refactor: Make QueryBlocklistRule an interface for extensibility

[abhishekrb19]

Summary

• Converts QueryBlocklistRule from a concrete class to an interface with two methods: getRuleName() and matches(Query<?>).
• Moves the existing implementation to DefaultQueryBlocklistRule with a default impl on the interface @JsonTypeInfo(defaultImpl = DefaultQueryBlocklistRule.class) so existing JSON configs without a "type" field continue to deserialize correctly.
• Extensions can register new rule types as Jackson subtypes via SimpleModule.registerSubtypes(...) and we can even introduce new union rule types with different logic w/o polluting existing default implementation.

Motivation

The current blocklist logic is fixed to datasource/queryType/context matching. Making it an interface allows adding new rule types (e.g., regex-based, time-window-based, user-based) without modifying the existing code.

This PR has:

• been self-reviewed.
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.

[FrankChen021]

Severity	Findings
P0	0
P1	0
P2	1
P3	0
Total	1

Reviewed 7 of 7 changed files.

---

This is an automated review by Codex GPT-5.5

[FrankChen021]

[P2] Fail unknown blocklist rule types instead of defaulting

Using defaultImpl = DefaultQueryBlocklistRule.class makes Jackson use the default rule not only when type is absent, but also when a non-empty type id cannot be resolved. For the new extension point, a broker or coordinator that lacks a registered custom rule subtype can silently deserialize that rule as DefaultQueryBlocklistRule; with Druid's unknown-property-tolerant mapper, extension-specific fields are ignored and any standard fields such as dataSources can turn into a broader block than intended. Please distinguish the legacy missing-type case from an unrecognized explicit type, so unknown extension rules fail fast rather than being interpreted as default rules.

[abhishekrb19]

Addressed in this commit by adding a StrictTypeIdResolver similar to DimensionSchema

[jtuglu1]

@mshahid6 want to take a look here?

[FrankChen021]

I rechecked the follow-up on QueryBlocklistRule type-id handling. The latest change uses StrictTypeIdResolver, preserves legacy missing-type defaulting, and adds coverage for missing, explicit default, and unknown explicit types. This addresses the previous concern; no inline reply is needed.

Reviewed 7 of 7 changed files.

---

This is an automated review by Codex GPT-5.5

[mshahid6]

LGTM, do we need to update the docs too?

[abhishekrb19]

Thanks @mshahid6! Given that this is already backward compatible without type, I was thinking we can leave the docs as-is for now and include some notes about the default type when we introduce more block list rule types. What do you think?

[abhishekrb19]

I went ahead and added some doc updates for completeness - ptal

[FrankChen021]

Severity	Findings
P0	0
P1	0
P2	0
P3	1
Total	1

Severity	Findings
P0	0
P1	0
P2	0
P3	1
Total	1

Reviewed 8 of 8 changed files.

---

This is an automated review by Codex GPT-5.5

[FrankChen021]

[P3] Scope the matching semantics to the default rule type

After introducing custom QueryBlocklistRule types, the following Rule matching behavior section still reads as if every blocklist rule must use datasource/query-type/context criteria with AND semantics and at least one of those criteria. The interface now lets extensions define arbitrary matches logic and fields, so this is only guaranteed for the built-in default rule. Please scope those bullets to the default type or add wording that custom types define their own matching behavior.

[abhishekrb19]

Thanks - updated the docs to clarify this in the latest commit.

[FrankChen021]

I rechecked the follow-up on the docs scoping comment. The latest docs move the default-rule matching bullets under the default type and explicitly state that custom types define their own matching semantics, so no inline reply is needed.

Reviewed 8 of 8 changed files.

---

This is an automated review by Codex GPT-5.5