Build the application

## Build
Correlate supervisory flowmeter telemetry with minimum night flow benchmarks to identify hidden pipeline leaks. Generate prioritized acoustic sounding inspection orders for municipal water leak detection teams.

## User
Water Distribution Loss Control Specialist

## Workflow
1. Ingest hourly inflow telemetry and customer automated meter reading feeds for designated District Metered Areas (DMAs).
2. Compute Minimum Night Flow (MNF) between 02:00 and 04:00 to isolate legitimate baseline consumption.
3. Model expected hydraulic background losses using WNTR pressure-dependent demand calculations.
4. Synthesize field inspection work orders targeting pipe clusters where unexplained night flows exceed threshold tolerances.

## Interface
Time-series chart showing DMA inlet flows, night flow thresholds, and an inspection dispatch queue with map overlays.

## Data
Synthesize 15-minute resolution DMA flowmeter logs, meter billing sums, and EPANET/WNTR pipe models.
Starter examples below are synthetic. They are not downloaded records.

## Source
[Water Network Tool for Resilience](https://raw.githubusercontent.com/USEPA/WNTR/main/README.md)
Applies WNTR hydraulic analysis routines to calculate theoretical background leakage based on average operating pressure across the DMA.

## Tools
langfuse, wntr, scipy.
Langfuse provides observability and operational tracing for multi-step hydraulic anomaly detection and dispatch pipelines.
Local Python execution runtime with hosted LLM for structuring water loss inspection briefs.

## Deliverables
Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.

## Acceptance
- Calculates Net Night Flow by subtracting customer AMR night usage from bulk meter night inflow.
- Triggers leak inspection alerts when unexplained night loss exceeds 20% of minimum inflow.
- Outputs prioritized street inspection vectors with localized pressure monitoring points.

## Starter Examples
```json
[
  {
    "input": {
      "dma_id": "DMA-DEIRA-02",
      "bulk_night_flow_m3h": 68.5,
      "customer_amr_night_m3h": 18.2,
      "average_pressure_bar": 3.8,
      "total_pipe_length_km": 14.5
    },
    "expected_output": {
      "net_night_loss_m3h": 50.3,
      "allowable_background_loss_m3h": 12.4,
      "unexplained_loss_m3h": 37.9,
      "leak_alert": true,
      "recommended_action": "Deploy acoustic correlators along sector mains A and B"
    }
  },
  {
    "input": {
      "dma_id": "DMA-DEIRA-05",
      "bulk_night_flow_m3h": 22,
      "customer_amr_night_m3h": 14.5,
      "average_pressure_bar": 3.2,
      "total_pipe_length_km": 8
    },
    "expected_output": {
      "net_night_loss_m3h": 7.5,
      "allowable_background_loss_m3h": 6.8,
      "unexplained_loss_m3h": 0.7,
      "leak_alert": false,
      "recommended_action": "Routine monitoring, no immediate dispatch required"
    }
  }
]
```

## Safety
Leak detection dispatches are diagnostic surveys; no valve operations or system pressure adjustments should be executed without utility authorization.
Use mock external writes. Require human review before any real deployment.