Improve new CLI testing ensuring complete coverage of arguments cases

[parthosa]

Fixes #562.

1. Changes:

1. Adds unit test to verify the behavior and decisions of the new CLI based on user's input arguments.
2. Ensures that tests have covered all possible states of the platform, cluster, event logs fields.

2. New CLI User Input:

ascli --platform <platform> --cluster <cluster> --eventlogs <eventlogs> [remainings args..]

3. Possible States:

• platform:undefined or actual value.
• cluster: undefined, cluster name, cluster property file or auto tuner file.
• event logs: undefined or actual value.

4. Cases Covered:

We have a total of 16 cases (2 X 4 X 2). However, among these, there are only 8 unique test cases. Each of the 8 test case has an argument (platform or event logs) for which the result (pass or fail) remains same irrespective of the argument's value. Thus, each of the 8 test cases has two sub-cases: (a) argument is provided, and (b)argument is not provided.

For example, consider the user input where cluster argument is not provided but event logs are provided. It should pass whether the platform is provided or not.

List of Test Cases:

Platform	Cluster	Event Logs	Result
Does Not Matter	Undefined	Undefined	Fail
Does Not Matter	Undefined	Defined	Pass
Defined	Cluster Name	Does Not Matter	Pass (except onPrem)
Undefined	Cluster Name	Does Not Matter	Fail
Does Not Matter	Cluster Props	Undefined	Pass (except onPrem)
Does Not Matter	Cluster Props	Defined	Pass
Does Not Matter	Auto Tuner	Undefined	Fail
Does Not Matter	Auto Tuner	Defined	Pass

5. Code Changes:

user_tools/tests/spark_rapids_tools_ut/test_tool_argprocessor.py:

• In each unit test, we use the @register_triplet_test decorator with the parameters <platform, cluster, eventlog>. This decorator is responsible for adding the triplet value to a global dictionary.

6. Correctness Check:

The final unit test, test_arg_cases_coverage, checks if the dictionary contains all 16 triplets, thus ensuring that these unit tests cover all 16 cases.

[nartal1]

LGTM. Thanks @parthosa !

[cindyyuanjiang]

Thanks @parthosa!

[nartal1]

Thanks @parthosa !