'news reader' is the saving of pages for offline viewing/long term preserving. So it's most like bookmarks, but bookmarks don't preserve the page. So what if from a UX perspective it's a "Save Offline" or "Preserve Web Page" feature for bookmarks? This could sync to Wallabag, and vice versa. Maybe for non-Wallabag users it could simply save the page to the users files.
violoncelloCH commentedon May 8, 2018
An other possible solution I thought of could be to just make wallabag installable as a nextcloud app (like it's done with rainloop)...
the advantage would be that we can just use wallabag as it is (with all options and hopefully mobile apps too) on top of nextcloud in an easy way to install...
the downside is that possibly serving the wallabag on top of nextcloud could be difficult and wallabag wouldn't fit into the design sheme of nextcloud either
So maybe just sharing the api and building and or copying wallabag features in a standalone nextcloud app is easier
jancborchardt commentedon May 8, 2018
@violoncelloCH which is exactly why @marcelklehr suggested the integration. ;)
Shadesss commentedon May 11, 2018
Would love some wallabag on my nextcloud.
1337sup3rh4x0r commentedon May 14, 2018
Another thing to keep in mind, at least for my workflow, is the "Starred" section of nextcloud News. It would be great if "starring" an article would send it to wallabag/bookmarks/that-combined-api, or if news integraded an additional button for that.
It should however be ensured that we don't go in circles / break the matrix, when you subscribe to the bookmarks feed and star an item form it ;)
sunjam commentedon May 26, 2019
Bounty has been added for Wallabag support in Nextcloud Bookmarks app. Please consider putting some money on the bounty here to encourage devs.
edit: $35 as of 2/28/2020
nicosomb commentedon Sep 2, 2021
There is no more discussion about this possibility of collaboration. I close this issue.
menelic commentedon Nov 18, 2024
this would still be a great integration, with omnivore having closed and wallabag offering omnivore import, this would be a great time to finally get Nextcloud users a full read it later solution
ChimeraFire commentedon May 23, 2025
And Pocket gone now too... This is needed.