# NEWS_CATEGORIES 字典定义了分类名称和对应的网易新闻URL
NEWS_CATEGORIES = {
    "国内": "https://news.163.com/domestic/",
    # ... 其他分类
}

# HEADERS 模拟浏览器请求，避免被网站拒绝
HEADERS = {
    'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
    # ... 其他头部信息
}

# 禁用不安全请求警告
urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

import json
import requests
import urllib3
import time
import threading
from concurrent.futures import ThreadPoolExecutor, as_completed
from datetime import datetime
from bs4 import BeautifulSoup
import tkinter as tk
from tkinter import ttk, scrolledtext, messagebox, filedialog

urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

NEWS_CATEGORIES = {
    "国内": "https://news.163.com/domestic/",
    "国际": "https://news.163.com/world/",
    "体育": "https://sports.163.com/",
    "NBA": "https://sports.163.com/nba/",
    "娱乐": "https://ent.163.com/",
    "电影": "https://ent.163.com/movie/",
    "财经": "https://money.163.com/",
    "科技": "https://tech.163.com/",
    "互联网": "https://tech.163.com/internet/",
    "汽车": "https://auto.163.com/",
    "教育": "https://edu.163.com/",
}

HEADERS = {
    'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
    'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
    'Accept-Language': 'zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2',
    'Accept-Encoding': 'gzip, deflate, br',
    'Connection': 'keep-alive',
    'Upgrade-Insecure-Requests': '1',
}

def get_category_links(category_url, max_links):
    valid_links = []
    try:
        resp = requests.get(category_url, headers=HEADERS, verify=False, timeout=10)
        resp.encoding = "utf-8"

        if resp.status_code != 200:
            return valid_links

        html = BeautifulSoup(resp.text, "lxml")

        content_selectors = [
            "div.hidden", "div.news_list", "div.news-data",
            "ul.newsList", "div.post_list", "div[class*='news']",
            "div[class*='list']", "div.list_wrap"
        ]

        content = None
        for selector in content_selectors:
            content = html.select_one(selector)
            if content:
                break

        if content:
            links = content.find_all("a", href=True)

            for a in links:
                href = a.get("href", "")
                title = a.get_text(strip=True)
                if (href and (href.startswith('https://www.163.com/dy/article/') or '163.com' in href) and
                        title and len(title) > 5 and len(valid_links) < max_links):
                    if not any(l[1] == href for l in valid_links):
                        valid_links.append((title, href))

    except Exception:
        pass

    return valid_links[:max_links]


def get_news_content(url):
    try:
        subResponse = requests.get(url, headers=HEADERS, verify=False, timeout=10)
        subResponse.encoding = "utf-8"

        if subResponse.status_code != 200:
            return ""

        subHtml = BeautifulSoup(subResponse.text, "lxml")

        content_selectors = [
            "div.post_body", "div.content", "div.article-content",
            "div.art-content", "div.main-content", "div[class*='content']",
            "div[class*='body']", ".post_text"
        ]

        for selector in content_selectors:
            subContent = subHtml.select_one(selector)
            if subContent:
                text_content = subContent.get_text(separator='\n', strip=True)
                if len(text_content) > 50:
                    return text_content
        return ""

    except Exception:
        return ""


def summarize_with_ollama(title, content):
    try:
        url = "http://localhost:11434/api/generate"
        content_preview = content[:1500]
        prompt = f"请对以下新闻进行简要总结，不超过150字，重点突出，用清晰的中文口吻：\n标题：{title}\n内容：{content_preview}"

        data = {
            "model": "qwen2.5:7b",
            "prompt": prompt,
            "stream": True,
            "options": {"temperature": 0.3}
        }

        response = requests.post(url, json=data, stream=True, timeout=60)
        response.raise_for_status()

        full_response = ""
        for item in response.iter_lines():
            if item:
                try:
                    json_data = json.loads(item.decode())
                    if "response" in json_data:
                        full_response += json_data["response"]
                    if json_data.get("done"):
                        break
                except json.JSONDecodeError:
                    continue

        return full_response.strip()

    except requests.exceptions.ConnectionError:
        return "🚨 Ollama连接失败。请检查服务是否运行在 http://localhost:11434"
    except requests.exceptions.HTTPError as e:
        return f"🚨 AI总结HTTP错误: {e.response.status_code}"
    except Exception as e:
        return f"🚨 AI总结失败: {str(e)}"


class NewsCrawlerApp:
    def __init__(self, root):
        self.root = root
        self.root.title("网易新闻AI总结系统 (优化版)")
        self.root.geometry("1200x800")
        self.news_data = []
        self.is_crawling = False
        self.thread_executor = ThreadPoolExecutor(max_workers=10)
        self.crawl_futures = []
        self.setup_ui()

