import os
import io
import requests
from bs4 import BeautifulSoup
import gradio as gr
import pandas as pd
import time

def get_papers_since(url, since_year=2023):
    """
    给定一个 Google Scholar citations page URL，返回指定年份（默认2023）及之后的论文列表。
    
    每篇论文用一个字典表示：{"title": ..., "year": ...}
    """
    headers = {
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"
    }
    
    papers = []
    start = 0
    pagesize = 20  # Google Scholar 通常每页显示 20 条记录
    
    while True:
        params = {"cstart": start, "pagesize": pagesize}
        response = requests.get(url, params=params, headers=headers)
        if response.status_code != 200:
            break
        
        soup = BeautifulSoup(response.text, "html.parser")
        rows = soup.find_all("tr", class_="gsc_a_tr")
        if not rows:
            break
        
        for row in rows:
            title_tag = row.find("a", class_="gsc_a_at")
            title = title_tag.text.strip() if title_tag else "Unknown Title"
            year_tag = row.find("span", class_="gsc_a_h")
            try:
                year = int(year_tag.text.strip()) if year_tag else None
            except ValueError:
                year = None
            
            if year and year >= since_year:
                papers.append({"title": title, "year": year})
        
        next_button = soup.find("button", id="gsc_bpf_more")
        if next_button and "disabled" not in next_button.attrs:
            start += pagesize
        else:
            break
    
    return papers

def search_paper_by_title(paper_title: str):
    """
    使用 Semantic Scholar "match" endpoint 根据标题搜索论文，返回 top-1 匹配结果。
    
    返回:
    dict: 包含 'paperId', 'title', 'abstract', 'venue' 四个键的字典。
          若找不到对应论文（HTTP 404 或无结果），则返回 None。
          若出现其它错误则抛出异常。
    """
    base_url = "https://api.semanticscholar.org/graph/v1/paper/search/match"
    params = {
        "query": paper_title,
        "fields": "paperId,abstract,title,venue"
    }
    headers = {
        "x-api-key": os.getenv("S2_API_KEY")
    }
    
    response = requests.get(base_url, params=params, headers=headers)

    if response.status_code == 404:
        return None
    elif response.status_code != 200:
        raise Exception(f"错误 {response.status_code}: {response.text}")
    
    data_list = response.json().get("data", [])
    if not data_list:
        return None
    
    paper_data = data_list[0]
    return {
        "paperId": paper_data.get("paperId"),
        "title": paper_data.get("title"),
        "abstract": paper_data.get("abstract"),
        "venue": paper_data.get("venue")
    }

def process_profiles(profiles_text, wechat, progress=None):
    """
    1. 将用户输入的多个 Google Scholar profile links (以换行分隔) 转为列表
    2. 爬取所有链接在 2023 年及之后的论文
    3. 对每篇论文调用 Semantic Scholar 获取匹配信息，若无匹配则丢弃
    4. 根据 paperId 去重
    5. 返回一个 CSV 文件内容（bytes），用于用户下载
    """
    log_messages = ["开始处理..."]

    def update_log(message):
        log_messages.append(message)
        return "\n".join(log_messages)
    
    if not profiles_text.strip():
        return update_log("错误: 未提供任何链接"), None

    if not wechat.strip():
        return update_log("错误: 未提供微信号"), None

    # 将用户输入按换行拆分
    profile_urls = [line.strip() for line in profiles_text.splitlines() if line.strip()]
    
    message = f"已识别 {len(profile_urls)} 个档案链接，开始处理..."
    yield update_log(message), None

    all_papers = []
    for i, url in enumerate(profile_urls):
        message = f"处理第 {i+1}/{len(profile_urls)} 个档案: {url}"
        yield update_log(message), None
        
        papers = get_papers_since(url, 2023)
        all_papers.extend(papers)
        
        message = f"已从档案 {i+1}/{len(profile_urls)} 中收集 {len(papers)} 篇论文"
        yield update_log(message), None

    message = f"共收集到 {len(all_papers)} 篇论文，开始获取详细信息..."
    yield update_log(message), None

    paperid_map = {}
    total_papers = len(all_papers)
    
    for i, paper in enumerate(all_papers):
        title = paper["title"]
        year = paper["year"]
        
        if i % 10 == 0 or i == total_papers - 1:
            percent = round((i+1)/total_papers*100)
            message = f"正在处理论文 {i+1}/{total_papers} ({percent}% 完成)"
            yield update_log(message), None
        
        if i > 0 and i % 10 == 0:
            time.sleep(1)
        
        try:
            time.sleep(0.3)
            paper_info = search_paper_by_title(title)
        except Exception as e:
            continue
        
        if not paper_info or not paper_info.get("paperId"):
            continue

        pid = paper_info["paperId"]
        if pid not in paperid_map:
            paperid_map[pid] = {
                "paperId": pid,
                "title": paper_info["title"],
                "abstract": paper_info["abstract"],
                "venue": paper_info["venue"],
                "year": year
            }
    
    if not paperid_map:
        message = "错误: 未找到任何匹配的论文"
        return update_log(message), None

    message = f"去重后共有 {len(paperid_map)} 篇论文，正在生成CSV..."
    yield update_log(message), None

    df = pd.DataFrame(list(paperid_map.values()))

    # 使用微信号作为 CSV 文件名
    temp_csv_path = f"{wechat.strip()}.csv"
    df.to_csv(temp_csv_path, index=False, encoding='utf-8')

    message = f"任务完成! 已生成包含 {len(paperid_map)} 篇论文的CSV文件: {temp_csv_path}"
    yield update_log(message), temp_csv_path

def build_app():
    """
    使用 gradio 搭建一个小型 Demo app。
    """
    with gr.Blocks() as demo:
        gr.Markdown("## Google Scholar & Semantic Scholar 信息整合工具")
        gr.Markdown("在下方输入任意多个 Google Scholar Profile 链接，每行一个，然后输入微信号，点击 **开始爬取**。")

        profile_input = gr.Textbox(
            lines=5,
            placeholder="粘贴或输入多个 Google Scholar 个人页面链接（每行一个，需要下面格式 'https://scholar.google.com/citations?user=NVii64oAAAAJ'）"
        )
        # 新增微信号输入框
        wechat_input = gr.Textbox(
            label="微信号",
            placeholder="请输入您的微信号"
        )
        
        progress_output = gr.Textbox(
            label="进度", 
            value="等待开始...", 
            lines=10,
            interactive=False
        )

        download_output = gr.File(label="下载结果 CSV")

        run_button = gr.Button("开始爬取")

        # 传入两个输入值
        run_button.click(
            fn=process_profiles,
            inputs=[profile_input, wechat_input],
            outputs=[progress_output, download_output]
        )

    return demo

if __name__ == "__main__":
    app = build_app()
    app.launch()