Update app.py

app.py

@@ -6,67 +6,611 @@ import plotly.graph_objects as go
 import io
 import base64
 from PIL import Image
+import matplotlib.pyplot as plt
+import seaborn as sns
+from plotly.subplots import make_subplots
 
-# 初始化默認數據
-default_data = pd.DataFrame({
-    "類別": ["A", "B", "C", "D", "E"],
-    "數值": [10, 20, 15, 25, 30]
-})
+# 擴展的圖表類型
+CHART_TYPES = [
+    "長條圖", "堆疊長條圖", "群組長條圖", "水平長條圖", 
+    "折線圖", "多重折線圖", "階梯折線圖", 
+    "圓餅圖", "環形圖", "散點圖", "氣泡圖", 
+    "區域圖", "堆疊區域圖", "雷達圖", "熱力圖", 
+    "箱型圖", "小提琴圖", "漏斗圖", "樹狀圖", 
+    "直方圖", "極座標圖", "甘特圖"
+]
 
-# 可用的顏色方案
+# 可用的顏色方案 (美觀且有實用價值的顏色選擇)
 COLOR_SCHEMES = {
     "默認": px.colors.qualitative.Plotly,
+    "Pastel": px.colors.qualitative.Pastel,
+    "Safe": px.colors.qualitative.Safe, 
+    "Vivid": px.colors.qualitative.Vivid,
+    "Prism": px.colors.qualitative.Prism,
+    "Antique": px.colors.qualitative.Antique,
+    "Bold": px.colors.qualitative.Bold,
+    "Pastel1": px.colors.qualitative.Pastel1,
+    "Pastel2": px.colors.qualitative.Pastel2,
+    "Set1": px.colors.qualitative.Set1,
+    "Set2": px.colors.qualitative.Set2,
+    "Set3": px.colors.qualitative.Set3,
     "藍綠色系": px.colors.sequential.Blues,
     "紅色系": px.colors.sequential.Reds,
     "綠色系": px.colors.sequential.Greens,
-    "彩虹色": px.colors.sequential.Turbo
+    "紫色系": px.colors.sequential.Purples,
+    "灰度": px.colors.sequential.Greys,
 }
 
-def create_plot(df, chart_type, x_column, y_column, color_scheme, title, width, height, show_grid, show_legend):
+# 圖案填充類型 (黑白印刷用)
+PATTERN_TYPES = [
+    "無", "/", "\\", "x", "-", "|", "+", ".", "*"
+]
+
+# 統計函數選項
+AGGREGATION_FUNCTIONS = [
+    "求和", "平均值", "最大值", "最小值", "計數", "中位數", "標準差", "變異數"
+]
+
+def agg_function_map(func_name):
+    """映射中文統計函數名稱到Pandas函數"""
+    mapping = {
+        "求和": "sum",
+        "平均值": "mean",
+        "最大值": "max",
+        "最小值": "min",
+        "計數": "count",
+        "中位數": "median",
+        "標準差": "std",
+        "變異數": "var"
+    }
+    return mapping.get(func_name, "sum")
+
+def create_plot(df, chart_type, x_column, y_column, group_column=None, size_column=None, 
+               color_scheme="默認", patterns=[], title="", width=700, height=500, 
+               show_grid=True, show_legend=True, agg_function="求和", custom_colors={}):
     """創建圖表函數"""
     
     # 數據預處理
     if df is None or df.empty:
-        df = default_data
+        return go.Figure()
     
     # 確保列存在
-    if x_column not in df.columns:
-        x_column = df.columns[0] if len(df.columns) > 0 else "類別"
-    
-    if y_column not in df.columns:
-        y_column = df.columns[1] if len(df.columns) > 1 else "數值"
+    if x_column not in df.columns or y_column not in df.columns:
+        return go.Figure()
     
     # 獲取選擇的顏色方案
     colors = COLOR_SCHEMES[color_scheme]
     
+    # 將非數值列轉換為類別型
+    for col in df.columns:
+        if df[col].dtype == 'object' or df[col].dtype == 'string':
+            df[col] = df[col].astype('category')
+    
+    # 使用pandas的groupby進行數據聚合
+    agg_func = agg_function_map(agg_function)
+    
     # 設置圖表參數
     fig_params = {
-        "width": width,
-        "height": height,
         "title": title
     }
     
