Date   

API for database access

Andrew Theaker <andrew@...>
 

Hi,

We are currently thinking about upgrading our group to premium and I'm trying to understand what benefits we could get from the API access.

The document is not really clear on access to a groups databases. We have a catalogue of all board games we own. It would be great if I could access this via API and render it in a nice view. I was thinking we could also create a database to tracking our results from events (wins/losses/attendance etc) and extract this information for compiling stats in graphs etc. Ideally I'd like to implement these things through a javascript front-end and just query the API as required.

Would these be possible with the current version of the API?


Re: Issue with /v1/updatesub

Eric Searcy
 

Thanks! (confirmed on my end, too :-) )

-Eric

On 4/25/19 9:30 AM, Mark Fletcher wrote:
Hi Eric,

It should now accept sub_id. If group_id is passed in as well, it will
be ignored.

Thanks,
Mark

On Thu, Apr 25, 2019 at 8:56 AM Eric Searcy <eric@linuxfoundation.org
<mailto:eric@linuxfoundation.org>> wrote:

Hi,

I'm calling updatesub as described in the documentation, and I'm
getting an error.

curl -v -u $GROUPIO_TOKEN: -d sub_id=##### -d sub_notify=sub_notify_none

=> {"object":"error","type":"bad_request","extra":"group_id"}

(Where ##### is the returned "id" field of the membership info
object when I ran /v1/getsub against a group_id I'm subscribed to.)

I can work around the error by adding group_id to the request. 
Subscriptions seem to be globally unique identifiers, so the way the
API documents it seems like it should work.

Two questions:

1) Can either the docs be updated, or the API be changed to match
the docs?

2) If you change the API to match the docs, can you ensure that any
passed group_id is just ignored (so it doesn't break backwards
compatibility).

Thanks!

-Eric


Re: Issue with /v1/updatesub

 

Hi Eric,

It should now accept sub_id. If group_id is passed in as well, it will be ignored.

Thanks,
Mark

On Thu, Apr 25, 2019 at 8:56 AM Eric Searcy <eric@...> wrote:

Hi,

I'm calling updatesub as described in the documentation, and I'm getting an error.

curl -v -u $GROUPIO_TOKEN: -d sub_id=##### -d sub_notify=sub_notify_none

=> {"object":"error","type":"bad_request","extra":"group_id"}

(Where ##### is the returned "id" field of the membership info object when I ran /v1/getsub against a group_id I'm subscribed to.)

I can work around the error by adding group_id to the request.  Subscriptions seem to be globally unique identifiers, so the way the API documents it seems like it should work.

Two questions:

1) Can either the docs be updated, or the API be changed to match the docs?

2) If you change the API to match the docs, can you ensure that any passed group_id is just ignored (so it doesn't break backwards compatibility).

Thanks!

-Eric


Re: Recommend cleaning up quoting of non-string values in API docs

 

On Thu, Apr 25, 2019 at 8:49 AM Eric Searcy <eric@...> wrote:
The groups.io API doesn't appear to use quotes around numbers and booleans.  At the very beginning of https://groups.io/api (under Pagination) it shows a correct response, however all subsequent sections (objects and the various calls that use them) show these numbers incorrectly quoted, not how they actually are presented by the API.

Good catch. It should be fixed now.

Thanks,
Mark 


Issue with /v1/updatesub

Eric Searcy
 

Hi,

I'm calling updatesub as described in the documentation, and I'm getting an error.

curl -v -u $GROUPIO_TOKEN: -d sub_id=##### -d sub_notify=sub_notify_none

=> {"object":"error","type":"bad_request","extra":"group_id"}

(Where ##### is the returned "id" field of the membership info object when I ran /v1/getsub against a group_id I'm subscribed to.)

I can work around the error by adding group_id to the request.  Subscriptions seem to be globally unique identifiers, so the way the API documents it seems like it should work.

Two questions:

1) Can either the docs be updated, or the API be changed to match the docs?

