app.py

import requests
from bs4 import BeautifulSoup
import pandas as pd
import folium
from folium.plugins import MarkerCluster, HeatMap
import plotly.graph_objects as go
import plotly.express as px
from geopy.geocoders import Nominatim
import re
import streamlit as st

# Streamlit title and description
st.title("米其林餐廳指南爬蟲")
st.write("Extract restaurant data, visualize with charts, and display locations on maps.")

# Read data from Google Sheets
sheet_id = "1xUfnD1WCF5ldqECI8YXIko1gCpaDDCwTztL17kjI42U"
df1 = pd.read_csv(f"https://docs.google.com/spreadsheets/d/{sheet_id}/export?format=csv")

# Convert "網址" column to a Python list
urls = df1["網址"].tolist()

# Create a DataFrame to store all restaurant data
df = pd.DataFrame(columns=["Store Name", "Address", "Phone", "Latitude", "Longitude", "Region"])

# Initialize Nominatim geocoder
geolocator = Nominatim(user_agent="my_app")

# Function to extract region (區域) from the address using regex
def extract_region(address):
    match = re.search(r'(.*?)區|縣|市', address)
    if match:
        return match.group(0)
    else:
        return "Unknown"

# Function to fetch and parse data
def fetch_data():
    global df
    # Progress bar in Streamlit
    progress_bar = st.progress(0)
    total_urls = len(urls)

    # Iterate through each URL
    for idx, url in enumerate(urls):
        response = requests.get(url)
        soup = BeautifulSoup(response.content, "html.parser")

        try:
            store_name = soup.find("h2", class_="restaurant-details__heading--title").text.strip()
        except AttributeError:
            store_name = None

        try:
            address = soup.find("li", class_="restaurant-details__heading--address").text.strip()
            region = extract_region(address)
        except AttributeError:
            address = None
            region = "Unknown"
        
        try:
            phone = soup.find("a", {"data-event": "CTA_tel"}).get("href").replace("tel:", "")
        except AttributeError:
            phone = None

        try:
            location = geolocator.geocode(address)
            if location:
                latitude = location.latitude
                longitude = location.longitude
            else:
                latitude = None
                longitude = None
        except:
            latitude = None
            longitude = None

        new_row = pd.DataFrame({
            "Store Name": [store_name],
            "Address": [address],
            "Phone": [phone],
            "Latitude": [latitude],
            "Longitude": [longitude],
            "Region": [region]
        })

        df = pd.concat([df, new_row], ignore_index=True)

        # Update progress bar
        progress_bar.progress((idx + 1) / total_urls)

# Button to trigger data fetching
if st.button("爬取餐廳資料"):
    fetch_data()
    
    # Save the DataFrame to CSV with UTF-8 encoding, including latitude and longitude
    csv_file = "restaurants_data.csv"
    df.to_csv(csv_file, encoding="utf-8-sig", index=False)

    # Display the DataFrame as a table at the top
    st.subheader("Restaurant Data")
    st.dataframe(df)

    # Display download button for the CSV
    st.download_button(
        label="Download restaurant data as CSV",
        data=open(csv_file, "rb").read(),
        file_name=csv_file,
        mime="text/csv"
    )

    # Group the data by region and sum the number of restaurants
    region_group = df.groupby("Region").size().reset_index(name='Count')

    # Plot enlarged pie chart with custom colors and labels
    pie_chart = go.Figure(go.Pie(
        labels=region_group["Region"],
        values=region_group["Count"],
        textinfo="label+percent",
        hoverinfo="label+value",
        textfont=dict(size=18),
        marker=dict(colors=px.colors.qualitative.Set3, line=dict(color='#000000', width=2))
    ))

    pie_chart.update_layout(
        title="Restaurant Distribution by Region",
        title_x=0.5,
        title_font=dict(size=24, family="Arial"),
        height=600,
        margin=dict(t=50, b=50, l=50, r=50)
    )
    st.subheader("Restaurant Distribution by Region (Enlarged Pie Chart)")
    st.plotly_chart(pie_chart)

    # Plot bar chart with custom colors and labels
    bar_chart = go.Figure(go.Bar(
        x=region_group["Region"],
        y=region_group["Count"],
        text=region_group["Count"],
        textposition='auto',
        marker=dict(color=px.colors.qualitative.Set2)
    ))

    bar_chart.update_layout(
        title="Restaurant Count by Region",
        title_x=0.5,
        title_font=dict(size=24, family="Arial"),
        height=400,
        margin=dict(t=50, b=50, l=50, r=50),
        xaxis_title="Region",
        yaxis_title="Number of Restaurants",
        xaxis=dict(tickangle=-45)
    )
    st.subheader("Restaurant Count by Region (Bar Chart)")
    st.plotly_chart(bar_chart)

    # Display a map using Folium
    st.subheader("Restaurant Locations Map")

    # Create map centered around the mean latitude and longitude
    m = folium.Map(location=[df['Latitude'].mean(), df['Longitude'].mean()], zoom_start=10)

    # Add marker cluster to the map
    marker_cluster = MarkerCluster().add_to(m)
    for index, row in df.iterrows():
        if pd.notnull(row["Latitude"]) and pd.notnull(row["Longitude"]):
            folium.Marker(
                location=[row["Latitude"], row["Longitude"]],
                popup=f"{row['Store Name']} ({row['Phone']})",
                tooltip=row["Address"]
            ).add_to(marker_cluster)

    # Display the map in Streamlit
    st.components.v1.html(m._repr_html_(), height=600)

    # New section for heatmap
    st.header("餐廳分布熱力圖")

    # Prepare data for heatmap
    heat_data = [[row['Latitude'], row['Longitude']] for index, row in df.iterrows() if pd.notnull(row['Latitude']) and pd.notnull(row['Longitude'])]

    # Create a new map for the heatmap
    heatmap = folium.Map(location=[df['Latitude'].mean(), df['Longitude'].mean()], zoom_start=10)

    # Add heatmap to the map
    HeatMap(heat_data).add_to(heatmap)

    # Display the heatmap in Streamlit
    st.components.v1.html(heatmap._repr_html_(), height=600)

    # Regional restaurant count analysis
    st.header("各區域餐廳數量分析")

    # Create bar chart for restaurant count by region using Plotly Express
    fig_bar = px.bar(region_group, x='Region', y='Count', 
                     title="各區域餐廳數量比較", 
                     color='Count',
                     color_continuous_scale=px.colors.sequential.Viridis)
    st.plotly_chart(fig_bar)

    # Create a scatter mapbox for individual restaurant locations
    fig_scatter = px.scatter_mapbox(df, lat="Latitude", lon="Longitude", 
                                    hover_name="Store Name", 
                                    hover_data=["Address", "Phone"],
                                    zoom=10, height=600,
                                    title="餐廳位置分布圖")
    fig_scatter.update_layout(mapbox_style="open-street-map")
    st.plotly_chart(fig_scatter)
    #