    def setup_ui(self):
        main_frame = ttk.Frame(self.root, padding="10")
        main_frame.pack(fill=tk.BOTH, expand=True)

        control_frame = ttk.Frame(main_frame)
        control_frame.pack(fill=tk.X, pady=(0, 10))

        row1_frame = ttk.Frame(control_frame)
        row1_frame.pack(fill=tk.X)

        ttk.Label(row1_frame, text="选择分类:").pack(side=tk.LEFT, padx=(0, 5))
        self.category_var = tk.StringVar(value="全部")
        self.category_combo = ttk.Combobox(row1_frame, textvariable=self.category_var,
                                           values=["全部"] + list(NEWS_CATEGORIES.keys()),
                                           state="readonly", width=15)
        self.category_combo.pack(side=tk.LEFT, padx=(0, 20))

        ttk.Label(row1_frame, text="每类数量:").pack(side=tk.LEFT, padx=(0, 5))
        self.count_var = tk.StringVar(value="10")
        self.count_combo = ttk.Combobox(row1_frame, textvariable=self.count_var,
                                        values=[str(i) for i in [5, 10, 20, 30, 50]],
                                        state="readonly", width=8)
        self.count_combo.pack(side=tk.LEFT, padx=(0, 20))

        self.ai_var = tk.BooleanVar(value=True)
        self.ai_check = ttk.Checkbutton(row1_frame, text="启用AI总结 (耗时)", variable=self.ai_var)
        self.ai_check.pack(side=tk.LEFT, padx=(0, 20))

        row2_frame = ttk.Frame(control_frame)
        row2_frame.pack(fill=tk.X, pady=(10, 0))

        self.start_btn = ttk.Button(row2_frame, text="🚀 开始爬取", command=self.toggle_crawling)
        self.start_btn.pack(side=tk.LEFT, padx=(0, 10))

        self.save_btn = ttk.Button(row2_frame, text="💾 保存结果", command=self.save_results, state='disabled')
        self.save_btn.pack(side=tk.LEFT, padx=(0, 10))

        self.clear_btn = ttk.Button(row2_frame, text="🗑️ 清空列表", command=self.clear_list)
        self.clear_btn.pack(side=tk.LEFT)

        self.progress = ttk.Progressbar(main_frame, mode='determinate')
        self.progress.pack(fill=tk.X, pady=(0, 5))

        self.status_var = tk.StringVar(value="准备就绪")
        self.status_label = ttk.Label(main_frame, textvariable=self.status_var, foreground="#007bff")
        self.status_label.pack(fill=tk.X)

        content_frame = ttk.Frame(main_frame)
        content_frame.pack(fill=tk.BOTH, expand=True)

        list_frame = ttk.LabelFrame(content_frame, text="新闻列表", padding="5")
        list_frame.pack(side=tk.LEFT, fill=tk.BOTH, expand=True, padx=(0, 5))

        self.news_list = tk.Listbox(list_frame, font=("Microsoft YaHei", 10))
        scrollbar_list = ttk.Scrollbar(list_frame, orient=tk.VERTICAL, command=self.news_list.yview)
        self.news_list.configure(yscrollcommand=scrollbar_list.set)

        self.news_list.pack(side=tk.LEFT, fill=tk.BOTH, expand=True)
        scrollbar_list.pack(side=tk.RIGHT, fill=tk.Y)

        self.news_list.bind('<<ListboxSelect>>', self.on_list_select)

        detail_frame = ttk.LabelFrame(content_frame, text="新闻详情", padding="5")
        detail_frame.pack(side=tk.RIGHT, fill=tk.BOTH, expand=True)

        self.detail_text = scrolledtext.ScrolledText(detail_frame, wrap=tk.WORD,
                                                     font=("Microsoft YaHei", 10),
                                                     width=60)
        self.detail_text.pack(fill=tk.BOTH, expand=True)

        self.stats_var = tk.StringVar(value="总新闻数: 0 | 已处理内容/AI: 0")
        self.stats_label = ttk.Label(main_frame, textvariable=self.stats_var)
        self.stats_label.pack(fill=tk.X, pady=(5, 0))

    def toggle_crawling(self):
        if self.is_crawling:
            self.stop_crawling()
        else:
            self.start_crawling()

    def start_crawling(self):
        selected_category = self.category_var.get()
        categories_to_crawl = list(NEWS_CATEGORIES.keys()) if selected_category == "全部" else [selected_category]

        max_news = int(self.count_var.get())
        enable_ai = self.ai_var.get()

        if self.is_crawling:
            return

        self.clear_list(silent=True)
        self.is_crawling = True
        self.start_btn.config(text="停止爬取", state='normal')
        self.save_btn.config(state='disabled')
        self.progress['value'] = 0
        self.crawl_futures = []

        self.update_status("🚀 **开始目录爬取**...")

