gogo2/COBY/utils/validation.py

"""
Data validation utilities for the COBY system.
"""

import re
from typing import List, Optional
from decimal import Decimal, InvalidOperation


def validate_symbol(symbol: str) -> bool:
    """
    Validate trading symbol format.
    
    Args:
        symbol: Trading symbol to validate
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not symbol or not isinstance(symbol, str):
        return False
    
    # Basic symbol format validation (e.g., BTCUSDT, ETH-USD)
    pattern = r'^[A-Z0-9]{2,10}[-/]?[A-Z0-9]{2,10}$'
    return bool(re.match(pattern, symbol.upper()))


def validate_price(price: float) -> bool:
    """
    Validate price value.
    
    Args:
        price: Price to validate
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not isinstance(price, (int, float, Decimal)):
        return False
    
    try:
        price_decimal = Decimal(str(price))
        return price_decimal > 0 and price_decimal < Decimal('1e10')  # Reasonable upper bound
    except (InvalidOperation, ValueError):
        return False


def validate_volume(volume: float) -> bool:
    """
    Validate volume value.
    
    Args:
        volume: Volume to validate
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not isinstance(volume, (int, float, Decimal)):
        return False
    
    try:
        volume_decimal = Decimal(str(volume))
        return volume_decimal >= 0 and volume_decimal < Decimal('1e15')  # Reasonable upper bound
    except (InvalidOperation, ValueError):
        return False


def validate_exchange_name(exchange: str) -> bool:
    """
    Validate exchange name.
    
    Args:
        exchange: Exchange name to validate
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not exchange or not isinstance(exchange, str):
        return False
    
    # Exchange name should be alphanumeric with possible underscores/hyphens
    pattern = r'^[a-zA-Z0-9_-]{2,20}$'
    return bool(re.match(pattern, exchange))


def validate_timestamp_range(start_time, end_time) -> List[str]:
    """
    Validate timestamp range.
    
    Args:
        start_time: Start timestamp
        end_time: End timestamp
        
    Returns:
        List[str]: List of validation errors (empty if valid)
    """
    errors = []
    
    if start_time is None:
        errors.append("Start time cannot be None")
    
    if end_time is None:
        errors.append("End time cannot be None")
    
    if start_time and end_time and start_time >= end_time:
        errors.append("Start time must be before end time")
    
    return errors


def validate_bucket_size(bucket_size: float) -> bool:
    """
    Validate price bucket size.
    
    Args:
        bucket_size: Bucket size to validate
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not isinstance(bucket_size, (int, float, Decimal)):
        return False
    
    try:
        size_decimal = Decimal(str(bucket_size))
        return size_decimal > 0 and size_decimal <= Decimal('1000')  # Reasonable upper bound
    except (InvalidOperation, ValueError):
        return False


def validate_speed_multiplier(speed: float) -> bool:
    """
    Validate replay speed multiplier.
    
    Args:
        speed: Speed multiplier to validate
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not isinstance(speed, (int, float)):
        return False
    
    return 0.01 <= speed <= 100.0  # 1% to 100x speed


def sanitize_symbol(symbol: str) -> str:
    """
    Sanitize and normalize symbol format.
    
    Args:
        symbol: Symbol to sanitize
        
    Returns:
        str: Sanitized symbol
    """
    if not symbol:
        return ""
    
    # Remove whitespace and convert to uppercase
    sanitized = symbol.strip().upper()
    
    # Remove invalid characters
    sanitized = re.sub(r'[^A-Z0-9/-]', '', sanitized)
    
    return sanitized


def validate_percentage(value: float, min_val: float = 0.0, max_val: float = 100.0) -> bool:
    """
    Validate percentage value.
    
    Args:
        value: Percentage value to validate
        min_val: Minimum allowed value
        max_val: Maximum allowed value
        
    Returns:
        bool: True if valid, False otherwise
    """
    if not isinstance(value, (int, float)):
        return False
    
    return min_val <= value <= max_val


def validate_connection_config(config: dict) -> List[str]:
    """
    Validate connection configuration.
    
    Args:
        config: Configuration dictionary
        
    Returns:
        List[str]: List of validation errors (empty if valid)
    """
    errors = []
    
    # Required fields
    required_fields = ['host', 'port']
    for field in required_fields:
        if field not in config:
            errors.append(f"Missing required field: {field}")
    
    # Validate host
    if 'host' in config:
        host = config['host']
        if not isinstance(host, str) or not host.strip():
            errors.append("Host must be a non-empty string")
    
    # Validate port
    if 'port' in config:
        port = config['port']
        if not isinstance(port, int) or not (1 <= port <= 65535):
            errors.append("Port must be an integer between 1 and 65535")
    
    return errors