Make negation normal form apply the propagated not statements in the leaf nodes

[uakyol]

This PR fixes this bug : #1132

[wfa-reviewable]

This change is 

[SanjayVas]

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @riemanli)

[riemanli]

Reviewed 1 of 2 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @uakyol)

---

src/test/kotlin/org/wfanet/measurement/eventdataprovider/eventfiltration/validation/EventFilterValidatorTest.kt line 408 at r1 (raw file):

    val expectedNormalizedExpression =
      "(!(person.gender == 2)  && true) && (!(person.age_group == 1) && " +

I know for non-operative node we always convert it to True, ex: video.viewed_fraction > 0.25. But for the case that the input expression is !(person.gender == 2 && video.viewed_fraction > 0.25), the code outputs (!(person.gender == 2) || true) which is always evaluated as True. Is that ok for the PBM?

[uakyol]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @riemanli)

---

src/test/kotlin/org/wfanet/measurement/eventdataprovider/eventfiltration/validation/EventFilterValidatorTest.kt line 408 at r1 (raw file):

Previously, riemanli (Rieman) wrote…

I know for non-operative node we always convert it to True, ex: video.viewed_fraction > 0.25. But for the case that the input expression is !(person.gender == 2 && video.viewed_fraction > 0.25), the code outputs (!(person.gender == 2) || true) which is always evaluated as True. Is that ok for the PBM?

Yes, the design is done this way. Basically, worst case we will charge more buckets but never less than intended.

[riemanli]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @uakyol)

---

src/test/kotlin/org/wfanet/measurement/eventdataprovider/eventfiltration/validation/EventFilterValidatorTest.kt line 408 at r1 (raw file):

Previously, uakyol wrote…

Yes, the design is done this way. Basically, worst case we will charge more buckets but never less than intended.

Got it. Thanks for clarifying!

[stevenwarejones]

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @uakyol)