{"data":{"kind":"file","path":"README.md","version_id":"o48nqzwbx6hinx2f7v15gbv8","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":1572,"modified_at":"2026-08-12T01:41:41.551000","content_hash":"b6b9f9bf60658b3c11631653c9eebead992563f6e80c6dd358a52882e2527aab"},"entries":[],"content":"# Astronomical Observation Planner\n\nAn environment that evaluates an LLM's ability to plan astronomical observations by reasoning about celestial mechanics, telescope scheduling, atmospheric conditions, and target prioritization.\n\n## Overview\n\nThe model is presented with astronomical observation scenarios and must produce detailed, actionable observation plans. Tasks include:\n\n- **Target Selection**: Choosing optimal celestial targets based on season, time of night, and equipment\n- **Scheduling**: Planning observation windows considering moon phase, weather, and object visibility\n- **Condition Assessment**: Evaluating seeing conditions, light pollution, and atmospheric transparency\n- **Equipment Matching**: Recommending appropriate telescope/filter combinations for specific targets\n- **Priority Ranking**: Ordering observation targets by scientific value, visibility, and feasibility\n\n## Scoring\n\nThe environment uses multiple rubric functions:\n\n| Metric | Weight | Description |\n|--------|--------|-------------|\n| `accuracy_score` | 1.0 | Correctness of astronomical facts and calculations |\n| `completeness_score` | 0.8 | Coverage of all required planning elements |\n| `feasibility_score` | 0.6 | Whether the plan is physically executable |\n| `reasoning_score` | 0.5 | Quality of justification for choices made |\n\n## Usage\n\n```python\nimport verifiers as vf\n\nenv = vf.load_environment(\"astro-obs-planner\")\n# or\nfrom astro_obs_planner import load_environment\nenv = load_environment()\n```\n\n## Tags\n\n`astronomy` `planning` `observation` `scheduling` `reasoning`\n","encoding":"utf-8","truncated":false,"total_bytes":1572},"status":null}