-    # 基於選擇的圖表類型創建圖表
-    if chart_type == "長條圖":
-        fig = px.bar(df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
-    
-    elif chart_type == "折線圖":
-        fig = px.line(df, x=x_column, y=y_column, markers=True, color_discrete_sequence=colors, **fig_params)
-    
-    elif chart_type == "圓餅圖":
-        fig = px.pie(df, names=x_column, values=y_column, color_discrete_sequence=colors, **fig_params)
+    # 創建基本圖形
+    fig = go.Figure()
     
-    elif chart_type == "散點圖":
-        fig = px.scatter(df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
-    
-    elif chart_type == "區域圖":
-        fig = px.area(df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
+    try:
+        # 基於選擇的圖表類型創建圖表
+        if chart_type == "長條圖":
+            # 進行數據分組和聚合
+            if group_column and group_column in df.columns and group_column != x_column:
+                grouped_df = df.groupby([x_column, group_column])[y_column].agg(agg_func).reset_index()
+                fig = px.bar(grouped_df, x=x_column, y=y_column, color=group_column, 
+                             color_discrete_sequence=colors, **fig_params)
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.bar(grouped_df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
+                
+            # 應用自定義顏色和圖案
+            if patterns and len(patterns) > 0:
+                for i, bar in enumerate(fig.data):
+                    pattern_index = i % len(patterns)
+                    if patterns[pattern_index] != "無":
+                        bar.marker.pattern = {
+                            'shape': patterns[pattern_index],
+                            'solidity': 0.5
+                        }
+        
+        elif chart_type == "堆疊長條圖":
+            if group_column and group_column in df.columns:
+                grouped_df = df.pivot_table(index=x_column, columns=group_column, 
+                                           values=y_column, aggfunc=agg_func).reset_index()
+                grouped_df = grouped_df.fillna(0)
+                
+                # 取得所有類別
+                categories = grouped_df.columns.tolist()
+                categories.remove(x_column)
+                
+                for i, category in enumerate(categories):
+                    color = colors[i % len(colors)]
+                    if category in custom_colors:
+                        color = custom_colors[category]
+                        
+                    pattern_shape = None
+                    if patterns and i < len(patterns) and patterns[i] != "無":
+                        pattern_shape = patterns[i]
+                        
+                    fig.add_trace(go.Bar(
+                        x=grouped_df[x_column],
+                        y=grouped_df[category],
+                        name=str(category),
+                        marker_color=color,
+                        marker_pattern_shape=pattern_shape
+                    ))
+                
+                fig.update_layout(barmode='stack')
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.bar(grouped_df, x=x_column, y=y_column, **fig_params)
+        
+        elif chart_type == "群組長條圖":
+            if group_column and group_column in df.columns:
+                grouped_df = df.pivot_table(index=x_column, columns=group_column, 
+                                           values=y_column, aggfunc=agg_func).reset_index()
+                grouped_df = grouped_df.fillna(0)
+                
+                # 取得所有類別
+                categories = grouped_df.columns.tolist()
+                categories.remove(x_column)
+                
+                for i, category in enumerate(categories):
+                    color = colors[i % len(colors)]
+                    if category in custom_colors:
+                        color = custom_colors[category]
+                        
+                    pattern_shape = None
+                    if patterns and i < len(patterns) and patterns[i] != "無":
+                        pattern_shape = patterns[i]
+                        
+                    fig.add_trace(go.Bar(
+                        x=grouped_df[x_column],
+                        y=grouped_df[category],
+                        name=str(category),
+                        marker_color=color,
+                        marker_pattern_shape=pattern_shape
+                    ))
+                
+                fig.update_layout(barmode='group')
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.bar(grouped_df, x=x_column, y=y_column, **fig_params)
+                
+        elif chart_type == "水平長條圖":
+            if group_column and group_column in df.columns:
+                grouped_df = df.groupby([x_column, group_column])[y_column].agg(agg_func).reset_index()
+                fig = px.bar(grouped_df, y=x_column, x=y_column, color=group_column, 
+                           color_discrete_sequence=colors, orientation='h', **fig_params)
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.bar(grouped_df, y=x_column, x=y_column, orientation='h', 
+                           color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "折線圖":
+            if group_column and group_column in df.columns and group_column != x_column:
+                grouped_df = df.groupby([x_column, group_column])[y_column].agg(agg_func).reset_index()
+                fig = px.line(grouped_df, x=x_column, y=y_column, color=group_column, 
+                            color_discrete_sequence=colors, markers=True, **fig_params)
+                
+                # 根據自定義顏色和線型
+                for i, trace in enumerate(fig.data):
+                    if i < len(patterns) and patterns[i] != "無":
+                        if patterns[i] == '/':
+                            trace.line.dash = 'dash'
+                        elif patterns[i] == '\\':
