Port correct streaming SQL TVF patterns into restructured SDP skill

[calreynolds]

Context

PR #320 by @sienkiewiczPat fixed several real issues in the SDP streaming SQL documentation, but the skill was restructured in #368 and the target files no longer exist. Closing #320 but porting the fixes here.

Supersedes #304 (which reported the same broken TVFs).

What needs to be added

The following corrections from #320 should be applied to the current databricks-skills/databricks-spark-declarative-pipelines/references/sql/ directory:

1. Kafka ingestion: Use read_kafka() TVF (not read_stream()), with correct parameter names (bootstrapServers, backtick-quoted dot-options like `kafka.security.protocol`), and serviceCredential auth option
2. Kinesis ingestion: Use read_kinesis() TVF with correct parameter names (streamName, initialPosition) and all three auth patterns (explicit credentials, IAM role, environment variables)
3. Event Hub ingestion: Use read_kafka() with SASL/SSL JAAS config (no read_eventhub TVF exists; Event Hub uses Kafka-compatible endpoint)
4. Secret syntax: Use secret('scope', 'key') function — the {{secrets/scope/key}} template syntax is not valid in SDP SQL
5. Advanced config: Add pipelines.reset.allowed config key

Reference

See PR #320 diff for the exact corrections: https://github.com/databricks-solutions/ai-dev-kit/pull/320/files