Change Confirmations to Confidence

[EasterTheBunny]

Confidence is a different approach to mapping a chain agnostic concept of confidence that
an event or transaction can or should be acted on. Confidence is intended to map to defined
levels of confirmations on EVM chains and other states on other chains.

Only two levels of Confidence exist by default to define the boundaries: Highest and
Lowest. This creates an inclusive range for all chain configurable Confidence levels and
can be treated as % confidence that an event or transaction can or should be acted on.

[github-actions]

I see you updated files related to core. Please run pnpm changeset in the root directory to add a changeset as well as in the text include at least one of the following tags:

• #added For any new functionality added.
• #breaking_change For any functionality that requires manual action for the node to boot.
• #bugfix For bug fixes.
• #changed For any change to the existing functionality.
• #db_update For any feature that introduces updates to database schema.
• #deprecation_notice For any upcoming deprecation functionality.
• #internal For changesets that need to be excluded from the final changelog.
• #nops For any feature that is NOP facing and needs to be in the official Release Notes for the release.
• #removed For any functionality/config that is removed.
• #updated For any functionality that is updated.
• #wip For any change that is not ready yet and external communication about it should be held off till it is feature complete.

[mateusz-sekara]

Hmm, I think I have mentioned that in a previous PR, but I believe reusing this instance doesn't give us much value, but can lead to major isues. I'm not an expert when it comes to Go GC, but most of the collectors are blazingly fast and optimized for short-living objects, memory footprint shouldn't be the case here. It would be safer to always create a new pgDSLParser on every query parsing.

For instance, you can have multiple threads (e.g. plugin) running LogPoller read queries, when reusing the same pgDSLParser instance it's not gonna work. I believe you should be able to write a concurrent test that spawns multiple routines calling FilteredLogs

[EasterTheBunny]

Good callout. I'll test concurrency and maybe the best path forward is to benchmark both options.

[EasterTheBunny]

I ran some quick benchmarks on both options with a query containing nested expressions. The benchmark was run with the command go test -bench=BenchmarkParser -benchtime=10s -benchmem each time.

wordFilter := NewEventByWordFilter(common.HexToHash("0x42"), 8, []primitives.ValueComparator{
	{Value: "a", Operator: primitives.Gt},
	{Value: "b", Operator: primitives.Lte},
})

expressions := []query.Expression{
	{BoolExpression: query.BoolExpression{
		Expressions: []query.Expression{
			query.Timestamp(10, primitives.Eq),
			{BoolExpression: query.BoolExpression{
				Expressions: []query.Expression{
					query.TxHash(common.HexToHash("0x84").Hex()),
					{BoolExpression: query.BoolExpression{
						Expressions: []query.Expression{
							query.Confirmation(primitives.Unconfirmed),
							wordFilter,
						},
						BoolOperator: query.AND,
					}},
				},
				BoolOperator: query.OR,
			}},
		},
		BoolOperator: query.AND,
	}},
}

No changes to pgDSLParser. The expression and err are reset for each build of an expression.
goos: darwin
goarch: arm64
pkg: github.com/smartcontractkit/chainlink/v2/core/chains/evm/logpoller
BenchmarkParser-10 3622035 3270 ns/op 5600 B/op 81 allocs/op
PASS
ok github.com/smartcontractkit/chainlink/v2/core/chains/evm/logpoller 20.203s

Split pgDSLVistor from pgDSLParser and created a new visitor instance for each expression in the build loop.
goos: darwin
goarch: arm64
pkg: github.com/smartcontractkit/chainlink/v2/core/chains/evm/logpoller
BenchmarkParser-10 3278098 4672 ns/op 5744 B/op 84 allocs/op
PASS
ok github.com/smartcontractkit/chainlink/v2/core/chains/evm/logpoller 38.635s

I don't see a huge difference between the two, but there does appear to still be some time loss on allocations.

[mateusz-sekara]

For debugging purposes, it could be worth logging what expression was built. However, maybe in the ORM, not the parser itself

[mateusz-sekara]

Some godoc or tests showing the expected cursor format would be very helpful here

[mateusz-sekara]

Does it make sense to sort that slice on every call? We should sort it on-write, right? When e.confidenceSteps is initialized you can just sort it once.

[mateusz-sekara]

Shouldn't that be an error? If step doesn't exist then someone is trying to do something that is not supported and might be not aware of that because code silently fallbacks to 0

[cl-sonarqube-production]

Quality Gate passed

Issues
0 New issues
0 Fixed issues
0 Accepted issues

Measures
0 Security Hotspots
73.5% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube