# my_app/tasks.py
from surreal_commands import command, submit_command
from pydantic import BaseModel

class ProcessInput(BaseModel):
    message: str
    uppercase: bool = False

class ProcessOutput(BaseModel):
    result: str
    length: int

@command("process_text")  # Auto-detects app name as "my_app"
def process_text(input_data: ProcessInput) -> ProcessOutput:
    result = input_data.message.upper() if input_data.uppercase else input_data.message
    return ProcessOutput(result=result, length=len(result))

# Alternative: explicit app name
@command("analyze", app="analytics")
def analyze_data(input_data: ProcessInput) -> ProcessOutput:
    return ProcessOutput(result=f"Analyzed: {input_data.message}", length=len(input_data.message))

from surreal_commands import submit_command, wait_for_command_sync

# Submit a command
cmd_id = submit_command("my_app", "process_text", {
    "message": "hello world", 
    "uppercase": True
})

print(f"Submitted command: {cmd_id}")

# Wait for completion
result = wait_for_command_sync(cmd_id, timeout=30)
if result.is_success():
    print(f"Result: {result.result}")

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

class AnalysisInput(BaseModel):
    data: List[float]
    method: str = Field(default="mean", description="Analysis method")
    threshold: Optional[float] = None

class AnalysisOutput(BaseModel):
    result: float
    method_used: str
    items_processed: int
    warnings: List[str] = []

def analyze_data(input_data: AnalysisInput) -> AnalysisOutput:
    # Your analysis logic here
    pass

async def async_process(input_data: MyInput) -> MyOutput:
    # Async processing
    await some_async_operation()
    return MyOutput(...)

# LangChain handles both sync and async
command = RunnableLambda(async_process)

from surreal_commands import command, CommandInput, CommandOutput, ExecutionContext

# Input that can access execution context
class ProcessInput(CommandInput):
    message: str
    uppercase: bool = False

# Output that includes execution metadata
class ProcessOutput(CommandOutput):
    result: str
    # command_id, execution_time, and execution_metadata are inherited

@command("process_with_context")
def process_with_context(input_data: ProcessInput) -> ProcessOutput:
    # Access execution context from input
    ctx = input_data.execution_context
    
    if ctx:
        command_id = ctx.command_id
        app_name = ctx.app_name
        user_context = ctx.user_context or {}
        user_id = user_context.get("user_id", "anonymous")
    else:
        command_id = "unknown"
        user_id = "anonymous"
    
    # Process the message
    result = input_data.message.upper() if input_data.uppercase else input_data.message
    result = f"Processed by {user_id}: {result}"
    
    # Return output - framework automatically populates:
    # - command_id (from execution context)
    # - execution_time (measured by framework)
    # - execution_metadata (additional context info)
    return ProcessOutput(result=result)

# Option 1: Access context only
class MyInput(CommandInput):
    data: str

@command("context_input_only")
def my_command(input_data: MyInput) -> RegularOutput:
    ctx = input_data.execution_context
    # ... use context ...
    return RegularOutput(result="done")

# Option 2: Output metadata only
class MyOutput(CommandOutput):
    result: str

@command("context_output_only")
def my_command(input_data: RegularInput) -> MyOutput:
    # ... process ...
    return MyOutput(result="done")  # Gets auto-populated metadata

# Option 3: Full context support
class MyInput(CommandInput):
    data: str

class MyOutput(CommandOutput):
    result: str

@command("full_context")
def my_command(input_data: MyInput) -> MyOutput:
    # ... access context and get metadata ...

# OLD (causes errors):
def old_command(input_data: MyInput, execution_context: ExecutionContext):
    command_id = execution_context.command_id
    # ...

# NEW (recommended):
class MyInput(CommandInput):
    # your fields here

def new_command(input_data: MyInput):
    command_id = input_data.execution_context.command_id if input_data.execution_context else None
    # ...

# check_results.py
import asyncio
from repository import db_connection

async def check_status():
    async with db_connection() as db:
        commands = await db.query(
            "SELECT * FROM command ORDER BY created DESC LIMIT 10"
        )
        for cmd in commands:
            print(f"{cmd['id']}: {cmd['status']}")

asyncio.run(check_status())

from surreal_commands import command

# Simple retry with 3 attempts
@command("process_data", retry={"max_attempts": 3, "wait_strategy": "fixed", "wait_time": 1})
def process_data(input_data: MyInput) -> MyOutput:
    # This command will retry up to 3 times with 1 second wait between attempts
    return MyOutput(result="processed")

# Fixed wait time (good for rate-limited APIs)
@command("api_call", retry={"max_attempts": 3, "wait_strategy": "fixed", "wait_time": 2})
def api_call(input_data: MyInput) -> MyOutput:
    pass

# Exponential backoff (recommended for network calls)
@command("network_call", retry={
    "max_attempts": 5,
    "wait_strategy": "exponential",
    "wait_min": 1,
    "wait_max": 60,
    "wait_multiplier": 2
})
def network_call(input_data: MyInput) -> MyOutput:
    # Wait times: 1s, 2s, 4s, 8s, 16s (capped at 60s)
    pass

# Exponential with jitter (prevents thundering herd)
@command("distributed_call", retry={"wait_strategy": "exponential_jitter"})
def distributed_call(input_data: MyInput) -> MyOutput:
    pass

# Retry only on specific exceptions
@command("selective_retry", retry={
    "max_attempts": 3,
    "retry_on": [ConnectionError, TimeoutError]
})
def selective_retry(input_data: MyInput) -> MyOutput:
    pass

# Don't retry on permanent errors
@command("skip_permanent", retry={
    "max_attempts": 3,
    "stop_on": [ValueError, TypeError, KeyError]
})
def skip_permanent(input_data: MyInput) -> MyOutput:
    pass

# Reduce logging for bulk operations with expected transient failures
@command("bulk_embed", retry={
    "max_attempts": 5,
    "wait_strategy": "exponential",
    "retry_log_level": "debug"  # Only show retries in debug mode
})
def bulk_embed(input_data: MyInput) -> MyOutput:
    pass

# Suppress all retry logs for high-volume background tasks
@command("background_sync", retry={
    "max_attempts": 3,
    "retry_log_level": "none"  # No retry logging
})
def background_sync(input_data: MyInput) -> MyOutput:
    pass

# Use warning level for critical operations
@command("critical_api_call", retry={
    "max_attempts": 3,
    "retry_log_level": "warning"  # Make retries more visible
})
def critical_api_call(input_data: MyInput) -> MyOutput:
    pass

# Explicitly disable retry for a command
@command("no_retry", retry=None)
def no_retry(input_data: MyInput) -> MyOutput:
    # Fails immediately without retry
    pass

# Or use enabled flag
@command("also_no_retry", retry={"enabled": False})
def also_no_retry(input_data: MyInput) -> MyOutput:
    pass

from surreal_commands import command, RetryConfig, RetryStrategy, RetryLogLevel

config = RetryConfig(
    enabled=True,
    max_attempts=5,
    wait_strategy=RetryStrategy.EXPONENTIAL,
    wait_min=1,
    wait_max=60,
    retry_log_level=RetryLogLevel.DEBUG  # Control log verbosity
)

@command("typed_retry", retry=config)
def typed_retry(input_data: MyInput) -> MyOutput:
    pass