+                            trace.line.dash = 'dot'
+                        elif patterns[i] == 'x':
+                            trace.line.dash = 'dashdot'
+                        elif patterns[i] == '-':
+                            trace.line.dash = 'longdash'
+                        else:
+                            trace.line.dash = 'solid'
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.line(grouped_df, x=x_column, y=y_column, markers=True, 
+                            color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "多重折線圖":
+            if group_column and group_column in df.columns:
+                grouped_df = df.pivot_table(index=x_column, columns=group_column, 
+                                           values=y_column, aggfunc=agg_func).reset_index()
+                grouped_df = grouped_df.fillna(0)
+                
+                # 取得所有類別
+                categories = grouped_df.columns.tolist()
+                categories.remove(x_column)
+                
+                for i, category in enumerate(categories):
+                    color = colors[i % len(colors)]
+                    if category in custom_colors:
+                        color = custom_colors[category]
+                    
+                    line_dash = 'solid'
+                    if patterns and i < len(patterns) and patterns[i] != "無":
+                        if patterns[i] == '/':
+                            line_dash = 'dash'
+                        elif patterns[i] == '\\':
+                            line_dash = 'dot'
+                        elif patterns[i] == 'x':
+                            line_dash = 'dashdot'
+                        elif patterns[i] == '-':
+                            line_dash = 'longdash'
+                    
+                    fig.add_trace(go.Scatter(
+                        x=grouped_df[x_column],
+                        y=grouped_df[category],
+                        mode='lines+markers',
+                        name=str(category),
+                        line=dict(color=color, dash=line_dash),
+                        marker=dict(color=color)
+                    ))
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.line(grouped_df, x=x_column, y=y_column, markers=True, **fig_params)
+        
+        elif chart_type == "階梯折線圖":
+            if group_column and group_column in df.columns:
+                grouped_df = df.pivot_table(index=x_column, columns=group_column, 
+                                           values=y_column, aggfunc=agg_func).reset_index()
+                grouped_df = grouped_df.fillna(0)
+                
+                # 取得所有類別
+                categories = grouped_df.columns.tolist()
+                categories.remove(x_column)
+                
+                for i, category in enumerate(categories):
+                    color = colors[i % len(colors)]
+                    if category in custom_colors:
+                        color = custom_colors[category]
+                    
+                    line_dash = 'solid'
+                    if patterns and i < len(patterns) and patterns[i] != "無":
+                        if patterns[i] == '/':
+                            line_dash = 'dash'
+                        elif patterns[i] == '\\':
+                            line_dash = 'dot'
+                        elif patterns[i] == 'x':
+                            line_dash = 'dashdot'
+                        elif patterns[i] == '-':
+                            line_dash = 'longdash'
+                    
+                    fig.add_trace(go.Scatter(
+                        x=grouped_df[x_column],
+                        y=grouped_df[category],
+                        mode='lines',
+                        name=str(category),
+                        line=dict(shape='hv', color=color, dash=line_dash)
+                    ))
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig.add_trace(go.Scatter(
+                    x=grouped_df[x_column],
+                    y=grouped_df[y_column],
+                    mode='lines',
+                    line=dict(shape='hv', color=colors[0])
+                ))
+        
+        elif chart_type == "圓餅圖":
+            # 圓餅圖只需要分類和對應的數值
+            grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+            
+            # 設置自定義顏色
+            pie_colors = colors
+            if custom_colors and len(custom_colors) > 0:
+                pie_colors = [custom_colors.get(cat, colors[i % len(colors)]) 
+                             for i, cat in enumerate(grouped_df[x_column])]
+            
+            # 設置自定義圖案
+            pattern_shapes = None
+            if patterns and len(patterns) > 0:
+                pattern_shapes = [p if p != "無" else None for p in patterns]
+                if len(pattern_shapes) < len(grouped_df):
+                    # 重複圖案以匹配數據長度
+                    pattern_shapes = pattern_shapes * (len(grouped_df) // len(pattern_shapes) + 1)
+                    pattern_shapes = pattern_shapes[:len(grouped_df)]
+            
+            fig = px.pie(grouped_df, names=x_column, values=y_column, 
+                       color_discrete_sequence=pie_colors, **fig_params)
+            
+            # 應用圖案填充
+            if pattern_shapes:
+                for i, trace in enumerate(fig.data):
+                    for j, path in enumerate(trace.marker.pattern):
+                        if j < len(pattern_shapes) and pattern_shapes[j]:
+                            path.shape = pattern_shapes[j]
+                            path.solidity = 0.5
+        
+        elif chart_type == "環形圖":
+            grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+            
+            # 設置自定義顏色
+            pie_colors = colors
+            if custom_colors and len(custom_colors) > 0:
+                pie_colors = [custom_colors.get(cat, colors[i % len(colors)]) 
