name: DevOpsEnv
version: "1.0.0"
description: >
  An OpenEnv-compliant Linux DevOps and SRE troubleshooting environment.
  Agents act as a junior SRE to diagnose and fix broken simulated server states
  using bash commands and file edits.

domain: software_engineering
tags:
  - openenv
  - devops
  - sre
  - troubleshooting
  - linux

license: MIT
author: SolvorLabs

# -------------------------------------------------
# Environment interface
# -------------------------------------------------
interface:
  reset:
    method: POST
    path: /reset
    request:
      task_id: string # task1 | task2 | task3
    response: Observation

  step:
    method: POST
    path: /step
    request:
      episode_id: string
      action: Action
    response: StepResult

  state:
    method: GET
    path: /state
    params:
      episode_id: string
    response: State

  tasks:
    method: GET
    path: /tasks
    response: List[TaskInfo]

  grader:
    method: POST
    path: /grader
    request:
      episode_id: string
    response: GraderResponse

  health:
    method: GET
    path: /health

# -------------------------------------------------
# Typed models (Matches your models.py)
# -------------------------------------------------
models:
  Observation:
    task_id: string
    task_description: string
    episode_id: string
    system_state: SystemState
    thread_history: list[dict]
    available_actions: list[string]
    step_number: integer
    max_steps: integer
    hint: string | null

  SystemState:
    task_id: string
    available_commands: list[string]
    filesystem_snapshot: string
    running_processes: list[dict]
    service_status: dict[string, string]
    logs: string
    http_ports_open: list[integer]
    docker_containers: list[dict]
    cpu_usage: float
    memory_usage_mb: integer

  Action:
    action_type: string # bash_cmd | file_edit | submit
    command: string | null
    file_path: string | null
    file_content: string | null
    summary: string | null

  Reward:
    step_reward: float
    total_reward: float
    explanation: string

  StepResult:
    observation: Observation
    reward: Reward
    done: boolean
    info: dict

  State:
    task_id: string
    episode_id: string
    step_number: integer
    max_steps: integer
    done: boolean
    total_reward: float
    history: list[dict]
    final_score: float | null

  GraderResponse:
    episode_id: string
    task_id: string
    score: float # 0.0 – 1.0
    breakdown: dict[string, float]
    feedback: string

# -------------------------------------------------
# Tasks (Matches your data.py)
# -------------------------------------------------
tasks:
  task1:
    name: "Nginx Service Recovery"
    difficulty: easy
    max_steps: 10
    description: >
      The Nginx web server has crashed. The agent must restart the service
      and ensure port 80 is returning an HTTP 200 OK status.
      
  task2:
    name: "Docker Compose Fix"
    difficulty: medium
    max_steps: 15
    description: >
      A Docker container is misconfigured with the wrong port mapping. 
      The agent must edit docker-compose.yml to fix the port and spin the container up.

  task3:
    name: "API Memory Leak"
    difficulty: hard
    max_steps: 20
    description: >
      A Python mock API is experiencing a memory leak. The agent must diagnose 
      the running processes, kill the faulty process, patch the python file, and restart.

# -------------------------------------------------
# Deployment
# -------------------------------------------------
deployment:
  framework: FastAPI
  python: ">=3.11"
  port: 7860
  dockerfile: Dockerfile
  huggingface_space: true
  space_sdk: docker