Seeds API follow ups

[puffo]

A couple of possible follow ups from the Seeds API PR

---

1. Verify mempool processing handles invalid Foreign Ids

Allow invalid seeds through, let them fail when processed and have no effect
We can then follow up with adjusting the filter to be more selective and robust afterwards - especially seeing as using this internally for now and will likely be pretty careful with the json payloads.

---

2. Deprecate ArtistProfile

For this PR, I don't think we need to support artist profiles, just artists, where there shouldn't be any foreign keys (and so then we'll probably move stuff like avatarURL and websiteURL up to artist instead of artistProfile maybe after this PR).

---

3. Ensure that slug adjustments are preserved for backward-compatibility

Best solution would be to support legacy identifiers/slugs and store them so that legacy urls are preserved

---

4. Handle side effects of changing a field

but, for example if we allow the audioURL to be changed after this has occurred, it will not necessarily re-upload the new track nor generate the new ipfsHash.

Ideas:
Investigate in-memory postgres for tests and previews of seed adjustments