# Implement Missing Features (Phase 4 & 5)

I will implement **Advanced Metrics** for Edge Functions and **pgvector Support** for the Data API.

## 1. pgvector Support (`data_api`)
**Goal**: Ensure `vector` columns are returned as native JSON arrays instead of strings in the Data API.

-   **Modification**: Update `rows_to_json` in `data_api/src/handlers.rs`.
-   **Logic**:
    -   Check if column type is "VECTOR".
    -   If yes, parse the string representation (e.g., `"[1.0,2.0,3.0]"`) into a `serde_json::Value::Array`.
    -   This provides seamless integration for clients using embeddings.

## 2. Advanced Metrics (`functions`)
**Goal**: Capture and log detailed execution metrics for Edge Functions.

-   **Modification**: Update `functions/src/runtime.rs` and `functions/src/handlers.rs`.
-   **Metrics**:
    -   `execution_time_ms`: Duration of the WASM execution.
    -   `memory_usage_bytes`: Approximate memory usage (if obtainable) or payload size.
-   **Implementation**:
    -   Use `tracing` with structured fields (e.g., `tracing::info!(target: "function_metrics", duration_ms = 123, ...)`).
    -   This allows the existing **Logs Viewer** (Loki-based) to aggregate and visualize these metrics in the future.

## Execution Steps
1.  **Refactor Data API**: Modify `rows_to_json` to handle `VECTOR` type.
2.  **Instrument Functions**: Add timing and logging to `WasmRuntime::execute`.
3.  **Verify**: Ensure compilation and correct logic.