chore: cleanup duplicated filter logic

[panpan0000]

Closes #

Clean up duplicated code of original https://github.com/k8sgpt-ai/k8sgpt/blob/main/pkg/analysis/analysis.go#L136 and https://github.com/k8sgpt-ai/k8sgpt/blob/main/pkg/analysis/analysis.go#L161

📑 Description

✅ Checks

• My pull request adheres to the code style of this project
• My code requires changes to the documentation
• I have updated the documentation as required
• All the tests have passed

1)I've run make test make fmt make lint make build
2)then run the generated binary with below command:
k8sgpt analyze --filter=VulnerabilityReport
k8sgpt analyze
the result outputs are the same with released binary from github.

ℹ Additional Information

[panpan0000]

@matthisholleville @AlexsJones can you pls review it ?

[panpan0000]

Hi, Maintainers, when you have time, can you please kindly help to review it ?
since the code get changed very frequently , this PR conflicts from time to time T_T....

[matthisholleville]

Thank you for your contribution. Great work, left a comment.

Shouldn't add some test about filter logic ?

[panpan0000]

Thank you for your contribution. Great work, left a comment.

Shouldn't add some test about filter logic ?

T_T... @matthisholleville
I tried to add 2 of 3 UT, but the last one with active_filter requires run TrivyAnalyzer.Analyze()
I don't think I can fake a trivy client easily to finish this UT....
so I will have to comment the code for now...

[matthisholleville]

Active filters correspond to the filters added by the command k8sgpt filter add ... Why would we need Trivy?

[panpan0000]

Active filters correspond to the filters added by the command k8sgpt filter add ... Why would we need Trivy?

@matthisholleville
Per your ask, I added UT for analysis.RunAnalysis(), so there're 3 cases:

• 1 --filter specified
• 2 there's activeFilter in viper(config)
• 3 none of two above

For case 2, there's only Trivy plugin implemented. that's where my problem came from.

in analysis.RunAnalysis(), major code as below , the key is the map analyzerMap[]:

        for _, filter := range filtersCandidate {
                if analyzer, ok := analyzerMap[filter]; ok {
                                   analyzer.Analyze(..)
                }
         }

So I guess I should do below in UT:

1. add a new filter type ( just viper.SetDefault("active_filters", "newFilter"))
2. fake a dummy analyzer( maybe just return Result with kind == newFilter)
3. hack into the analyzerMap, to inject the dummy analyzer as above. But the new PROBLEM is : the additionalAnalyzerMap map is private variable in pkg/analyzer/analyzer.go, I cannot access it in my UT.

should I add a new public function to modify that private variable ?

[matthisholleville]

I don't quite understand point two. Active filters are the result of adding filters via the command k8sgpt filters add Ingress for example. This is not a feature only related to integrations. I think they are two different things.

For me there are 3 things to test:

1. The user adds --filter in his analysis command
2. We get the actives_filters (stored in the configuration file)
3. Neither of them is specified and we use the core analyzers

[panpan0000]

Thanks @matthisholleville . I thought the active_filter can only be added by integration and not knowing about k8sgpt filters add . Now it 's clear.

the PR has been updated. please kindly help to review again :-)

[matthisholleville]

There seems to be an issue with the filters. I'm not sure if it's related to this change, but it would be worth fixing it.

I don't have the HPA filter enabled, yet it is being analyzed. Could you please take a look?

[matthisholleville]

Fixed by : #431

[panpan0000]

Fixed by : #431

Hi, @matthisholleville , I think you meant to reply to PR #413 (which I will close) instead of this one ?

this PR has been updated with unit-test added per your ask. when you have time, can you review this ?

[matthisholleville]

We have just discussed your PR, and we believe that your modification adds more complexity than it resolves. Thank you very much for your time and contribution, but at the moment, we will not be proceeding with your PR.

[panpan0000]

@matthisholleville , Thanks for your time and response

• 1. Sorry, I don't catch that why the "duplicated logic" simplify will introduce complexity.
• 2. Even when the code is unnecessary, is my Unit-Test part still worthy ?

[panpan0000]

Mat explain that the filter part code is unstable now , it's better to keep it as it is.

I will create new PR for UT only.