+                             for i, cat in enumerate(grouped_df[x_column])]
+            
+            fig = px.pie(grouped_df, names=x_column, values=y_column, hole=0.4,
+                       color_discrete_sequence=pie_colors, **fig_params)
+            
+            # 應用圖案填充
+            if patterns and len(patterns) > 0:
+                for i, trace in enumerate(fig.data):
+                    trace.marker.pattern = {
+                        'shape': patterns[i % len(patterns)] if patterns[i % len(patterns)] != "無" else None,
+                        'solidity': 0.5
+                    }
+        
+        elif chart_type == "散點圖":
+            if group_column and group_column in df.columns:
+                fig = px.scatter(df, x=x_column, y=y_column, color=group_column,
+                              color_discrete_sequence=colors, **fig_params)
+                
+                if size_column and size_column in df.columns:
+                    fig = px.scatter(df, x=x_column, y=y_column, color=group_column,
+                                  size=size_column, color_discrete_sequence=colors, **fig_params)
+            else:
+                fig = px.scatter(df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
+            
+            # 應用散點圖符號
+            if patterns and len(patterns) > 0:
+                for i, trace in enumerate(fig.data):
+                    pattern_idx = i % len(patterns)
+                    if patterns[pattern_idx] != "無":
+                        if patterns[pattern_idx] == '/':
+                            trace.marker.symbol = 'diamond'
+                        elif patterns[pattern_idx] == '\\':
+                            trace.marker.symbol = 'square'
+                        elif patterns[pattern_idx] == 'x':
+                            trace.marker.symbol = 'x'
+                        elif patterns[pattern_idx] == '-':
+                            trace.marker.symbol = 'line-ew'
+                        elif patterns[pattern_idx] == '|':
+                            trace.marker.symbol = 'line-ns'
+                        elif patterns[pattern_idx] == '+':
+                            trace.marker.symbol = 'cross'
+                        elif patterns[pattern_idx] == '.':
+                            trace.marker.symbol = 'circle'
+                        elif patterns[pattern_idx] == '*':
+                            trace.marker.symbol = 'star'
+                        else:
+                            trace.marker.symbol = 'circle'
+        
+        elif chart_type == "氣泡圖":
+            if size_column and size_column in df.columns:
+                if group_column and group_column in df.columns:
+                    fig = px.scatter(df, x=x_column, y=y_column, color=group_column,
+                                  size=size_column, size_max=30, 
+                                  color_discrete_sequence=colors, **fig_params)
+                else:
+                    fig = px.scatter(df, x=x_column, y=y_column, size=size_column,
+                                  size_max=30, color_discrete_sequence=colors, **fig_params)
+            else:
+                # 如果沒有指定大小列，則退回到一般散點圖
+                fig = px.scatter(df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "區域圖":
+            grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+            fig = px.area(grouped_df, x=x_column, y=y_column, 
+                        color_discrete_sequence=colors, **fig_params)
+            
+            # 應用填充圖案
+            if patterns and len(patterns) > 0 and patterns[0] != "無":
+                for trace in fig.data:
+                    trace.fill = 'tozeroy'
+                    # Plotly的區域圖不直接支持填充圖案，但可以使用線條樣式來模擬
+                    if patterns[0] == '/':
+                        trace.line.dash = 'dash'
+                    elif patterns[0] == '\\':
+                        trace.line.dash = 'dot'
+                    elif patterns[0] == 'x':
+                        trace.line.dash = 'dashdot'
+        
+        elif chart_type == "堆疊區域圖":
+            if group_column and group_column in df.columns:
+                # 創建樞紐表以便於繪製堆疊面積圖
+                pivot_df = df.pivot_table(index=x_column, columns=group_column, 
+                                         values=y_column, aggfunc=agg_func).reset_index()
+                pivot_df = pivot_df.fillna(0)
+                
+                # 獲取類別列
+                categories = pivot_df.columns.tolist()
+                categories.remove(x_column)
+                
+                # 建立堆疊區域圖
+                for i, category in enumerate(categories):
+                    color = colors[i % len(colors)]
+                    if category in custom_colors:
+                        color = custom_colors[category]
+                    
+                    # 添加區域軌跡
+                    fig.add_trace(go.Scatter(
+                        x=pivot_df[x_column],
+                        y=pivot_df[category],
+                        mode='lines',
+                        line=dict(width=0.5, color=color),
+                        stackgroup='one',
+                        name=str(category)
+                    ))
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.area(grouped_df, x=x_column, y=y_column, **fig_params)
+        
+        elif chart_type == "雷達圖":
+            if group_column and group_column in df.columns:
+                # 對於每個組創建一個雷達圖的軌跡
+                groups = df[group_column].unique()
+                
+                for i, group in enumerate(groups):
+                    group_data = df[df[group_column] == group]
+                    grouped_df = group_data.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                    
+                    theta = grouped_df[x_column].tolist()
+                    r = grouped_df[y_column].tolist()