        self.master_thread = threading.Thread(
            target=self._run_master_crawl,
            args=(categories_to_crawl, max_news, enable_ai)
        )
        self.master_thread.daemon = True
        self.master_thread.start()

    def stop_crawling(self):
        if not self.is_crawling:
            return

        self.is_crawling = False

        for future in self.crawl_futures:
            future.cancel()

        self.start_btn.config(text="开始爬取")
        self.save_btn.config(state='normal')
        self.update_status("🛑 **爬取已停止**")
        self.update_progress(0)

    def _run_master_crawl(self, categories, max_news, enable_ai):
        initial_news_count = 0

        try:
            self.update_status("✅ **阶段一：正在快速获取新闻目录...**")

            for index, category in enumerate(categories):
                if not self.is_crawling:
                    break

                category_url = NEWS_CATEGORIES[category]
                links = get_category_links(category_url, max_news)

                progress = int(((index + 1) / len(categories)) * 30)
                self.root.after(0, self.update_progress, progress)

                for title, url in links:
                    news_data = {
                        'category': category,
                        'title': title,
                        'url': url,
                        'content': '等待获取内容...',
                        'summary': '等待AI总结...' if enable_ai else 'AI已禁用',
                        'timestamp': datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
                        'list_index': len(self.news_data)
                    }
                    self.news_data.append(news_data)
                    self.root.after(0, self._add_news_to_list, news_data)
                    initial_news_count += 1

                time.sleep(0.5)

            if not self.is_crawling:
                return

            self.root.after(0, self.update_stats, initial_news_count, 0)
            self.update_status(f"🎉 **阶段一完成！** 找到 {initial_news_count} 条新闻。开始内容/AI处理...")

            total_tasks = initial_news_count
            if total_tasks == 0:
                self.root.after(0, self.crawling_finished)
                return

            processed_count = 0

            for data in self.news_data:
                if not self.is_crawling:
                    break
                future = self.thread_executor.submit(
                    self._process_single_news,
                    data,
                    enable_ai
                )
                self.crawl_futures.append(future)

            for future in as_completed(self.crawl_futures):
                if not self.is_crawling:
                    break

                if future.cancelled():
                    continue

                processed_data = future.result()

                if processed_data:
                    self.root.after(0, self._update_list_item, processed_data)

                processed_count += 1

                progress = 30 + int((processed_count / total_tasks) * 70)
                self.root.after(0, self.update_progress, progress)
                self.root.after(0, self.update_stats, total_tasks, processed_count)

        except Exception as e:
            self.root.after(0, self.update_status, f"🚨 **主爬取线程出错**: {str(e)}")

        finally:
            self.root.after(0, self.crawling_finished)

    def _process_single_news(self, news_data, enable_ai):
        if not self.is_crawling:
            return None

        content = get_news_content(news_data['url'])
        news_data['content'] = content if content else '获取内容失败或内容过短。'

        if content and enable_ai:
            summary = summarize_with_ollama(news_data['title'], content)
            news_data['summary'] = summary
        elif not enable_ai:
            news_data['summary'] = 'AI总结已禁用。'
        else:
            news_data['summary'] = '未能获取内容，无法进行AI总结。'

        news_data['finish_time'] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

        return news_data

    def _add_news_to_list(self, news_data):
        display_text = f"[{news_data['category']}] {news_data['title']}"
        self.news_list.insert(tk.END, display_text)
        self.news_list.itemconfig(tk.END, {'fg': 'gray'})

    def _update_list_item(self, processed_data):
        index = processed_data['list_index']
        self.news_data[index] = processed_data

        if "🚨" in processed_data['summary'] or not processed_data['content']:
            color = 'red'
        elif "等待" in processed_data['summary'] or "已禁用" in processed_data['summary']:
            color = 'blue'
        else:
            color = 'darkgreen'

        self.news_list.itemconfig(index, {'fg': color})

        selection = self.news_list.curselection()
        if selection and selection[0] == index:
            self.show_news_detail(index)

    def on_list_select(self, event):
        selection = self.news_list.curselection()
        if selection:
            index = selection[0]
            if 0 <= index < len(self.news_data):
                self.show_news_detail(index)

    def show_news_detail(self, index):
        news = self.news_data[index]

        summary_color = 'darkgreen' if "🚨" not in news['summary'] and "等待" not in news['summary'] else 'red'

        detail_text = f"""【分类】{news['category']}
【标题】{news['title']}
【获取时间】{news['timestamp']}
【处理完成】{news.get('finish_time', '未完成')}
【链接】{news['url']}

