Hi Jimmy, Daniel, Catherine, Stephen, Brian,

Based on Catherine's presentation at 
https://wiki.onap.org/display/DW/PTL+2019-02-19 the AAI team has started to add 
more robot testcases into the testsuites repository for CSIT-style regression 
testing of each releases' use cases.
See also https://jira.onap.org/browse/AAI-2184 and 
https://jira.onap.org/browse/AAI-2208

For AAI, that means wanting to:

-        detect the presence/absence of schema elements and EdgeRule 
relationships as they change from version (n) to (n+1)

-        confirming that these changes stay functional for subsequent releases

-        being able to manipulate individual objects/sub-objects and 
relationships in a flexible and general-purpose manner

-        (eventually) achieve 100% coverage of schema elements and relationships

-        (perhaps) testing performance of AAI services across the board (some 
problem use cases have been identified in 
https://wiki.onap.org/display/DW/2019-04-04+AAI+Developers+Meeting )

The current templates and robot keywords in testsuite repository seems to be 
focussed on the demo scenario and healthchecks, to the point of having many 
hard-coded values, sub-objects and relationships, as well as being stuck on API 
v11 schema and minimal coverage of the bare necessities.

In the first gerrit review https://gerrit.onap.org/r/#/c/79583/ there was a bit 
of discussion as to whether the testsuites repository is the right place to put 
such things. The feeling was that the tests are only run on the latest release 
and don't need to test for regressions. But eventually it was accepted and 
merged.

In the most recent gerrit review https://gerrit.onap.org/r/#/c/84195/ we get a 
-2 on working code (that also doesn't break existing code) because it didn't 
reuse hard-coded demo testcases. We don't want to change the demo, but I think 
the demo code should be refactored so that demo-specific parts are separate 
from general-purpose reusable common base code.

For refactoring to happen, we still need more foundations that allow for 
manipulating sub-objects and relationship-lists individually, as well as being 
able to compose the relationships between schema elements in the same way they 
are modelled in the AAI services. We need to avoid the problems inherent in 
number of REST URL combinations as enumerated in the swagger-generated API 
documents, while still being able to test every part of the schema.

There are still more steps to be taken. We are still just setting the stage. 
There are no backwards steps in these commits.
We want to be a help rather than a hindrance to the ONAP projects. We do not 
want to be swimming against the current.

Right now, feels like there's a fraction too much friction...

So the questions are:

-        what is the architectural direction for testsuite repository?

-        Is the testsuites/aai sub-directory a place that the AAI team can use 
as required?

-        Should these AAI-specific CSIT-style robot testcases go into an AAI 
repository instead of the testsuite repository?

-        Do other ONAP projects need/want to share in the robot scripts that 
AAI team is writing?


Thanks,
Keong


-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.

View/Reply Online (#16478): https://lists.onap.org/g/onap-discuss/message/16478
Mute This Topic: https://lists.onap.org/mt/30919602/21656
Group Owner: [email protected]
Unsubscribe: https://lists.onap.org/g/onap-discuss/unsub  
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to