Code Gallery, API Explorer, other Developer Tools?

[#1]

Looks like what used to be known as "Developer Connection" (Home | Adobe Developer Connection ) has gone away.  The really helpful API Explorer has disappeared, all code samples are gone, documentation has moved to Github and is incomplete, and the developer forums appear to have gone away including all knowledge contained in them. 

I have encountered Adobe I/O  and of course some of the work on Adobe-Experience-Cloud · GitHub  will be helpful, someday.  But code samples, tips and tricks, and developer-to-developer knowledge sharing are also really important. 

Most missing, btw, is a welcoming intro for the beginning R or Python person who wants to scale up, the "junior data scientist" or "advanced analyst" who wants to expand their skills, something in writing akin to Trevor Paulsen and Tim Wilson's session at Summit Adobe Summit 2018 | The Digital Marketing Conference | March 25—29, 2018​.  They used R, but Python would also be appreciated as there are some great libraries for accessing the APIs out there as well.  

Perhaps Adobe can explain what's going to happen next to support analytic developers who rely on these tools, and provide some recommended sources or places to get help and share knowledge.  These need not be Adobe-owned, btw: links or pointers to Facebook or Slack groups, active git repositories, etc. would all be appreciated.  But if Adobe wants to keep the data scientists enthused about the power of Adobe's capabilities, they should work with the community to start building back up this support. 

I look forward to hearing what's next!

Best, Michael

[#2 · Accepted solution]

Hi Michael,

What's happening next is Analytics will release the new Analytics Reporting API in October, 2018.

The Analytics Reporting API V2 was previewed in an Adobe Summit 2018 lab session. It is the same reporting API that the Analysis Workspace UI is built on. It will be available via the Adobe I/O Console and Gateway and will support the following:

* oAuth authentication

* Automatic data-center routing

* No queueing/polling - true REST request/response

* Swagger interface and specification

Regarding the 1.4 API documentation in Github:  GitHub - AdobeDocs/analytics-1.4-apis: Documentation and helpful resources for calling the Analytics...

Please let us know if you think there is something missing from the 1.4 documentation in Github. We removed the documentation on oAuth through the 1.4 API gateway as it's shortly to become deprecated (but still supported for those callers currently calling 1.4 APIs - don't want anyone to panic that their code will start breaking in October) with the release of the V2 APIs under Adobe I/O. Other than the removal of the 1.4 oAuth documentation, all other 1.4 documentation should be in the Github repository.

Note that there is also a working Swagger interface and specification, Swagger UI​,  which can be used either directly or with a variety of publicly-available API tools to provide the same functionality as the API Explorer. It didn't make sense for Adobe to continue supporting and maintaining a custom API Explorer tool when there are so many readily-available alternatives.

We hope you enjoy using the new version of the Analytics Reporting API when it's officially released in a few months!

Thanks,

Brian

[#3]

Hi Brian,

I tried to use the Service Account integration in Adobe I/O for Adobe Analytics but it did not work as it created a technical account and somehow is not supported by the backend as it is not an Adobe ID.

Would this fix when the Analytics API v2 will be released ?

The main point of an API is to be able to use it via a script to pull some data and do some stuff and using the OAuth integration does not allow to do so as it seems that each time you request an access_token via OAuth it redirect you to login page where you need to specify your username and password.

Best regards.

Alexis Cazes

[#4]

Yeah I agree with you there - the documentation is lacking and probably will be for some time unfortunately, and the APIs are unforgiving at best - even the base tutorials are a head-scratcher.

- that being said,  I actually was able to get something going in Python (after many hours) to at least generate the headers needed to pull from the APIs if you want to use it.  I also put some resources that helped me in the post:

Better API Documentation

For R, Randy is awesome - helped me personally when I got stuck, and he built out a package that helps with API access:

https://cran.r-project.org/web/packages/RSiteCatalyst/RSiteCatalyst.pdf

For Python I liked some of this gentleman's work:

python-omniture/account.py at master · dancingcactus/python-omniture · GitHub

Unfortunately, I don't think Adobe treats R and Python as first class citizens just yet for doing the more rigorous analytics work, as they want you "in their tools" more often than not, but hopefully with the industry moving in the Python/R direction it will help pressure for well-supported features for the tools we want to use

Hope it helps, Happy Analyzing!

[#5]

Hi Alexis,

Service and technical accounts will not be supported in the October 2018 release. We will work to add support for those use cases in a future release. The initial lack of support for service and technical accounts is not strictly an Analytics limitation but is a limitation in various interoperability layers between Analytics, Adobe I/O, and Adobe Cloud Platform.

Thanks,

Brian