refactor: Extract deployment file model classes to sqrl-deployment-model

[ferenc-csaky]

Summary

Extract deployment-file models into a dependency-light sqrl-deployment-model module. Deployment JSON is now produced from explicit file models rather than planner physical plans, keeping compiler state and persisted contracts separate.

Motivation

Deployment-file consumers should not require Calcite, Flink, or planner implementation dependencies. Persisting physical plans also exposes compiler-only state and can accidentally serialize unsupported runtime types.

Implementation

• Add the sqrl-deployment-model Maven module, limited to JDK types and Jackson annotations.
• Add CombinedPlanModel, FlinkPlanModel, JdbcPlanModel, JdbcStatementModel, KafkaPlanModel, KafkaNewTopicModel, and MutationDatabaseModel.
• Add EnginePhysicalPlan.toFileModel() and update Packager to serialize that representation.
• Add explicit mappings for Flink, JDBC, Kafka, and combined database plans.
• Preserve the combined-plan plans wrapper while mapping every nested plan to its file model. This prevents Jackson from traversing Calcite RelDataType instances during deployment serialization.
• Keep planner lifecycle behavior in sqrl-planner; KafkaNewTopic remains the EngineCreateTable wrapper around the deployment DTO.
• Keep mutation-database creation and compatibility checks in planner helpers while persisting MutationDatabaseModel.
• Update config loaders and CLI execution to consume deployment models when reading postgres.json, kafka.json, and mutation-database output.

Compatibility

• Preserve existing deployment JSON structure, including the combined plans wrapper.
• Preserve JDBC CREATE TABLE metadata: fields, primary key, partition key/type/count, and TTL.
• Preserve Kafka topic convenience constructors and defaults.
• Mark derived KafkaPlanModel.isEmpty() as ignored so kafka.json only contains topics and testRunnerTopics.
• Retain Jackson deserialization support for persisted TestPlan.jdbcViews as GenericJdbcStatement.

Testing

• Expand PlanModelSerializationTest to cover JDBC generic, CREATE TABLE, and extension statements; Flink fields; Kafka topic mapping/defaults; Kafka JSON properties; and combined-plan serialization containing a Calcite BIGINT type.
• Add a CLI regression test for deserializing persisted TestPlan.jdbcViews.

[codecov]

Codecov Report

❌ Patch coverage is 63.07692% with 24 lines in your changes missing coverage. Please review.
✅ Project coverage is 17.44%. Comparing base (ea69f32) to head (5e7483c).
⚠️ Report is 1 commits behind head on main.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
.../com/datasqrl/engine/log/kafka/KafkaLogEngine.java	0.00%	9 Missing ⚠️
...li/src/main/java/com/datasqrl/cli/DatasqrlRun.java	0.00%	3 Missing ⚠️
.../java/com/datasqrl/compile/CompilationProcess.java	0.00%	2 Missing ⚠️
...engine/database/relational/AbstractJDBCEngine.java	0.00%	2 Missing ⚠️
...l/engine/database/relational/JdbcPhysicalPlan.java	93.54%	1 Missing and 1 partial ⚠️
...n/java/com/datasqrl/planner/FlinkPhysicalPlan.java	33.33%	2 Missing ⚠️
...om/datasqrl/planner/dag/plan/MutationDatabase.java	0.00%	2 Missing ⚠️
.../src/main/java/com/datasqrl/compile/DagWriter.java	0.00%	1 Missing ⚠️
...main/java/com/datasqrl/util/ConfigLoaderUtils.java	85.71%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #2238      +/-   ##
============================================
+ Coverage     17.13%   17.44%   +0.30%     
- Complexity     1054     1072      +18     
============================================
  Files           613      612       -1     
  Lines         17802    17807       +5     
  Branches       2166     2167       +1     
============================================
+ Hits           3051     3107      +56     
+ Misses        14435    14382      -53     
- Partials        316      318       +2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.