from flask import Flask, render_template, request, redirect, url_for, send_file, flash
import os
import json
from datetime import datetime
import pandas as pd
from yt_dlp import YoutubeDL

app = Flask(__name__)
app.secret_key = "change_this_to_something_random"

HISTORY_FILE = "history.json"
OUTPUT_FILE = "output.xlsx"
MAX_NEW = 30  # 每次最多采集最新 30 个视频

# 确保 history 文件存在
if os.environ.get("RUN_SCHEDULER", "false") == "true":
    scheduler = BackgroundScheduler()
    scheduler.add_job(run_scheduled_scrape, 'interval', hours=get_schedule_hours())
    scheduler.start()


def load_history():
    with open(HISTORY_FILE, "r", encoding="utf-8") as f:
        return json.load(f)

def save_history(history):
    with open(HISTORY_FILE, "w", encoding="utf-8") as f:
        json.dump(history, f, ensure_ascii=False, indent=2)

def append_to_excel(rows):
    """
    将新抓取的记录写入两个文件：
      1. output.xlsx（累计追加）
      2. output_YYYYMMDD_HHMMSS.xlsx（当前批次独立保存）
    """
    df = pd.DataFrame(rows)
    timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
    time_file = f"output_{timestamp}.xlsx"

    # 写出当前批次独立文件
    df.to_excel(time_file, index=False)

    # 同时更新主文件 output.xlsx（追加）
    if os.path.exists(OUTPUT_FILE):
        old = pd.read_excel(OUTPUT_FILE)
        combined = pd.concat([old, df], ignore_index=True)
        combined.to_excel(OUTPUT_FILE, index=False)
    else:
        df.to_excel(OUTPUT_FILE, index=False)

    return time_file

def fetch_latest_videos(profile_url, max_items=10):
    ydl_opts = {
        "ignoreerrors": True,
        "quiet": True,
        "skip_download": True,
    }
    with YoutubeDL(ydl_opts) as ydl:
        info = ydl.extract_info(profile_url, download=False)
    entries = []
    if not info:
        return []
    if "entries" in info and info["entries"]:
        for e in info["entries"]:
            if e:
                entries.append(e)
    else:
        entries = [info]

    normalized = []
    for e in entries:
        video_id = e.get("id") or e.get("display_id") or e.get("url")
        title = e.get("title", "")
        webpage_url = e.get("webpage_url") or e.get("url")
        upload_date = e.get("upload_date")
        timestamp = e.get("timestamp")
        if timestamp:
            dt = datetime.utcfromtimestamp(timestamp)
            date_str = dt.strftime("%Y-%m-%d")
        elif upload_date:
            try:
                date_str = datetime.strptime(str(upload_date), "%Y%m%d").strftime("%Y-%m-%d")
            except Exception:
                date_str = str(upload_date)
        else:
            date_str = ""
        view_count = e.get("view_count") if e.get("view_count") is not None else e.get("views")
        like_count = e.get("like_count") if e.get("like_count") is not None else e.get("likes")
        normalized.append({
            "id": video_id,
            "title": title,
            "url": webpage_url,
            "date": date_str,
            "views": view_count,
            "likes": like_count
        })

    def sort_key(x):
        try:
            return datetime.strptime(x["date"], "%Y-%m-%d")
        except:
            return datetime.min
    normalized.sort(key=sort_key, reverse=True)
    return normalized[:max_items]

def list_output_files():
    """列出当前目录下所有 output*.xlsx 文件（带时间和大小）"""
    files = []
    for f in os.listdir("."):
        if f.startswith("output") and f.endswith(".xlsx"):
            size_kb = os.path.getsize(f) / 1024
            mtime = datetime.fromtimestamp(os.path.getmtime(f)).strftime("%Y-%m-%d %H:%M:%S")
            files.append({
                "name": f,
                "mtime": mtime,
                "size": f"{size_kb:.1f} KB"
            })
    # 按时间倒序
    files.sort(key=lambda x: x["mtime"], reverse=True)
    return files

@app.route("/", methods=["GET", "POST"])
def index():
    if request.method == "POST":
        profile_url = request.form.get("profile_url", "").strip()
        if not profile_url:
            flash("请先输入 TikTok 或 YouTube 的主页链接。")
            return redirect(url_for("index"))
        try:
            entries = fetch_latest_videos(profile_url, max_items=MAX_NEW)
        except Exception as ex:
            flash(f"抓取失败：{ex}")
            return redirect(url_for("index"))

        if not entries:
            flash("未能提取到视频信息，请确认主页链接是否有效。")
            return redirect(url_for("index"))

        history = load_history()
        key = profile_url
        seen = set(history.get(key, []))

        new_rows = []
        new_ids = []
        skipped = 0
        added = 0

        for e in entries:
            vid = e.get("id") or e.get("url")
            if not vid:
                continue
            if vid in seen:
                skipped += 1
                continue
            row = {
                "source_profile": profile_url,
                "video_id": vid,
                "date": e.get("date", ""),
                "title": e.get("title", ""),
                "video_url": e.get("url", ""),
                "views": e.get("views"),
                "likes": e.get("likes")
            }
            new_rows.append(row)
            new_ids.append(vid)
            seen.add(vid)
            added += 1

        if new_rows:
            time_file = append_to_excel(new_rows)
        else:
            time_file = None

        history[key] = list(seen)
        save_history(history)

        if new_rows:
            flash(f"抓取完成：新增 {added} 条，跳过 {skipped} 条。"
                  f"已保存至 output.xlsx 及 {time_file}。")
        else:
            flash(f"没有发现新视频，跳过 {skipped} 条。")

        return redirect(url_for("index"))

    files = list_output_files()
    return render_template("index.html", files=files)

@app.route("/download/<filename>")
def download(filename):
    if os.path.exists(filename):
        return send_file(filename, as_attachment=True)
    else:
        flash("文件不存在。")
        return redirect(url_for("index"))

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=7860, debug=False)