from praisonaiagents import Agent

# Enable context management with safe defaults
agent = Agent(
    instructions="You are a helpful assistant.",
    context=True,  # Enable with defaults
)

# The agent now automatically:
# - Tracks token usage
# - Optimizes when approaching limits (80% threshold)
# - Uses smart optimization strategy

response = agent.chat("Hello!")

from praisonaiagents import Agent
from praisonaiagents.context import ManagerConfig

# Custom context configuration
config = ManagerConfig(
    auto_compact=True,
    compact_threshold=0.7,  # Trigger at 70%
    strategy="smart",
    monitor_enabled=True,
    monitor_path="./context.txt",
)

agent = Agent(
    instructions="You are helpful.",
    context=config,
)

from praisonaiagents.context import (
    ContextBudgeter,
    ContextLedgerManager,
    get_optimizer,
    OptimizerStrategy,
)

# Create budgeter for your model
budgeter = ContextBudgeter(model="gpt-4o-mini")
budget = budgeter.allocate()
print(f"Usable context: {budget.usable:,} tokens")

# Track token usage
ledger = ContextLedgerManager()
ledger.track_system_prompt("You are a helpful assistant.")
ledger.track_history(messages)
print(f"Total used: {ledger.get_total()} tokens")

# Optimize when needed
optimizer = get_optimizer(OptimizerStrategy.SMART)
optimized, stats = optimizer.optimize(messages, target_tokens=50000)

# Enable context monitoring
praisonai chat --context-monitor

# Use specific optimization strategy
praisonai chat --context-strategy smart --context-threshold 0.8

# View context stats in session
/context stats
/context budget
/context dump

PRAISONAI_CONTEXT_AUTO_COMPACT=true
PRAISONAI_CONTEXT_STRATEGY=smart
PRAISONAI_CONTEXT_THRESHOLD=0.8
PRAISONAI_CONTEXT_MONITOR=true
PRAISONAI_CONTEXT_MONITOR_PATH=./context.txt
PRAISONAI_CONTEXT_MONITOR_FORMAT=human
PRAISONAI_CONTEXT_REDACT=true

from praisonaiagents.context import MultiAgentLedger, MultiAgentMonitor

# Per-agent context isolation
multi_ledger = MultiAgentLedger()
researcher = multi_ledger.get_agent_ledger("researcher")
writer = multi_ledger.get_agent_ledger("writer")

# Per-agent monitoring
multi_monitor = MultiAgentMonitor(base_path="./context/")