+                    
+                    # 封閉雷達圖
+                    theta.append(theta[0])
+                    r.append(r[0])
+                    
+                    fig.add_trace(go.Scatterpolar(
+                        r=r,
+                        theta=theta,
+                        fill='toself',
+                        name=str(group),
+                        line_color=colors[i % len(colors)]
+                    ))
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                
+                theta = grouped_df[x_column].tolist()
+                r = grouped_df[y_column].tolist()
+                
+                # 封閉雷達圖
+                theta.append(theta[0])
+                r.append(r[0])
+                
+                fig.add_trace(go.Scatterpolar(
+                    r=r,
+                    theta=theta,
+                    fill='toself',
+                    line_color=colors[0]
+                ))
+            
+            fig.update_layout(
+                polar=dict(
+                    radialaxis=dict(
+                        visible=True
+                    )
+                )
+            )
+        
+        elif chart_type == "熱力圖":
+            # 熱力圖需要兩個分類變量和一個連續變量
+            if group_column and group_column in df.columns:
+                # 創建樞紐表
+                pivot_df = df.pivot_table(index=x_column, columns=group_column, 
+                                         values=y_column, aggfunc=agg_func)
+                
+                # 創建熱力圖
+                fig = px.imshow(pivot_df, color_continuous_scale=px.colors.sequential.Viridis, **fig_params)
+                fig.update_layout(coloraxis_showscale=True)
+            else:
+                # 如果沒有組列，則沒有足夠的維度來創建熱力圖
+                fig = go.Figure()
+                fig.add_annotation(
+                    text="熱力圖需要選擇一個分組列",
+                    showarrow=False,
+                    font=dict(size=16)
+                )
+        
+        elif chart_type == "箱型圖":
+            if group_column and group_column in df.columns:
+                fig = px.box(df, x=group_column, y=y_column, color=group_column,
+                          color_discrete_sequence=colors, **fig_params)
+            else:
+                fig = px.box(df, y=y_column, color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "小提琴圖":
+            if group_column and group_column in df.columns:
+                fig = px.violin(df, x=group_column, y=y_column, color=group_column,
+                             box=True, points="all", color_discrete_sequence=colors, **fig_params)
+            else:
+                fig = px.violin(df, y=y_column, box=True, points="all",
+                             color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "漏斗圖":
+            grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+            
+            # 按值排序
+            grouped_df = grouped_df.sort_values(by=y_column, ascending=False)
+            
+            # 創建漏斗圖
+            fig = go.Figure(go.Funnel(
+                y=grouped_df[x_column],
+                x=grouped_df[y_column],
+                textposition="inside",
+                textinfo="value+percent initial",
+                marker={"color": colors[:len(grouped_df)]}
+            ))
+            
+            fig.update_layout(title=title)
+        
+        elif chart_type == "樹狀圖":
+            if group_column and group_column in df.columns:
+                # 創建層次結構
+                fig = px.treemap(df, path=[group_column, x_column], values=y_column,
+                              color_discrete_sequence=colors, **fig_params)
+            else:
+                grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+                fig = px.treemap(grouped_df, path=[x_column], values=y_column,
+                              color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "直方圖":
+            # 直方圖顯示單一變量的分佈
+            fig = px.histogram(df, x=x_column, color=group_column if group_column else None,
+                            color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "極座標圖":
+            grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+            
+            # 創建極座標條形圖
+            fig = px.bar_polar(grouped_df, r=y_column, theta=x_column,
+                            color=x_column, color_discrete_sequence=colors, **fig_params)
+        
+        elif chart_type == "甘特圖":
+            # 甘特圖需要開始和結束時間
+            # 假設x_column是任務名稱，y_column是開始時間，group_column是結束時間
+            if group_column and group_column in df.columns:
+                # 確保日期時間格式
+                try:
+                    df[y_column] = pd.to_datetime(df[y_column])
+                    df[group_column] = pd.to_datetime(df[group_column])
+                    
+                    # 創建甘特圖
+                    fig = px.timeline(df, x_start=y_column, x_end=group_column, y=x_column,
+                                   color=size_column if size_column else None,
+                                   color_discrete_sequence=colors, **fig_params)
+                    
+                    fig.update_layout(xaxis_type="date")
+                except:
+                    fig = go.Figure()
+                    fig.add_annotation(
+                        text="無法將列轉換為日期格式，甘特圖需要日期時間格式的開始和結束列",
+                        showarrow=False,
+                        font=dict(size=14)
+                    )
+            else:
+                fig = go.Figure()
+                fig.add_annotation(
+                    text="甘特圖需要開始日期和結束日期列",
+                    showarrow=False,
+                    font=dict(size=16)
+                )
+        
+        else:
+            # 默認圖表類型
+            grouped_df = df.groupby(x_column)[y_column].agg(agg_func).reset_index()
+            fig = px.bar(grouped_df, x=x_column, y=y_column, **fig_params)
     