2) If you change the API to match the docs, can you ensure that any passed group_id is just ignored (so it doesn't break backwards compatibility).

Thanks!

-Eric


Recommend cleaning up quoting of non-string values in API docs

Eric Searcy
 

The groups.io API doesn't appear to use quotes around numbers and booleans.  At the very beginning of https://groups.io/api (under Pagination) it shows a correct response, however all subsequent sections (objects and the various calls that use them) show these numbers incorrectly quoted, not how they actually are presented by the API.

This is an important distinction for implementation in some languages (like using Go's json.Unmarshal).  Fixing will help any future people from running into extra work discovering the API doesn't match the docs.

Thanks!


Re: Impossible to download archives from cloudfoundry+cf-dev

Luis Cañas-Díaz
 

This is just a reminder.

Can u shed some light on this issue Mark? Our customers are waiting for their data :(

Thanks,
Luis.


Re: Impossible to download archives from cloudfoundry+cf-dev

Luis Cañas-Díaz
 

Hi again Mark,

does the test we included above (the one with curl) shed some light on the issue?


Re: Impossible to download archives from cloudfoundry+cf-dev

Luis Cañas-Díaz
 

Hi again Mark,
we've been doing several tests playing with the timeouts both with our tools and with standard tools such as 'curl'. None of them changed the result :(

Doing a test with curl for the group onap+onap-discuss:

curl "https://api.groups.io/v1/downloadarchives?group_id=19846" -u **** --output /tmp/xxx

we also get the error:

 

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  234M    0  234M    0     0   796k      0 --:--:--  0:05:01 --:--:--  588k
curl: (18) transfer closed with outstanding read data remaining

 


Re: Impossible to download archives from cloudfoundry+cf-dev

Luis Cañas-Díaz
 

We are having a look during today/tomorrow. Thanks for the info!


Re: Impossible to download archives from cloudfoundry+cf-dev

 

On Fri, Feb 15, 2019 at 3:51 AM Luis Cañas-Díaz <lcanas@...> wrote:

we are seeing this error with the following groups during the past week:

They are consistent and the last time I saw them was 5 minutes ago with a local test I just did with the group cloudfoundry+cf-dev

I checked one of the errors, and it was 'upstream prematurely closed connection while reading'. Can you look at increasing your network timeouts?

Thanks,
Mark 


Re: Impossible to download archives from cloudfoundry+cf-dev

Luis Cañas-Díaz
 

Hi Mark,

we are seeing this error with the following groups during the past week:

  • https://groups.io/g/cloudfoundry+cf-dev
  • https://groups.io/g/onap+onap-discuss
  • https://groups.io/g/OPNFV+opnfv-tech-discuss
  • https://groups.io/g/cloudfoundry+cf-dev

They are consistent and the last time I saw them was 5 minutes ago with a local test I just did with the group cloudfoundry+cf-dev

Thanks for your help.


Re: Impossible to download archives from cloudfoundry+cf-dev

Kenny Paul
 

Hi Mark,
I believe in ONAP it may be limited to one list (onap-discuss) looking at the info I'm seeing on the dashboard.


Re: Impossible to download archives from cloudfoundry+cf-dev

 

Hello,

On Thu, Feb 14, 2019 at 4:58 AM Luis Cañas-Díaz <lcanas@...> wrote:

we are trying to get data on behalf of some projects such as ONAP and Cloudfoundry and we got these errors. If you can clarify whether this is useful or not we will appreciate it.

The errors we get are quite similar to the ones my workmate included in the message above.


Are you seeing these errors with specific groups or all groups you try? Are they intermittent errors or are they consistent? When was the last error you saw?

Thanks,
Mark 


Re: Impossible to download archives from cloudfoundry+cf-dev

Luis Cañas-Díaz
 

Hi,
we are trying to get data on behalf of some projects such as ONAP and Cloudfoundry and we got these errors. If you can clarify whether this is useful or not we will appreciate it.

