{"data":{"kind":"file","path":"README.md","version_id":"zkapp35wpb3yz1w9hujvp3qe","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":1783,"modified_at":"2026-09-20T02:56:52.113000","content_hash":"a45b48748cebb7514ea9eeea7bbe0ddffde946e24986e31048987939b524409e"},"entries":[],"content":"# Smart Agriculture Evaluation Environment\n\nA rigorous verifiers-based benchmark environment evaluating Large Language Models on precision agriculture, agronomy diagnostics, IoT sensor analysis, and automated greenhouse management.\n\n## Overview\n\nModern smart farming relies on data-driven interventions. Models evaluated in this environment encounter real-world agricultural scenarios with complex telemetry:\n- **Irrigation & Soil Moisture Management**: Evapotranspiration models, tensiometer readings, deficit irrigation thresholds.\n- **Precision Plant Nutrition & Fertigation**: Electrical conductivity (EC), pH buffering, nitrogen-phosphorus-potassium balance.\n- **Integrated Pest & Disease Management (IPM)**: Degree-day modeling, economic injury levels (EIL), biological vs chemical control trade-offs.\n- **Controlled Environment Agriculture (CEA)**: Vapor pressure deficit (VPD) regulation, supplemental DLI (daily light integral), CO2 enrichment.\n\n## Evaluation Metrics\n\nThe environment implements a hardened multi-dimensional rubric assessing:\n1. **Decision Accuracy**: Correctness of the agricultural diagnosis, dosing calculation, or control setpoints against established agronomic standards.\n2. **Actionable Prescription**: Verification that numerical recommendations (liters/ha, kg/ha, target VPD, ppm) are precise and physically sound.\n3. **Safety & Risk Mitigation**: Evaluation of warnings regarding environmental runoff, phytotoxicity, or beneficial insect safety.\n4. **Reasoning Completeness**: Rigorous step-by-step agronomic justification incorporating the provided sensor telemetry.\n\n## Usage\n\n```python\nimport verifiers as vf\nfrom smart_agriculture import load_environment\n\nenv = load_environment()\nprint(f\"Loaded {len(env.dataset)} agricultural scenarios\")\n```\n","encoding":"utf-8","truncated":false,"total_bytes":1783},"status":null}