Prune cardano transactions in signer after block range roots computation

[Alenar]

Content

This PR add, to the signer only, a pruning of the cardano transactions after the computation of block ranges roots.

This is done using a decorator wrapping the existing CardanoTransactionsImporter.

Two new configuration parameters have been added to signer Configuration structure:

• network_security_parameter, default to 2160 (mainnet value).
• enable_transaction_pruning, default to true.

Since both have a default value our signer won't have to update their configurations.

Pre-submit checklist

• Branch
  • Tests are provided (if possible)
  • Crates versions are updated (if relevant)
  • CHANGELOG file is updated (if relevant)
  • Commit sequence broadly makes sense
  • Key commits have useful messages
• PR
  • No clippy warnings in the CI
  • Self-reviewed the diff
  • Useful pull request description
  • Reviewer requested
• Documentation
  • Update README file (if relevant)
  • Update documentation website (if relevant)
  • Add dev blog post (if relevant)

Comments

• As it was the third time that we needed to do a simple query to retrieve a single result from database, a system using a extension trait on the sqlite connection has been added to do that simply with reduced boilerplate.
• Some simplification has been done on the configuration system: a leaner way to hydrate default configuration (mithril-signer & mithril-aggregator) and generalized use of the test configuration values (new_sample(party_id), mithril-signer).

Issue(s)

Closes #1645

[github-actions]

Test Results

    3 files  ± 0     43 suites  ±0   8m 32s ⏱️ ±0s
  970 tests +11    970 ✅ +11  0 💤 ±0  0 ❌ ±0 
1 064 runs  +11  1 064 ✅ +11  0 💤 ±0  0 ❌ ±0 

Results for commit e258352. ± Comparison against base commit 2fe74c2.

♻️ This comment has been updated with latest results.

[dlachaume]

LGTM 👍

[dlachaume]

Is the use of the generic data type Q: Into<String> necessary here? The function is always called with &str parameters. Would it be simpler to directly use &str instead?

[Alenar]

Since it's a low level api that may be used in different kind of scenario I would like to have some flexibility with the input parameters, hence the generic.
That said after re-challenging it I don't need a owned string here but only a ref to a str, so I'm thinking of changing the generics to AsRef<str>, this avoid an allocation and simplify a bit the implementation:

 impl ConnectionExtensions for SqliteConnection {
-    fn query_single_cell<Q: Into<String>, T: ReadableWithIndex>(
+    fn query_single_cell<Q: AsRef<str>, T: ReadableWithIndex>(
         &self,
         sql: Q,
         params: &[Value],
     ) -> StdResult<T> {
-        let sql = sql.into();
         let mut statement = self.prepare(&sql).with_context(|| {
             format!(
                 "Prepare query error: SQL=`{}`",
-                &sql.replace('\n', " ").trim()
+                sql.as_ref().replace('\n', " ").trim()
             )
         })?;
         statement.bind(params)?;

[jpraynaud]

LGTM 👍

I left few comments and suggestions below.