-    else:  # 默認使用長條圖
-        fig = px.bar(df, x=x_column, y=y_column, color_discrete_sequence=colors, **fig_params)
+    except Exception as e:
+        # 處理創建圖表時的錯誤
+        fig = go.Figure()
+        fig.add_annotation(
+            text=f"創建圖表時出錯: {str(e)}",
+            showarrow=False,
+            font=dict(size=14)
+        )
     
-    # 設置網格
+    # 設置網格和圖例
     fig.update_layout(
+        width=width,
+        height=height,
         showlegend=show_legend,
         xaxis=dict(showgrid=show_grid),
         yaxis=dict(showgrid=show_grid)
@@ -106,8 +650,8 @@ def parse_data(csv_data):
         if ',' in first_line:
             # 優先使用逗號作為分隔符
             df = pd.read_csv(io.StringIO(csv_data), sep=',')
-        elif ' ' in first_line:
-            # 如果沒有逗號但有空格，使用空格作為分隔符
+        elif ' ' in first_line or '\t' in first_line:
+            # 如果沒有逗號但有空格或制表符，使用空格作為分隔符
             df = pd.read_csv(io.StringIO(csv_data), sep='\\s+')
         else:
             # 默認使用逗號
@@ -156,38 +700,61 @@ def update_columns(df):
     """更新列選擇下拉菜單"""
     if df is None or df.empty:
         # 默認列
-        return gr.Dropdown(choices=["類別", "數值"], value="類別"), gr.Dropdown(choices=["類別", "數值"], value="數值")
+        return gr.Dropdown(choices=["類別", "數值"], value="類別"), gr.Dropdown(choices=["類別", "數值"], value="數值"), gr.Dropdown(choices=["無", "類別", "數值"], value="無"), gr.Dropdown(choices=["無", "類別", "數值"], value="無")
     
