app.py

# file_path: app.py
import streamlit as st
import requests
from bs4 import BeautifulSoup
import pandas as pd
from google.oauth2.service_account import Credentials
import gspread
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots

# Google Sheets credentials
SCOPE = ['https://www.googleapis.com/auth/spreadsheets']
SERVICE_ACCOUNT_FILE = "realtime-441511-f5708eabdf26.json"
SPREADSHEET_URL = "https://docs.google.com/spreadsheets/d/1tIsXCbB8P6ZxdnZNnv7S7BBWbbT7lrSjW990zG-vQAA/edit?gid=0#gid=0"

# Streamlit app
st.title("Booking.com 台南飯店資料爬取與分析")
st.sidebar.header("功能選擇")
mode = st.sidebar.selectbox("選擇模式", ["資料爬取", "資料視覺化", "上傳至 Google Sheet"])

@st.cache_data
def scrape_booking_hotel():
    url = "https://www.booking.com/searchresults.zh-tw.html"
    headers = {
        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36',
        'Accept-Language': 'zh-TW,zh;q=0.9,en-US;q=0.8,en;q=0.7',
    }
    params = {
        'ss': '台南',
        'checkin': '2024-11-16',
        'checkout': '2024-11-17',
        'group_adults': '2',
        'no_rooms': '1',
        'group_children': '0',
        'dest_id': '-2637868',
        'dest_type': 'city'
    }
    try:
        response = requests.get(url, headers=headers, params=params)
        response.raise_for_status()
        soup = BeautifulSoup(response.text, 'html.parser')
        hotels_data = []
        hotel_cards = soup.find_all('div', {'data-testid': 'property-card'})

        for hotel in hotel_cards:
            try:
                name_elem = hotel.find('div', {'data-testid': 'title', 'class': 'f6431b446c'})
                name = name_elem.text.strip() if name_elem else "無資料"
                price_elem = hotel.find('span', {
                    'data-testid': 'price-and-discounted-price',
                    'class': 'f6431b446c'
                })
                price = price_elem.text.strip() if price_elem else "無資料"
                price = price.replace('TWD', '').replace(' ', '').replace(',', '').strip()
                rating_container = hotel.find('div', {'class': 'a3b8729ab1'})
                rating_elem = rating_container.find('div', {'class': 'ac4a7896c7'}) if rating_container else None
                rating = rating_elem.text.strip() if rating_elem else "無評分"
                description_elem = hotel.find('div', {'data-testid': 'recommended-units'})
                if description_elem:
                    room_type = description_elem.find('h4', {'class': 'abf093bdfe'})
                    room_type = room_type.text.strip() if room_type else ""
                    bed_info = description_elem.find('div', {'class': 'abf093bdfe'})
                    bed_info = bed_info.text.strip() if bed_info else ""
                    cancellation = description_elem.find('strong', text='可免費取消')
                    cancellation = "可免費取消" if cancellation else ""
                    payment = description_elem.find('strong', text='無需訂金')
                    payment = "無需訂金" if payment else ""
                    description = f"{room_type} | {bed_info} | {cancellation} | {payment}".strip(' |')
                else:
                    description = "無說明"
                hotels_data.append({
                    '飯店名稱': name,
                    '價格': price,
                    '評分': rating,
                    '說明': description
                })
            except AttributeError:
                continue

        df = pd.DataFrame(hotels_data).drop_duplicates()
        return df
    except requests.RequestException:
        return pd.DataFrame()

def clean_rating(x):
    if pd.isna(x) or x == '無評分':
        return 0
    return float(str(x).replace('分數', '').replace('分', ''))

def create_price_rating_scatter(df):
    fig = px.scatter(
        df,
        x='價格',
        y='評分',
        text='飯店名稱',
        size='價格',
        color='評分',
        title='台南飯店價格與評分關係圖',
        labels={'價格': '房價 (TWD)', '評分': '評分 (0-10)'}
    )
    fig.update_layout(height=600, title_x=0.5)
    return fig

def create_price_distribution(df):
    fig = go.Figure()
    fig.add_trace(go.Histogram(
        x=df['價格'],
        name='價格分布',
        nbinsx=10,
        marker_color='rgb(55, 83, 109)'
    ))
    fig.add_trace(go.Box(
        x=df['價格'],
        name='價格箱型圖',
        marker_color='rgb(26, 118, 255)'
    ))
    fig.update_layout(title_text='台南飯店價格分布', title_x=0.5, height=500)
    return fig

def upload_to_google_sheets(df):
    creds = Credentials.from_service_account_file(SERVICE_ACCOUNT_FILE, scopes=SCOPE)
    gs = gspread.authorize(creds)
    sheet = gs.open_by_url(SPREADSHEET_URL)
    worksheet = sheet.get_worksheet(0)
    df1 = df.astype(str)
    worksheet.update([df1.columns.values.tolist()] + df1.values.tolist())
    return "資料已成功上傳到 Google Sheet！"

if mode == "資料爬取":
    st.header("爬取台南飯店資料")
    if st.button("開始爬取"):
        df = scrape_booking_hotel()
        if not df.empty:
            st.dataframe(df)
            df.to_csv('booking_hotels_tainan.csv', index=False, encoding='utf-8-sig')
            st.success("資料爬取成功，已儲存至 booking_hotels_tainan.csv")
        else:
            st.error("未能成功爬取資料")

elif mode == "資料視覺化":
    st.header("分析與視覺化")
    try:
        df = pd.read_csv('booking_hotels_tainan.csv', encoding='utf-8-sig')
        df['價格'] = pd.to_numeric(df['價格'], errors='coerce')
        df['評分'] = df['評分'].apply(clean_rating)
        st.plotly_chart(create_price_rating_scatter(df))
        st.plotly_chart(create_price_distribution(df))
    except Exception as e:
        st.error(f"讀取或分析資料時發生錯誤：{e}")

elif mode == "上傳至 Google Sheet":
    st.header("上傳資料至 Google Sheet")
    try:
        df = pd.read_csv('booking_hotels_tainan.csv', encoding='utf-8-sig')
        result = upload_to_google_sheets(df)
        st.success(result)
    except Exception as e:
        st.error(f"上傳資料時發生錯誤：{e}")