I may be missing it, but is there an API endpoint or a methodology to update existing topics with hashtags? A group member has bravely (foolishly?) volunteered to work on classifying topics transferred from a Yahoo! group, but the prospect of having to manually update some 16,000 topics is daunting to say the least. I'd hoped to be able to develop some tooling to automate at least some of the process.

Regards,

Derek   

On Tue, Sep 7, 2021 at 01:04 PM, Mark Fletcher wrote:

On Tue, Sep 7, 2021 at 8:57 AM <groups@...> wrote:

I already searched the group and found one question similar to mine, but it was unfortunately unanswered. :-(

I am specifically trying to set "sort_field" in the "Search Archives" API call. Without setting it, the returned topics list is I suppose, in an order of "relevance". I want the order to be chronological.

This should be fixed now. You can use `relevance` or `created`.

Sweet -- I just tried it and it works! (And I see you've updated the docs as well. Thank you!)

-Dave

Ok, got it. The token is not a variable to pass, but rather the username. Thanks.

It would be helpful to future folks to probably provide this information in the API documentation. :-)

Thanks again.

Hi All,

I already searched the group and found one question similar to mine, but it was unfortunately unanswered. :-(

I am specifically trying to set "sort_field" in the "Search Archives" API call. Without setting it, the returned topics list is I suppose, in an order of "relevance". I want the order to be chronological.

I have reviewed the API documentation thoroughly, and "sort_field" is not defined consistently,

The usual definition for "sort_field" is: "The field to sort on. Valid values are dependent on the call." This is true for "Search Archives".

I have tried the following parameters for "sort_field" in the "Search Archives" API call, but unfortunately, none seemed to work: "id, created, subject, title, name, order". In all instances, the return messages were "sort_field": "relevance".

Is "sort_field" parameter simply un-implemented in the "Search Archives" API? Is there any chance it will be implemented, or will I simply have to perform my own sort on the returned objects?

Thanks,

-Dave

On Tue, Sep 7, 2021 at 04:37 AM, Derek Milliner wrote:

Meant to add to that that the example in the API docs (https://groups.io/api#login) is a bit misleading on that - it shows a cookie-based request with a token-based response :-)

Derek,

Thank you so much. That explains my confusion -- if the documentation is mixed up, that provides a shaky foundation to build on.

So I followed your example adding the "token=yes" POST parameters and was able to get a token back. The next thing I would like to do is to get a group's topics using token-based auth. The example in the documentation is:

curl "https://groups.io/api/v1/gettopics?groupid=2" \
-b "cookies.curl"

The documentation example above looks like a cookies-based auth request. How do I convert that into a token-based request?

For example, this below results in an "unauthorized" error:

dave@Moonquake:~ $ curl "https://groups.io/api/v1/gettopics?group_name=<group_name>" -d token="<redacted>"
{"object":"error","type":"unauthorized_error","extra":""}

 

*pulls out hair*

Likewise I tried this:

dave@Moonquake:~ $ curl "https://groups.io/api/v1/gettopics?group_name=<group_name>" -a 123456: -d token="<redacted>"
{"object":"error","type":"unauthorized_error","extra":""}

What is weird is I can get the cookie-based auth working in Postman (but not token-based auth) -- not sure where the differences is between Postman and curl, honestly.

I appreciate all your assistance with me figuring out how to use the API.

-Dave

Meant to add to that that the example in the API docs (https://groups.io/api#login) is a bit misleading on that - it shows a cookie-based request with a token-based response :-)

By default the login uses cookie-based so the 'token' as such is in the cookie. If you add 'token=true' to the request (so "email=&password=&token=true") you'll get the token stanza in the return:

I would like to install the program on drive D: It defaults to Drive C:.   My  reason for wanting to install to drive D: is that my drive C: is an SSD and is very small... only 115 GB.  Windows 10 has (over time) eaten up all of the free space and now I can't even up date windowz for a lack of free space.  My Drive D: is an HHD, and has a terabyte of storage.  I have been able to move most of my programs over to it but for some reason when I try to install N3FJP it doesn't give me an option to which drive I want to install it on.  I have searched all of the FAQs and can't find anything.   This may sound stupid but I just thought I would ask.  Back before the Plandemic, when work was good and steady, I would just buy a new computer or a larger SSD but money is tight and I was hoping for a less costly solution.  Thanks

Mark, did the ability to add a description to the picture ever get resolved?

Thanks,

Eric

Hello, all:

I am in charge of the lore.kernel.org service, which provides
developer-friendly mailing list archives for easy retrieval and
processing. It is of great importance to many developers to be able to
get access to messages as they were sent out by original authors, in
order to be able to verify DKIM signatures of the originating domain.

Unfortunately, this is not currently possible with groups.io when
using regular email subscriptions, so I am looking for a way to
retrieve messages directly via API or similar means. I know that these
raw messages are preserved somewhere on the backend, as I can get them
via an "archive export" zip files.

Anything you can suggest?

Regards,
--
Konstantin Ryabitsev
Director, IT Core Projects
The Linux Foundation
Montréal, Québec

Hello,

I am working on an application that makes use of the Groups.io API, and I have run into an issue I thought would be worth asking about here. There is a group we manage that when trying to get the group settings by calling https://groups.io/api/v1/getgroup?group_id=12184 returns an empty privacy setting of "", which is not what it should be; in the Groups.io UI, the privacy visibility setting is set to "sub_group_privacy_unlisted_public_archives" (Group not listed in parent group, publicly viewable messages). Any ideas as to why the API is returning an empty privacy setting would be much appreciated!

Thanks,
Andres

On Fri, May 28, 2021 at 02:16 PM, David Mazur-Goulet wrote:

Can we have two or three different applications interface with groups.io? Is it technically possible with the API?

Short answer; yes. Slightly longer answer, the API requires a valid login so an application will need to make the appropriate login calls and store/provide the resulting login token, but each login is distinct. You could if you wanted provide different login accounts for each application if you wanted to apply different permissions. A distinct login counts as a member, this may be an issue with the basic accounts for instance.

Is it possible to sync the directory of a subgroup with another application using the API?

If you mean as in extract the member directory of a subgroup, again yes, this would use the getmemberdirectory API call. Note that exactly what you can do through the API is controlled mainly by the permissions granted to the account you use for the API login, (in this case the "view_member_directory" permission), hence the comment above vis. different per-app accounts. Not all functionality is exposed through the API as yet; the reference docs can be found here: https://groups.io/api

Cheers

Derek