Can we have a variant of the endpoints which take the mempool into account?

[lemmon-714]

First of all, I hope this is the right place for this. If not, please redirect me.

I am currently working on transaction-chaining, which works now (with blockfrost, thank you!). However, in some edge case it fails due to the mempool not being included in the utxo-endpoint.

I am bringing this up here as well for two reasons:

1. Even though the issue above is a very edgy case, it prevents the "clean" solution, which might cause issues once the krakens are unleashed regarding tx-chaining and everyone starts using it everywhere in complex ways. Using the mempool-utxos will prevent the "let's keep the fingers crossed the user has enough utxos for collateral hanging around even when doing weird chain-straining things" problem source.
2. Fixing it at the blockfrost-level will make it easier to fix it in other wallets (which might not have that issue yet, have not checked). It also appears to be a bit "cleaner" in general.

I need to mention - especially regaring the "cleaner" bit - that simply including the mempool in the queried ledger-state is obviously incorrect for multiple reasons. But right now I am keeping track of my own pending modifications to the ledger state and modify the api-gotten utxos with that. Which is of course suboptimal. But having a simple variant to the endpoint(s) which include the mempool might make it easy to adjust the wallets downstream.