fix: Ignore false-y identity values and remove email sanitization

[rmi22186]

Summary

A customer was passing the following into an identity call

const ids = { 
   other: '', // blank hashed email),
   email: 'foo'
}

and this ended up in the Rokt Kit without removing the falsy other identity.

We perform several pieces of logic in our selectPlacements call, and so:

replaceOtherIdentityWithEmailsha256(ids) would result in:

{
emailsha256: '',
email: 'foo'
}

This is eventually passed to sanitizeEmailIdentities (line 197), which removes email if emailsha256 is present. We do this in order to avoid having a customer pass both an email and emailsha256 into selectPlacements. This then leads to

{
   emailsha256: ''
}

Any empty string gets sanitized by Rokt's SDK.

Solution: Ensuring other has a value, as opposed to just existing on the userIdentities object via hasOwnProperty resolves this.

---

Before: No email in the experiences call

After: email appears in the experiences call.

Testing Plan

Unit tests as well as overrode in customer site to ensure this worked properly.

[samdozor]

How was it decided that we would prefer hashed email over raw? The platform always prefers to receive a raw value - it increases ability to CTL and match rates

[rmi22186]

]> How was it decided that we would prefer hashed email over raw? The platform always prefers to receive a raw value - it increases ability to CTL and match rates

@samdozor - This decision was guided by the fact that Rokt's legacy GTM wrapper does this - https://github.com/ROKT/gtm_wrapper/blob/master/src/gtm_wrapper.ts#L69-L85 - With the specific warning being:

console.warn(
				'ROKT: Both raw and hashed email values were provided with the SHA-256 hashing option selected. Only the hashed value will be used for integration.',
			);

My inference is that if the partner goes through the additional step of passing the hashed email, then that is their preference. LMK your thoughts