------------------------------------------
【AI总结】
{news['summary']}
------------------------------------------

【内容预览】
{news['content'][:800] + '...\n\n(内容可能已被截断)' if len(news['content']) > 800 else news['content']}
"""

        self.detail_text.delete(1.0, tk.END)
        self.detail_text.insert(1.0, detail_text)

        start_index = self.detail_text.search("【AI总结】", "1.0", stopindex=tk.END)
        if start_index:
            end_index = self.detail_text.search("------------------------------------------", start_index + "+1c",
                                                stopindex=tk.END)
            self.detail_text.tag_config("summary_tag", foreground=summary_color, font=("Microsoft YaHei", 10, "bold"))
            self.detail_text.tag_add("summary_tag", start_index, end_index)

    def update_status(self, message):
        def _update():
            self.status_var.set(message)
            self.status_label.config(foreground='red' if '🚨' in message or '🛑' in message else '#007bff')
        self.root.after(0, _update)

    def update_progress(self, value):
        def _update():
            self.progress['value'] = value
        self.root.after(0, _update)

    def update_stats(self, total, processed):
        self.stats_var.set(f"总新闻数: {total} | 已处理内容/AI: {processed}")

    def crawling_finished(self):
        self.is_crawling = False
        self.start_btn.config(text="🚀 开始爬取", state='normal')
        self.save_btn.config(state='normal')
        self.update_status(f"🎉 **所有任务完成！** 总计 {len(self.news_data)} 条新闻已处理。")
        self.update_progress(100)

    def save_results(self):
        if not self.news_data:
            messagebox.showwarning("警告", "没有数据可保存！")
            return

        finished_data = [d for d in self.news_data if d.get('finish_time')]
        if not finished_data:
            messagebox.showwarning("警告", "没有已完成处理（内容和AI）的数据可保存！")
            return

        try:
            initial_filename = datetime.now().strftime("%Y%m%d_%H%M%S_news.md")
            file_path = filedialog.asksaveasfilename(
                defaultextension=".md",
                initialfile=initial_filename,
                filetypes=[("Markdown files", "*.md"), ("All files", "*.*")]
            )

            if not file_path:
                return

            with open(file_path, 'w', encoding='utf-8') as f:
                f.write(f"# 网易新闻AI总结报告\n\n")
                f.write(f"生成时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
                f.write(f"新闻数量: {len(finished_data)}\n\n")

                categories = {}
                for news in finished_data:
                    categories.setdefault(news['category'], []).append(news)

                for category, news_list in categories.items():
                    f.write(f"## {category}\n\n")
                    f.write("---\n\n")

                    for i, news in enumerate(news_list, 1):
                        f.write(f"### {i}. {news['title']}\n\n")
                        f.write(f"- **获取时间**: {news['timestamp']}\n")
                        f.write(f"- **处理完成**: {news.get('finish_time')}\n")
                        f.write(f"- **原文链接**: <{news['url']}>\n\n")

                        f.write("#### AI总结\n")
                        summary_lines = news['summary'].split('\n')
                        for line in summary_lines:
                            f.write(f"> {line.strip()}\n")
                        f.write("\n")

                        f.write("#### 内容预览 (前 200 字符)\n")
                        content_preview = news['content'][:200].replace('\n', ' ') + '...' if len(
                            news['content']) > 200 else news['content'].replace('\n', ' ')
                        f.write(f"**内容预览**: {content_preview}\n\n")
                        f.write("---\n\n")

            messagebox.showinfo("成功", f"结果已保存到: {file_path}")
            self.update_status(f"💾 结果已保存到: {file_path}")

        except Exception as e:
            messagebox.showerror("错误", f"保存文件时出错: {str(e)}")

    def clear_list(self, silent=False):
        self.news_data.clear()
        self.news_list.delete(0, tk.END)
        self.detail_text.delete(1.0, tk.END)
        self.update_stats(0, 0)

        if not silent:
            self.update_status("列表已清空")


def main():
    root = tk.Tk()

    try:
        root.tk.call('source', 'azure.tcl')
        root.tk.call('set_theme', 'dark')
    except Exception:
        pass

    app = NewsCrawlerApp(root)
    root.mainloop()


if __name__ == "__main__":
    main()