Add Data Lake Consumer

Event consumer that subscribes to RabbitMQ and writes to Azure Data Lake (Parquet). Pattern: logging’s DataLakeCollector for risk.evaluated (BI_SALES_RISK_IMPLEMENTATION_PLAN §3.5, §9.1). BI Sales Risk: Paths and Parquet columns MUST match documentation/requirements/BI_SALES_RISK_DATA_LAKE_LAYOUT.md (§2.1 risk.evaluated, §2.2 ml_outcomes, §4 config).

1. Consumer

Path: src/events/consumers/[Name]DataLakeCollector.ts

• EventConsumer with queue, exchange: coder_events, bindings: e.g. ['risk.evaluated','ml.prediction.completed','opportunity.updated','forecast.generated'].
• Handler: map event to row. For risk.evaluated use columns in Data Lake Layout §2.1. Build path: {path_prefix}/year={YYYY}/month={MM}/day={DD}/... (Layout §1).
• Write via @azure/storage-blob (BlockBlob) or @azure/storage-blob + parquetjs (or Arrow) for Parquet. Buffer/batch by time or count if needed.
• Config: data_lake.connection_string, data_lake.container, data_lake.path_prefix (e.g. /risk_evaluations).

2. Config