main.py

#!/usr/bin/env python3
"""
Ogle — 学术搜索爬虫
用法:
    python main.py                          # 启动 GUI
    python main.py --cli                    # 命令行模式（读取 keywords.txt）
    python main.py --cli "人工智能" "深度学习"  # 命令行模式 + 关键词
"""
import sys
from pathlib import Path

# 确保 src 目录在 sys.path 中
sys.path.insert(0, str(Path(__file__).parent))

from src.config_loader import load_config
from src.crawler import run_crawler
from src.html_writer import write_results


def parse_keywords_arg(args: list) -> list:
    """从命令行参数或 keywords.txt 读取关键词。"""
    keywords = []
    for a in args:
        if a.startswith("--"):
            continue
        keywords.append(a)

    if not keywords:
        txt_path = Path(__file__).parent / "keywords.txt"
        if txt_path.exists():
            lines = txt_path.read_text(encoding="utf-8").strip().splitlines()
            keywords = [line.strip() for line in lines
                        if line.strip() and not line.strip().startswith("#")]
    return keywords


def main():
    args = sys.argv[1:]

    # --cli 模式
    if "--cli" in args:
        keywords = parse_keywords_arg(args)
        if not keywords:
            print("❌ 没有提供关键词。用法: python main.py --cli 关键词1 关键词2")
            print("   或在 keywords.txt 中每行写一个关键词")
            return

        print(f"📋 关键词 ({len(keywords)}): {', '.join(keywords)}")
        config = load_config()
        print(f"🚀 开始搜索...\n")

        def cli_progress(collected, max_total, keyword):
            """CLI 进度回调：每获取一条结果立刻刷新。"""
            bar_width = 30
            ratio = collected / max_total if max_total > 0 else 1
            filled = int(bar_width * min(ratio, 1))
            bar = "█" * filled + "░" * (bar_width - filled)
            print(f"\r   [{bar}] {collected}/{max_total} 条  ← {keyword}", end="", flush=True)

        results = run_crawler(keywords, config, progress_callback=cli_progress)

        total = sum(len(v) for v in results.values())
        output_dir = write_results(results, config)
        print(f"\n✅ 完成！共 {total} 条结果，保存至: {output_dir.resolve()}")
        print(f"   汇总页: {output_dir / 'all_results' / 'index.html'}")
        return

    # 默认启动 GUI
    from src.gui import launch_gui
    launch_gui()


if __name__ == "__main__":
    main()