     columns = df.columns.tolist()
     x_dropdown = gr.Dropdown(choices=columns, value=columns[0] if columns else None)
     y_dropdown = gr.Dropdown(choices=columns, value=columns[1] if len(columns) > 1 else columns[0])
+    group_dropdown = gr.Dropdown(choices=["無"] + columns, value="無")
+    size_dropdown = gr.Dropdown(choices=["無"] + columns, value="無")
     
-    return x_dropdown, y_dropdown
+    return x_dropdown, y_dropdown, group_dropdown, size_dropdown
 
-def download_figure(fig):
+def download_figure(fig, format_type="PNG"):
     """導出圖表為圖像"""
     if fig is None:
         return None, "沒有圖表可以導出"
     
     try:
-        # 將Plotly圖表轉換為PNG圖像
-        img_bytes = fig.to_image(format="png")
+        # 選擇導出格式
+        if format_type == "PNG":
+            img_bytes = fig.to_image(format="png")
+            mime_type = "image/png"
+            ext = "png"
+        elif format_type == "SVG":
+            img_bytes = fig.to_image(format="svg")
+            mime_type = "image/svg+xml"
+            ext = "svg"
+        elif format_type == "PDF":
+            img_bytes = fig.to_image(format="pdf")
+            mime_type = "application/pdf"
+            ext = "pdf"
+        elif format_type == "JPEG":
+            img_bytes = fig.to_image(format="jpeg")
+            mime_type = "image/jpeg"
+            ext = "jpg"
+        else:
+            img_bytes = fig.to_image(format="png")
+            mime_type = "image/png"
+            ext = "png"
         
-        # 創建PIL圖像對象
-        img = Image.open(io.BytesIO(img_bytes))
+        # 創建文件對象
+        filename = f"chart_export.{ext}"
         
-        return img, "圖表已成功導出為PNG圖像"
+        return (img_bytes, filename, mime_type), f"圖表已成功導出為{format_type}格式"
     
     except Exception as e:
         return None, f"導出圖表時出錯: {str(e)}"
 
 # 建立Gradio界面
-with gr.Blocks(title="數據可視化工具") as demo:
-    gr.Markdown("# 數據可視化工具")
-    gr.Markdown("上傳CSV或Excel文件，或直接在下方輸入數據來創建各種圖表")
+with gr.Blocks(title="進階數據可視化工具") as demo:
+    gr.Markdown("# 進階數據可視化工具")
+    gr.Markdown("上傳CSV或Excel文件，或直接在下方輸入數據來創建各種專業圖表")
     
     # 狀態變量
     data_state = gr.State(None)
+    custom_colors_state = gr.State({})
+    patterns_state = gr.State([])
     
     with gr.Tabs():
         # 數據輸入頁籤
@@ -199,7 +766,9 @@ with gr.Blocks(title="數據可視化工具") as demo:
                     upload_status = gr.Textbox(label="上傳狀態")
                 
                 with gr.Column():
-                    csv_input = gr.Textbox(label="或直接輸入數據（逗號或空格分隔）", placeholder="類別,數值\nA,10\nB,20\nC,15\nD,25\nE,30\n\n或\n\n類別 數值\nA 10\nB 20\nC 15\nD 25\nE 30", lines=10)
+                    csv_input = gr.Textbox(label="或直接輸入數據（逗號或空格分隔）", 
+                                          placeholder="類別,數值\nA,10\nB,20\nC,15\nD,25\nE,30\n\n或\n\n類別 數值\nA 10\nB 20\nC 15\nD 25\nE 30", 
+                                          lines=10)
                     parse_button = gr.Button("解析數據")
                     parse_status = gr.Textbox(label="解析狀態")
             
@@ -215,15 +784,19 @@ with gr.Blocks(title="數據可視化工具") as demo:
         # 圖表創建頁籤
         with gr.TabItem("圖表創建"):
             with gr.Row():
-                with gr.Column():
+                with gr.Column(scale=1):
                     chart_type = gr.Dropdown(
-                        ["長條圖", "折線圖", "圓餅圖", "散點圖", "區域圖"], 
+                        CHART_TYPES, 
                         label="圖表類型", 
                         value="長條圖"
                     )
                     
-                    x_column = gr.Dropdown(["類別"], label="X軸（或類別）")
-                    y_column = gr.Dropdown(["數值"], label="Y軸（或數值）")
+                    agg_function = gr.Dropdown(
+                        AGGREGATION_FUNCTIONS,
+                        label="聚合函數",
+                        value="求和",
+                        info="選擇如何彙總數據"
+                    )
                     
                     chart_title = gr.Textbox(label="圖表標題", placeholder="我的數據圖表")
                     