The errors we get are quite similar to the ones my workmate included in the message above.

Thanks in advance.


Impossible to download archives from cloudfoundry+cf-dev

valcos
 

Hi,

I'm trying to download the archives of the group cloudfoundry+cf-dev using the downloadarchives endpoint , however I get the error below, which seems to be related to a server issue:
RuntimeError: ('Connection broken: IncompleteRead(34069340 bytes read)', IncompleteRead(34069340 bytes read))
Is there any solution to get this archive?

Cheers,
Valerio


Re: Getting archives from group-specific domains

valcos
 

Thank you very much Mark! This change in the API helped me a lot!

Cheers,
Valerio


Re: Getting archives from group-specific domains

 

HI Valerio,

I just changed it so that getsub and getsubs will now return group_name. Hopefully this will let you proceed.

Cheers,
Mark

On Wed, Oct 10, 2018 at 2:48 AM valcos <valcos@...> wrote:
Hi Mark,

Thank you for answering! I'm not sure I've understood your question. I'm going to comment below the steps I did.

- I manually subscribed to 6 groups (e.g., beta | main, beta | api, announce@..., main@...)

- I got my access token --> curl "https://api.groups.io/v1/login" -u 123456: -d "email=xxx@xxx&password=yyy"

- I retrieved my subscriptions --> curl "https://api.groups.io/v1/getsubs" -u ...
  the output contains all subscriptions, but the attribute group_name is not returned. However, this is coherent with https://groups.io/api#get-subscriptions

- I retrieved the subscription for a specific group id --> - curl "https://api.groups.io/v1/getsub?group_id=1" -u ...
  the output contains only one subscription, the attribute group_name is not returned, this is not aligned with https://groups.io/api#get-subscription

As possible enhancements of the API:
- Is it possible to include the group_name in what is returned by the endpoint getsub?
- It would be possible to include domain as parameter of the endpoint downloadarchives, thus one could download archives using the params group_name and domain (instead of using the group_id)?

Thank you,
valerio


Re: Getting archives from group-specific domains

valcos
 

Hi Mark,

Thank you for answering! I'm not sure I've understood your question. I'm going to comment below the steps I did.

- I manually subscribed to 6 groups (e.g., beta | main, beta | api, announce@..., main@...)

- I got my access token --> curl "https://api.groups.io/v1/login" -u 123456: -d "email=xxx@xxx&password=yyy"

- I retrieved my subscriptions --> curl "https://api.groups.io/v1/getsubs" -u ...
  the output contains all subscriptions, but the attribute group_name is not returned. However, this is coherent with https://groups.io/api#get-subscriptions

- I retrieved the subscription for a specific group id --> - curl "https://api.groups.io/v1/getsub?group_id=1" -u ...
  the output contains only one subscription, the attribute group_name is not returned, this is not aligned with https://groups.io/api#get-subscription

As possible enhancements of the API:
- Is it possible to include the group_name in what is returned by the endpoint getsub?
- It would be possible to include domain as parameter of the endpoint downloadarchives, thus one could download archives using the params group_name and domain (instead of using the group_id)?

Thank you,
valerio


Re: Getting archives from group-specific domains

 

Hi Valerio,

Is there any reason why you can't be subscribed to the group ahead of time? Then you could get your subscriptions, which contains the group id.

Mark

On Tue, Oct 9, 2018 at 9:31 AM valcos <valcos@...> wrote:

Actually the fact that an organization has its own domain is not a problem, if you pass the group_id to the endpoint downloadarchives (note that the attribute view_archives has to be enabled in that group permissions). The procedure to download an archive is the following:

The problem I'm having now is how to download the archives if I have only the group_name (and not the group_id). The endpoint getsub, getsubs don't return the group name, and the method getgroup doesn't work if the attribute manage_group_settings is set to false.

Any hints are welcomed.

Cheers,
Valerio

141 - 160 of 232