# -*- coding: utf-8 -*-
import streamlit as st
import requests
import json
import pandas as pd
import time
import plotly.express as px
import plotly.graph_objects as go
import matplotlib.pyplot as plt
import matplotlib.font_manager as fm
import matplotlib as mpl
import os
import tempfile

# ────────────────────────────────────────────────
# 頁面設定
# ────────────────────────────────────────────────
st.set_page_config(
    page_title="PChome 商品價格分析",
    page_icon="🛒",
    layout="wide",
)

st.title("🛒 PChome 電商商品價格分析")
st.markdown("輸入關鍵字與筆數，自動爬取 PChome 商品資料並視覺化分析。")

# ────────────────────────────────────────────────
# 側邊欄：使用者輸入
# ────────────────────────────────────────────────
with st.sidebar:
    st.header("⚙️ 搜尋設定")
    keyword = st.text_input("搜尋關鍵字", value="耳機", placeholder="例：耳機、筆電、滑鼠")
    max_items = st.slider("最多筆數", min_value=20, max_value=200, value=60, step=20)
    sort_option = st.selectbox(
        "排序方式",
        options=["sale/dc", "price/ac", "price/dc", "new/dc"],
        format_func=lambda x: {
            "sale/dc": "銷量高→低",
            "price/ac": "價格低→高",
            "price/dc": "價格高→低",
            "new/dc": "最新上架",
        }[x],
    )
    sleep_sec = st.slider("每頁請求間隔（秒）", min_value=1, max_value=10, value=2)
    run_btn = st.button("🚀 開始爬取", use_container_width=True)

# ────────────────────────────────────────────────
# 爬蟲函式
# ────────────────────────────────────────────────
ITEMS_PER_PAGE = 20  # PChome 每頁固定 20 筆

def fetch_pchome(keyword: str, max_items: int, sort: str, sleep_sec: int) -> pd.DataFrame:
    total_pages = -(-max_items // ITEMS_PER_PAGE)  # 無條件進位除法
    all_data = pd.DataFrame()

    progress = st.progress(0, text="爬取中…")
    for i in range(1, total_pages + 1):
        url = (
            f"https://ecshweb.pchome.com.tw/search/v3.3/all/results"
            f"?q={keyword}&page={i}&sort={sort}"
        )
        try:
            resp = requests.get(url, timeout=15)
            resp.raise_for_status()
            data = json.loads(resp.content)
            if "prods" not in data or not data["prods"]:
                break
            df_page = pd.DataFrame(data["prods"])
            all_data = pd.concat([all_data, df_page], ignore_index=True)
        except Exception as e:
            st.warning(f"第 {i} 頁爬取失敗：{e}")
            break

        progress.progress(i / total_pages, text=f"已爬取第 {i}/{total_pages} 頁…")
        if i < total_pages:
            time.sleep(sleep_sec)

    progress.empty()
    # 截斷到使用者要求的筆數
    return all_data.head(max_items)

# ────────────────────────────────────────────────
# 主流程
# ────────────────────────────────────────────────
if run_btn:
    if not keyword.strip():
        st.error("請輸入搜尋關鍵字！")
        st.stop()

    with st.spinner("正在爬取資料，請稍候…"):
        raw_df = fetch_pchome(keyword, max_items, sort_option, sleep_sec)

    if raw_df.empty:
        st.error("未取得任何資料，請確認關鍵字或稍後再試。")
        st.stop()

    # 只保留需要的欄位（有些商品可能缺欄）
    cols_needed = [c for c in ["name", "price", "brand", "storeId", "picS"] if c in raw_df.columns]
    df = raw_df[cols_needed].copy()
    df["price"] = pd.to_numeric(df["price"], errors="coerce")
    df = df.dropna(subset=["price"])
    df = df.reset_index(drop=True)
    df.index += 1  # 從 1 開始

    avg_price = df["price"].mean()
    max_price = df["price"].max()
    min_price = df["price"].min()

    # ── 統計指標 ──────────────────────────────
    st.subheader("📊 價格統計")
    c1, c2, c3, c4 = st.columns(4)
    c1.metric("筆數", f"{len(df)} 筆")
    c2.metric("平均價格", f"NT$ {avg_price:,.0f}")
    c3.metric("最高價格", f"NT$ {max_price:,.0f}")
    c4.metric("最低價格", f"NT$ {min_price:,.0f}")

