import os
import glob
from typing import Dict, Any, List, Optional


def count_lines(file_path: str) -> int:
    """Count lines in a file."""
    try:
        with open(file_path, "r", encoding="utf-8", errors="ignore") as f:
            return len(f.readlines())
    except Exception:
        return 0


def get_language(file_path: str) -> str:
    """Determine language from file extension."""
    ext = os.path.splitext(file_path)[1].lower()
    language_map = {
        ".py": "Python",
        ".cs": "C#",
        ".js": "JavaScript",
        ".ts": "TypeScript",
        ".html": "HTML",
        ".css": "CSS",
        ".md": "Markdown",
        ".json": "JSON",
        ".xml": "XML",
        ".java": "Java",
        ".cpp": "C++",
        ".c": "C",
        ".h": "C/C++ Header",
        ".go": "Go",
        ".rs": "Rust",
        ".php": "PHP",
        ".rb": "Ruby",
        ".sh": "Shell",
        ".bat": "Batch",
        ".ps1": "PowerShell",
        ".sql": "SQL",
        ".yaml": "YAML",
        ".yml": "YAML",
    }
    return language_map.get(ext, "Other")


def analyze_code_stats(
    root_dir: str = ".",
    exclude_dirs: List[str] = None,
    source_extensions: List[str] = None,
    top_files_limit: int = 20,
) -> Dict[str, Any]:
    """Analyze code statistics for the project."""
    if exclude_dirs is None:
        exclude_dirs = ["node_modules", "venv", ".git", ".vs", "bin", "obj"]

    if source_extensions is None:
        source_extensions = [
            ".py",
            ".cs",
            ".js",
            ".ts",
            ".html",
            ".css",
            ".md",
        ]

    stats = {
        "total_lines": 0,
        "file_count": 0,
        "languages": {},
        "top_files": [],
    }

    file_stats = []

    # Convert exclude_dirs to absolute paths
    exclude_paths = [os.path.join(root_dir, d) for d in exclude_dirs]

    # Find all source files
    for ext in source_extensions:
        pattern = os.path.join(root_dir, "**", f"*{ext}")
        for file_path in glob.glob(pattern, recursive=True):
            # Skip excluded directories
            if any(
                file_path.startswith(exclude_path)
                for exclude_path in exclude_paths
            ):
                continue

            # Count lines
            lines = count_lines(file_path)

            # Get language
            language = get_language(file_path)

            # Update statistics
            stats["total_lines"] += lines
            stats["file_count"] += 1

            # Update language count
            if language not in stats["languages"]:
                stats["languages"][language] = 0
            stats["languages"][language] += 1

            # Add to file stats
            file_stats.append(
                {
                    "path": file_path,
                    "name": os.path.basename(file_path),
                    "lines": lines,
                    "language": language,
                }
            )

    # Sort file stats by line count
    file_stats.sort(key=lambda x: x["lines"], reverse=True)

    # Get top files
    stats["top_files"] = file_stats[:top_files_limit]

    return stats