#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
compare_manifests.py

读取三个 filesystem.manifest 文件，比较包的并集，并输出包含发行版勾选标记、Priority、Section、Description 的 Markdown 表格。
增加详细日志输出，并使用多线程加速 apt show 调用。
修复：自动剥离包名中的 ":<arch>" 后缀。
"""
import sys
import subprocess
import logging
import concurrent.futures
import time

# 日志配置
logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s [%(levelname)s] %(message)s",
    datefmt="%Y-%m-%d %H:%M:%S"
)


def parse_manifest(path):
    """
    读取 manifest 文件，返回其中所有包名的集合。
    自动剥离包名中的 ":<arch>" 后缀。
    """
    pkgs = set()
    logging.info(f"Parsing manifest: {path}")
    try:
        with open(path, encoding='utf-8') as f:
            for line in f:
                line = line.strip()
                if not line or line.startswith('#'):
                    continue
                raw_pkg = line.split()[0]
                pkg = raw_pkg.split(':', 1)[0]
                pkgs.add(pkg)
        logging.info(f"Found {len(pkgs)} unique packages in {path}")
    except Exception as e:
        logging.error(f"Error reading manifest {path}: {e}")
        sys.exit(1)
    return pkgs


def get_pkg_info(pkg):
    """
    调用 `apt show pkg`，解析 Priority、Section、Description（首段），
    返回 dict 包含信息。
    若调用失败，返回空字段。
    """
    logging.debug(f"Fetching apt info for package: {pkg}")
    try:
        p = subprocess.run(
            ["apt", "show", pkg],
            stdout=subprocess.PIPE,
            stderr=subprocess.DEVNULL,
            text=True,
            encoding='utf-8',
            timeout=30
        )
    except subprocess.TimeoutExpired:
        logging.warning(f"Timeout fetching info for {pkg}")
        return {"Priority": "", "Section": "", "Description": ""}

    if p.returncode != 0:
        logging.debug(f"apt show failed for {pkg}")
        return {"Priority": "", "Section": "", "Description": ""}

    priority = ""
    section = ""
    desc_lines = []
    in_desc = False

    for line in p.stdout.splitlines():
        if in_desc:
            if not line.strip():
                break
            desc_lines.append(line.strip())
        else:
            if line.startswith("Priority:"):
                priority = line.split(":", 1)[1].strip()
            elif line.startswith("Section:"):
                section = line.split(":", 1)[1].strip()
            elif line.startswith("Description:"):
                desc = line.split(":", 1)[1].strip()
                desc_lines.append(desc)
                in_desc = True

    full_desc = "  ".join(desc_lines)
    logging.debug(f"Obtained info for {pkg}: priority={priority}, section={section}")
    return {"Priority": priority, "Section": section, "Description": full_desc}


def main():
    if len(sys.argv) != 4:
        print(f"Usage: {sys.argv[0]} <ubuntu_manifest> <zorin_manifest> <anduin_manifest>")
        sys.exit(1)

    start_time = time.time()
    paths = {
        "Ubuntu": sys.argv[1],
        "Zorin": sys.argv[2],
        "Anduin": sys.argv[3],
    }

    # 解析 manifests
    sets = {name: parse_manifest(path) for name, path in paths.items()}
    all_pkgs = sorted(set.union(*sets.values()))
    total = len(all_pkgs)
    logging.info(f"Total unique packages to process: {total}")

    # 多线程获取 apt 信息
    pkg_infos = {}
    max_workers = min(10, total)
    logging.info(f"Starting ThreadPoolExecutor with {max_workers} workers")

    with concurrent.futures.ThreadPoolExecutor(max_workers=max_workers) as executor:
        future_to_pkg = {executor.submit(get_pkg_info, pkg): pkg for pkg in all_pkgs}
        for idx, future in enumerate(concurrent.futures.as_completed(future_to_pkg), 1):
            pkg = future_to_pkg[future]
            try:
                pkg_infos[pkg] = future.result()
                logging.info(f"[{idx}/{total}] Processed {pkg}")
            except Exception as e:
                logging.error(f"Error processing {pkg}: {e}")
                pkg_infos[pkg] = {"Priority": "", "Section": "", "Description": ""}

    # 写入 Markdown
    out_path = "comp_result.md"
    logging.info(f"Writing results to {out_path}")
    with open(out_path, "w", encoding="utf-8") as out:
        # 新表头：检查标记 -> Priority -> Section -> Description
        out.write("| Package | Ubuntu | Zorin | Anduin | Priority | Section | Description |\n")
        out.write("|---------|--------|-------|--------|----------|---------|-------------|\n")

        for pkg in all_pkgs:
            info = pkg_infos.get(pkg, {"Priority": "", "Section": "", "Description": ""})
            mark = lambda name: "√" if pkg in sets[name] else ""
            desc = info["Description"].replace("|", "\\|")
            out.write(
                f"| {pkg} | {mark('Ubuntu')} | {mark('Zorin')} | {mark('Anduin')} |"
                f" {info['Priority']} | {info['Section']} | {desc} |\n"
            )

    elapsed = time.time() - start_time
    logging.info(f"Completed in {elapsed:.2f}s")

if __name__ == "__main__":
    main()