feat: add full support of Query API (pre-rc.20)

[0x009922]

Queries

Context

Existing Query API implementation in Iroha JS is not complete:

• It doesn't account for "Live Queries". Iroha by default returns query results as batches, and clients must make subsequent requests with the given query_id and cursor retrieve more results.
• It doesn't support pagination and sorting parameters.

Solution

The new API looks like this:

// iterable query
const domains = client
  .query(pre, datamodel.QueryBox('FindDomains'), {
    // optional params
    start: 10,
    limit: 20,
    sortByMetadataKey: 'sorting-key'
  })
  .then((response) => response.as('Iter').all())

// singular query
const transaction = client
  .query(pre, datamodel.QueryBox(
    'FindTransactionByHash', 
    datamodel.FindTransactionByHash({ hash: ... })
  ))
  .then((response) => response.as('Value').enum.as('TransactionQueryOutput'))

The response is a enumeration of possible query responses:

• Iter for iterable query response. This is a live query, and it contains a handle to fetch subsequent batches of data. As in the example above, there is a helper method .all(): Promise<Value[]> that simply fetches all batches and collects them into a single array.
• Value for singular query response. This is returned by queries that return only a single result.
• Failure with decoded ValidationFail reason.

computeTransactionHash

While investigating #200, I found that the transaction hash emitted by pipeline events is different from the hash used for querying transactions: the first is HashOf<TransactionPayload>, and the second is HashOf<SignedTransaction>.

To reduce this ambiguity for SDK users, I soft-deprecated computeTransactionHash and added new helpers: computeTransactionPayloadHash and computeSignedTransactionHash.

[0x009922]

Guess we don't have proper CI for now...

I swear I have run the checks locally!