feat: add iceberg storage consumer support

[tenfyzhong]

Summary

• add a parallel Iceberg consumer path with metadata-version replay, row conversion, checkpoints, and shared row staging while keeping the existing CSV path intact
• wire --source-format=iceberg and --iceberg.source-uri into the warehouse commands for Snowflake, Redshift, BigQuery, and Databricks
• add warehouse-specific row appliers, Iceberg metrics, docs, and replace the local TiCDC worktree override with a reproducible fork reference

Changes

• introduce pkg/icebergconsumer for config, checkpoints, schema diffing, and row conversion
• add ReplicateIceberg(...) runtime wiring in cmd/core.go and source flag validation for Iceberg mode
• add row appliers for Snowflake, Redshift, BigQuery, and Databricks
• migrate compatible imports from legacy TiFlow paths to TiCDC paths
• update docs for Iceberg source usage and staging semantics

Testing

• GOTOOLCHAIN=go1.25.8 go test ./... -count=1 in an environment that rewrites https://github.com/ to SSH and sets GOPROXY=direct, GOPRIVATE=github.com/tenfyzhong/ticdc, GONOSUMDB=github.com/tenfyzhong/ticdc, and GONOPROXY=github.com/tenfyzhong/ticdc so the forked TiCDC replacement can be fetched reproducibly

Dependency

• Depends on sink,storage: add iceberg protocol for cloud storage sink pingcap/ticdc#4565