[5] Improve test coverage in fabric8-analytics/fabric8-analytics-data-model repository

[msrb]

Description

Data importer is a very important piece in data ingestion pipeline, but @tuxdna recently discovered a lot of bugs there. Test coverage seems to be rather low. We want to add more tests, especially for parts which directly deal with data transformation and gremlin queries.

Acceptance criteria

• data importer test suite is greatly improved and our QA guy (@tisnik) gives us thumbs up
• all data importer tests run in CI, for every PR and merge to master branch
• code coverage is measured and reported as well (Run test coverage fabric8-analytics/fabric8-analytics-data-model#90)

[sivaavkd]

@msrb @tisnik - please estimate story points for this

[msrb]

I think it would be the best if somebody who is familiar with the codebase could start and add at least some tests so others can then join and help. Unfortunately, the only person in the stream that is really familiar with the codebase si Saleem.

@tuxdna Do you think you could give up one of the Sentry integration tasks and focus on this issue instead? I think one of the barriers why people tend to avoid graph ingestion is that it's very easy to break something as there are currently no (or very few?) tests. I think having the codebase covered by tests would lower the barrier significantly.

[tisnik]

@msrb 👍

[tuxdna]

Let me revisit the test setup for Data Model Importer now.

[tuxdna]

• Remove unused code fabric8-analytics/fabric8-analytics-data-model#129

[msrb]

@tisnik what's the desired threshold here?

@tuxdna do you plan to push more tests?

[msrb]

Moving to the current sprint.

[tuxdna]

• Add test for an EPV that doesn't exist fabric8-analytics/fabric8-analytics-data-model#135

• Better test environment using deployment scripts fabric8-analytics/fabric8-analytics-data-model#136

[tuxdna]

We are at 76% in Data Model Importer - https://ci.centos.org/job/devtools-fabric8-analytics-data-model-fabric8-analytics/152/console

Earlier it was 65%.

[tuxdna]

• Add tests for rest api fabric8-analytics/fabric8-analytics-data-model#137 - This is approaching 83% code coverage now.

Overall the code coverage has been improved from 24% to 83% as a part of current task.

[tuxdna]

• Add tests for graph manager fabric8-analytics/fabric8-analytics-data-model#138

[tuxdna]

• Add tests for graph populator fabric8-analytics/fabric8-analytics-data-model#139

At 87% now.

[tuxdna]

@tisnik What is the threshold for us to move past this task ?

[tuxdna]

• Fix exceptions and typos with date time parsing fabric8-analytics/fabric8-analytics-data-model#140

• Add test with valid input for graph import fabric8-analytics/fabric8-analytics-data-model#141

With above two PRs we will reach 90% code coverage.

[tisnik]

Code coverage so far:

Name                                      Stmts   Miss  Cover   Missing
-----------------------------------------------------------------------
src/__init__.py                               2      0   100%
src/config.py                                22      0   100%
src/data_importer.py                        195     32    84%   86, 97, 120-131, 158, 177-178, 215-217, 264-266, 274-275, 282-295, 307, 312, 317, 322, 327, 343, 348, 353
src/data_source/__init__.py                   0      0   100%
src/data_source/abstract_data_source.py       6      0   100%
src/data_source/s3_data_source.py            36      7    81%   29-31, 38, 43, 53, 65-66
src/graph_manager.py                         38      0   100%
src/graph_populator.py                      176      1    99%   385
src/rest_api.py                              97     21    78%   35-39, 97, 100-102, 113-115, 124, 143-157, 161
-----------------------------------------------------------------------
TOTAL                                       572     61    89%

-> ok on my side 👍

[tuxdna]

@msrb @tisnik All checks are complete, and we will keep adding more tests. Feel free to close this issue now. 👍

Thanks!

[tuxdna]

• Add one more test for invalid package insertion fabric8-analytics/fabric8-analytics-data-model#142

https://ci.centos.org/job/devtools-fabric8-analytics-data-model-fabric8-analytics/170/console