shared/utils.py

"""Utility functions for LLM Code Deployment system."""
import base64
import hashlib
import re
from datetime import datetime
from typing import Any

from uuid_utils import uuid7


def generate_nonce() -> str:
    """Generate a unique nonce using UUID7."""
    return str(uuid7())


def generate_task_id(template_id: str, brief: str, attachments: list[dict[str, Any]]) -> str:
    """Generate a unique task ID based on template and content."""
    content = f"{brief}{str(attachments)}"
    hash_value = hashlib.sha256(content.encode()).hexdigest()[:5]
    return f"{template_id}-{hash_value}"


def decode_data_uri(data_uri: str) -> tuple[str, bytes]:
    """Decode a data URI into mime type and content."""
    # Format: data:mime/type;base64,content
    match = re.match(r"data:([^;]+);base64,(.+)", data_uri)
    if not match:
        raise ValueError(f"Invalid data URI format: {data_uri[:50]}...")

    mime_type = match.group(1)
    encoded_content = match.group(2)
    content = base64.b64decode(encoded_content)

    return mime_type, content


def encode_data_uri(mime_type: str, content: bytes) -> str:
    """Encode content as a data URI."""
    encoded = base64.b64encode(content).decode("utf-8")
    return f"data:{mime_type};base64,{encoded}"


def sanitize_repo_name(name: str) -> str:
    """Sanitize a string for use as a GitHub repository name."""
    # Replace spaces and special chars with hyphens
    sanitized = re.sub(r"[^a-zA-Z0-9-_.]", "-", name)
    # Remove leading/trailing hyphens
    sanitized = sanitized.strip("-")
    # Collapse multiple hyphens
    sanitized = re.sub(r"-+", "-", sanitized)
    # Ensure lowercase
    return sanitized.lower()


def get_timestamp() -> datetime:
    """Get current UTC timestamp."""
    return datetime.utcnow()