src/utils/__init__.py

"""Language Detection Utility"""
import re
from typing import Optional

def detect_language(code: str) -> str:
    """
    Auto-detect programming language from code content.
    
    Args:
        code: Source code string
        
    Returns:
        str: Detected language (python, javascript, typescript, or unknown)
    """
    # Python indicators
    python_patterns = [
        r'\bdef\s+\w+\s*\(',
        r'\bclass\s+\w+\s*:',
        r'\bimport\s+\w+',
        r'\bfrom\s+\w+\s+import',
        r'__init__',
        r'self\.',
    ]
    
    # JavaScript indicators
    js_patterns = [
        r'\bfunction\s+\w+\s*\(',
        r'\bconst\s+\w+\s*=',
        r'\blet\s+\w+\s*=',
        r'\bvar\s+\w+\s*=',
        r'=>',
        r'console\.log',
    ]
    
    # TypeScript indicators
    ts_patterns = [
        r':\s*(string|number|boolean|any|void)',
        r'\binterface\s+\w+',
        r'\btype\s+\w+\s*=',
        r'<\w+>',
    ]
    
    # Count matches
    python_score = sum(1 for pattern in python_patterns if re.search(pattern, code))
    js_score = sum(1 for pattern in js_patterns if re.search(pattern, code))
    ts_score = sum(1 for pattern in ts_patterns if re.search(pattern, code))
    
    # TypeScript is superset of JavaScript, so add JS score
    if ts_score > 0:
        ts_score += js_score
    
    # Determine winner
    scores = {
        "python": python_score,
        "javascript": js_score,
        "typescript": ts_score
    }
    
    max_score = max(scores.values())
    if max_score == 0:
        return "unknown"
    
    return max(scores, key=scores.get)