LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows
Python ai-agentsdata-engineeringdocument-aigenerative-aiidpjson-extractionllmmcp-serverocrpdf-extractionprompt-engineeringstructured-output
"""Security Configuration Constants
Security and validation configuration.
"""
class SecurityConfig:
"""Security and validation configuration."""
# Input validation patterns
VALID_UUID_PATTERN = r"^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$"
VALID_ORGANIZATION_ID_PATTERN = r"^[a-zA-Z0-9_-]+$"
VALID_FILE_NAME_PATTERN = r'^[^<>:"/\\|?*\x00-\x1f]+$'
# Maximum field lengths
MAX_ERROR_MESSAGE_LENGTH = 2048
MAX_TASK_NAME_LENGTH = 100
MAX_FILE_NAME_LENGTH = 255
MAX_FILE_PATH_LENGTH = 4096
# Allowed characters
SAFE_FILENAME_CHARS = (
"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789._- "
)
# Security headers for API requests
REQUIRED_HEADERS = [
"Authorization",
"Content-Type",
]
# Rate limiting
API_RATE_LIMIT_PER_MINUTE = 1000
WEBHOOK_RATE_LIMIT_PER_MINUTE = 100