Reintroduce db_write_timestamp wield to avoid breaking changes

[JonoPrest]

The integration tests were failing because I removed db_write_timestamp field from user entities.

I did this because during refactor I though why are we enforcing this field for users. (We did it early on to try and capture some metrics of the indexer)

So I removed it but I realise this is a breaking change if anyone is actually using that field with Hasura/postgres queries.

We can remove in v2 but not in v1

[JonoPrest]

Also some tests were broken. I'm not sure how, they were passing in the PR before merge as far as I'm aware.

[DZakh]

Maybe they didn't run because a merge conflict or something

[DZakh]

And they are failing even now 🤔

[JonoPrest]

Nice! Thank you 👍🏼

[DZakh]

I don't like that integration tests are also failing. According to the logs greetings is a string instead of an array of strings

Maybe the db column type has changed. I'm not familiar with the part of the code.

[JonoPrest]

I don't like that integration tests are also failing. According to the logs greetings is a string instead of an array of strings

Maybe the db column type has changed. I'm not familiar with the part of the code.

I'm looking at it now. Definitely something funky

[JonoPrest]

@DZakh array flags were missing! 🙏🏼 thank you integration tests 👼🏼 and thanks for spotting the issue. Let's see if we pass now.