    # ── 下載 CSV ─────────────────────────────
    import datetime
    today = datetime.date.today().strftime("%Y%m%d")
    csv_bytes = df.to_csv(index=False, encoding="utf-8-sig").encode("utf-8-sig")
    st.download_button(
        label="📥 下載 CSV",
        data=csv_bytes,
        file_name=f"{today}_PCHOME_{keyword}.csv",
        mime="text/csv",
    )

    # ── 資料表 ───────────────────────────────
    with st.expander("🔍 查看原始資料", expanded=False):
        st.dataframe(df, use_container_width=True)

    # ════════════════════════════════════════
    # 圖表區
    # ════════════════════════════════════════
    st.subheader("📈 視覺化分析")
    tab1, tab2, tab3 = st.tabs(["折線圖", "圓餅圖", "旭日圖"])

    # ── Tab1：折線圖 ─────────────────────────
    with tab1:
        fig_line = go.Figure()
        fig_line.add_trace(
            go.Scatter(
                x=df.index,
                y=df["price"],
                mode="lines+markers",
                name="售價",
                line=dict(color="#4C9BE8", width=2),
                marker=dict(size=5),
                hovertext=df["name"] if "name" in df.columns else None,
                hovertemplate="<b>%{hovertext}</b><br>售價：NT$ %{y:,.0f}<extra></extra>",
            )
        )
        fig_line.add_hline(
            y=avg_price,
            line_dash="dash",
            line_color="red",
            annotation_text=f"平均 NT$ {avg_price:,.0f}",
            annotation_position="top left",
        )
        fig_line.update_layout(
            title=f"{today} PChome 電商「{keyword}」售價走勢",
            xaxis_title="商品編號",
            yaxis_title="價格（NT$）",
            hovermode="x unified",
            height=500,
        )
        st.plotly_chart(fig_line, use_container_width=True)

    # ── Tab2：圓餅圖（依價格區間） ────────────
    with tab2:
        bins = [0, 500, 1000, 3000, 5000, 10000, float("inf")]
        labels = ["≤500", "501~1000", "1001~3000", "3001~5000", "5001~10000", ">10000"]
        df["price_range"] = pd.cut(df["price"], bins=bins, labels=labels, right=True)
        pie_data = df["price_range"].value_counts().reset_index()
        pie_data.columns = ["price_range", "count"]
        pie_data = pie_data.sort_values("price_range")

        fig_pie = px.pie(
            pie_data,
            names="price_range",
            values="count",
            title=f"「{keyword}」各價格區間商品占比",
            hole=0.35,
            color_discrete_sequence=px.colors.qualitative.Set3,
        )
        fig_pie.update_traces(textposition="inside", textinfo="percent+label")
        fig_pie.update_layout(height=500)
        st.plotly_chart(fig_pie, use_container_width=True)

    # ── Tab3：旭日圖（品牌 → 價格區間） ──────
    with tab3:
        if "brand" in df.columns and df["brand"].notna().any():
            sun_df = df[["brand", "price_range"]].dropna()
            sun_df["brand"] = sun_df["brand"].fillna("未知品牌").replace("", "未知品牌")
            sun_df["count"] = 1
            sun_df = sun_df.groupby(["brand", "price_range"], as_index=False)["count"].sum()

            fig_sun = px.sunburst(
                sun_df,
                path=["brand", "price_range"],
                values="count",
                title=f"「{keyword}」品牌 × 價格區間旭日圖",
                color="count",
                color_continuous_scale="Blues",
            )
            fig_sun.update_layout(height=600)
            st.plotly_chart(fig_sun, use_container_width=True)
        else:
            # brand 欄不存在時，改用商品名稱前綴 × 價格區間
            df["name_short"] = df["name"].str[:6] + "…" if "name" in df.columns else "商品"
            sun_df = df[["name_short", "price_range"]].dropna()
            sun_df["count"] = 1
            sun_df = sun_df.groupby(["name_short", "price_range"], as_index=False)["count"].sum()
            fig_sun = px.sunburst(
                sun_df,
                path=["name_short", "price_range"],
                values="count",
                title=f"「{keyword}」商品 × 價格區間旭日圖",
                color="count",
                color_continuous_scale="Blues",
            )
            fig_sun.update_layout(height=600)
            st.plotly_chart(fig_sun, use_container_width=True)

else:
    st.info("👈 請在左側設定關鍵字與筆數，然後點擊「開始爬取」。")