Blockfrost Implementation

[errfrom]

Closes #1119
Closes #734
Closes #1234

Includes #1364: Some tests for getTxMetadata, getScriptByHash, and getScriptsByHashes
Includes #1371: Framework for implementing Blockfrost queries
Includes #1385: Foundation for generating Blockfrost fixtures
Includes #1368: getTxMetadata and isTxConfirmed queries
Includes #1378: getDatumByHash and getScriptByHash queries
Includes #1366: utxosAt and getUtxoByOref queries
Includes #1372: getProtocolParameters and getCurrentEpoch queries
Includes #1376: submitTx and evaluateTx queries
Includes #1377: getChainTip, getEraSummaries, and getSystemStart queries
Includes #1389: Revised Contract testing utilities
Includes #1401: Various fixes, bring back submitE, trace log blockfrost requests
includes #1412: Introduce confirmTxDelay to smooth over issues with Blockfrost API

Notable changes:

• ConfigParams r -> ContractParams (outside of QueryM)
• Contract r -> Contract
• Removed unused bindings for ODC and Ogmios. Namely utxo lookup in Ogmos, and Datum fetching in ODC. We no longer use them and are not exposed in the API.
  • ODC has been removed completely
  • Postgres too
    • This will only be for the CTL backend, it turns out blockfrost requires it
• Fleshed out the docs a little to use the new types/contract params, and mention there's multiple backends.
• getTxByHash is currently CTLBackend only. Will be removed when Kupo replaces ODC fully.
  • Actually this PR uses kupo for isTxConfirmed, so we can do so now. But I'm not sure I trust the implementation at the moment (tracked in Can't use Kupo to confirm Txs with no outputs #1293)
• getTxByHash has been removed, and getTxMetadata has been added. This, as well as utxo, datum and script lookup, should cover all cases needed for getting a full transaction.
• KeyWallet now receives a networkId on construction (for Contract based construction, this is asked)
  • It makes no sense to re-use a wallet for multiple networks, and this simplifies implementation
  • mkKeyWalletFromFiles/PrivateKeys is now in Contract and returns a KeyWallet instead of Wallet
  • Reverted the above, except for mkKeyWalletFromFiles still returns KeyWallet
• reindexSpentScriptRedeemers is no longer in Contract (it's pure)
• Lots of logic that was previously QueryM is now in Contract, polymorphic to the backend.
• Removed submitE - submission is only in maybe internally. Readded
• All 'cardano' actions are now implemented via the abstract handlers, other than those relating to staking
• Test.Ctl.AffInterface -> Test.Ctl.QueryM.AffInterface
• Moved ProtocolParameters, EraSummaries, Epoch, and SystemStartto Ctl.Internal.Types
• Moved ClientError to Ctl.Internal.Service.Error

Remaining issues:

• ContractEnv/QueryHandle mention Ogmios types.
  • Not a huge issue, but conceptually a bit odd.
  • ProtocolParameters (and EraSummaries) should ideally be placed in Cardano.Types. One issue is the Epoch type, which already exists in Cardano.Types.Transaction. The underlying representations are different however.
  • One problem with ProtocolParameters and EraSummaries, is the responses are different between Ogmios and Blockfrost, but I think only by camel case vs snake case. I don't think this is the case for tx evaluation though, which is explicitly implemented via ogmios by all backends.
• Similar to other types, ClientError is defined by QueryM yet is used in Contract
• In general I think backend-agnostic error responses/handling may need more thought
• No metadata submission + fetching example
• There's an issue with log supression in the Plutip tests

Pre-review checklist

• All code has been formatted using our config (make format)
• Any new API features or modification of existing behavior is covered as defined in our test plan
• The changelog has been updated under the ## Unreleased header, using the appropriate sub-headings (### Added, ### Removed, ### Fixed), and the links to the appropriate issues/PRs have been included

done