from pydantic import BaseModel, Field
from typing import List, Optional


class Resource(BaseModel):
    id: str = Field(description="Unique resource identifier")
    type: str = Field(description="Instance type (e.g., t3.small, m5.large)")
    cpu_usage: float = Field(description="CPU usage percentage")
    mem_usage: float = Field(description="Memory usage percentage")
    monthly_cost: float = Field(description="Monthly cost in USD")


class Metrics(BaseModel):
    avg_latency_ms: float = Field(description="Average latency in milliseconds")
    error_rate: float = Field(description="Error rate (0-1)")
    throughput_rps: float = Field(description="Requests per second")


class SLA(BaseModel):
    max_latency_ms: float = Field(description="Maximum allowed latency")
    max_budget: float = Field(description="Maximum monthly budget in USD")
    min_uptime_pct: float = Field(description="Minimum uptime percentage")


class Observation(BaseModel):
    inventory: List[Resource] = Field(description="List of active cloud resources")
    metrics: Metrics = Field(description="Current system metrics")
    sla: SLA = Field(description="Service Level Agreement requirements")
    echoed_message: str = Field(default="System ready", description="Feedback from last action")
    task_id: str = Field(default="easy", description="Current task identifier")
    task_name: str = Field(default="Right-Sizing", description="Human-readable task name")
    difficulty: str = Field(default="easy", description="Task difficulty level")
    step: int = Field(default=0, description="Current step number")


class Action(BaseModel):
    message: str = Field(description="Agent's command to modify infrastructure")


class Reward(BaseModel):
    value: float = Field(description="Reward value between 0 and 1")
    reason: str = Field(default="", description="Explanation of the reward")
    cost_change_pct: float = Field(default=0.0, description="Percentage change in cost")
    latency_change_pct: float = Field(default=0.0, description="Percentage change in latency")


ObservationModel = Observation
ActionModel = Action
RewardModel = Reward