Update @value rule for qualitative measures

[samuele-mattiuzzo]

After our recent work on the Rules, we have omitted/misunderstood a piece of rule.

From Andy's comment we can clearly see how we are not explicitly checking for

[@measure=5] and not(@value) (pseudocode obviously) but instead we infer its correctness as a side effect.

After a few chats with @amy-silcock we realised how feeble this logic is, so we have to update the rules for that specific check, so that we avoid inferring correctnes by the fact that the other similar tests are Fail or Pass

[andylolz]

[@measure=5] and not(@value)

^^ Yes, exactly like this! 👍 Err maybe something like:

{
    "//result/indicator[@measure='5']/baseline": {
        "evaluates_to_true": {
            "cases": [
                { "eval": "not(@value)" }
            ]
        }
    }
}

but instead we infer its correctness as a side effect

I don’t think so… Currently, if @measure is 5 and @value is present, no ruleset rule fails / no error is raised. So the explicit test is good partly because it’s good to have an explicit test, but also I think it adds something that isn’t covered by current tests.

[akmiller01]

I think this is covered by 8.8.3 now: https://github.com/IATI/IATI-Rulesets/blob/version-2.03/rulesets/standard.json#L1638

But meta_tests are still insufficient to test it properly: https://github.com/IATI/IATI-Rulesets/blob/version-2.03/meta_tests/result_indicator_baseline_selector_with_conditions.json