Build the application

## Build
Turn unusual intake readings into proposed sampling visits and laboratory evidence requests. Track unresolved samples and prepare review packets without diagnosing algal blooms or recommending chemical doses.

## User
Desalination Environmental Compliance Engineer

## Workflow
1. Import synthetic intake readings and fictional sampling thresholds.
2. Identify exceedances and prepare sampling requests with cited readings.
3. Track sample collection, laboratory evidence, and unresolved follow-up tasks.
4. Export an intake review packet without operational commands.

## Interface
Reading timeline, sample request queue, laboratory attachment panel, and engineer review export.

## Data
Create normalized synthetic readings, fictional sampling thresholds, sample identifiers, and laboratory records. These are not measured GCC data.
Starter examples below are synthetic. They are not downloaded records.

## Source
[WaterTAP treatment modeling](https://raw.githubusercontent.com/watertap-org/watertap/main/README.md)
WaterTAP documents treatment modeling. This application prepares evidence for later modeling, without deriving organic loads or chemical doses.

## Tools
haystack, pydantic, fastapi.
Haystack retrieves linked sample evidence. Deterministic rules create requests from supplied thresholds.
Use a currently available model for evidence-linked summaries, not water-quality diagnosis.

## 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
- Uses thresholds supplied with each fixture.
- Creates a request when either supplied threshold is exceeded.
- Shows missing laboratory evidence instead of diagnosing blooms.
- Exports evidence packets and pending tasks.

## Starter Examples
```json
[
  {
    "input": {
      "intake_id": "A",
      "chlorophyll_ug_l": 15.5,
      "turbidity_ntu": 8,
      "fictional_thresholds": {
        "chlorophyll_ug_l": 10,
        "turbidity_ntu": 5
      },
      "lab_evidence": []
    },
    "expected_output": {
      "sampling_request_proposed": true,
      "exceeded_fields": [
        "chlorophyll_ug_l",
        "turbidity_ntu"
      ],
      "bloom_status": "unverified",
      "chemical_dose_recommendation": null,
      "approval_required": true
    }
  },
  {
    "input": {
      "intake_id": "B",
      "chlorophyll_ug_l": 1.2,
      "turbidity_ntu": 1.5,
      "fictional_thresholds": {
        "chlorophyll_ug_l": 10,
        "turbidity_ntu": 5
      },
      "lab_evidence": []
    },
    "expected_output": {
      "sampling_request_proposed": false,
      "exceeded_fields": [],
      "bloom_status": "unverified",
      "chemical_dose_recommendation": null,
      "approval_required": true
    }
  }
]
```

## Safety
Use synthetic records. Do not diagnose water safety, recommend chemical doses, or operate equipment.
Use mock external writes. Require human review before any real deployment.