@@ -233,19 +806,92 @@ with gr.Blocks(title="數據可視化工具") as demo:
                         value="默認"
                     )
                 
-                with gr.Column():
+                with gr.Column(scale=1):
+                    # 軸和分組選擇
+                    x_column = gr.Dropdown(["類別"], label="X軸（或類別）")
+                    y_column = gr.Dropdown(["數值"], label="Y軸（或數值）")
+                    group_column = gr.Dropdown(["無"], label="分組列（用於多系列圖表）")
+                    size_column = gr.Dropdown(["無"], label="大小列（用於氣泡圖等）")
+                    
+                    # 尺寸控制
                     chart_width = gr.Slider(300, 1200, 700, label="圖表寬度")
                     chart_height = gr.Slider(300, 800, 500, label="圖表高度")
                     
-                    show_grid = gr.Checkbox(label="顯示網格", value=True)
-                    show_legend = gr.Checkbox(label="顯示圖例", value=True)
-                    
-                    update_button = gr.Button("更新圖表")
-                    download_button = gr.Button("導出為PNG圖像")
+                    # 顯示選項
+                    with gr.Row():
+                        show_grid = gr.Checkbox(label="顯示網格", value=True)
+                        show_legend = gr.Checkbox(label="顯示圖例", value=True)
             
+            # 圖案和顏色自定義區
             with gr.Row():
-                chart_output = gr.Plot(label="圖表預覽")
-                download_output = gr.Image(label="導出的圖表", visible=False)
+                with gr.Column():
+                    gr.Markdown("### 圖案和顏色設置")
+                    gr.Markdown("為圖表元素設置特定的填充圖案（適用於黑白印刷）和顏色")
+                    
+                    # 動態添加圖案，先默認提供三個
+                    with gr.Row():
+                        pattern1 = gr.Dropdown(PATTERN_TYPES, label="圖案1", value="無")
+                        pattern2 = gr.Dropdown(PATTERN_TYPES, label="圖案2", value="無")
+                        pattern3 = gr.Dropdown(PATTERN_TYPES, label="圖案3", value="無")
+                    
+                    # 自定義顏色區域
+                    color_customization = gr.Textbox(
+                        label="自定義顏色 (格式: 類別1:#FF0000,類別2:#00FF00)", 
+                        placeholder="A:#1f77b4,B:#ff7f0e,C:#2ca02c", 
+                        info="輸入類別名稱和十六進制顏色代碼，用逗號分隔多個項目"
+                    )
+                
+                with gr.Column():
+                    # 按鈕區
+                    update_button = gr.Button("更新圖表", variant="primary")
+                    
+                    with gr.Row():
+                        export_format = gr.Dropdown(
+                            ["PNG", "SVG", "PDF", "JPEG"], 
+                            label="導出格式", 
+                            value="PNG"
+                        )
+                        download_button = gr.Button("導出圖表")
+                    
+                    export_chart = gr.File(label="導出的圖表")
+                    export_chart_status = gr.Textbox(label="導出狀態")
+            
+            # 圖表預覽區
+            chart_output = gr.Plot(label="圖表預覽")
+    
+    # 輔助函數
+    def parse_custom_colors(color_text):
+        """解析自定義顏色文本"""
+        custom_colors = {}
+        if color_text and color_text.strip():
+            try:
+                pairs = color_text.split(',')
+                for pair in pairs:
+                    if ':' in pair:
+                        key, value = pair.split(':', 1)
+                        custom_colors[key.strip()] = value.strip()
+            except:
+                pass
+        return custom_colors
+    
+    def update_patterns(p1, p2, p3):
+        """更新圖案列表"""
+        patterns = []
+        if p1:
+            patterns.append(p1)
+        if p2:
+            patterns.append(p2)
+        if p3:
+            patterns.append(p3)
+        return patterns
+    
+    def process_group_column(group_col):
+        """處理分組列選擇"""
+        return None if group_col == "無" else group_col
+    
+    def process_size_column(size_col):
+        """處理大小列選擇"""
+        return None if size_col == "無" else size_col
     
     # 事件處理
     upload_button.click(
@@ -259,7 +905,7 @@ with gr.Blocks(title="數據可視化工具") as demo:
     ).then(
         update_columns,
         inputs=[data_state],
-        outputs=[x_column, y_column]
+        outputs=[x_column, y_column, group_column, size_column]
     )
     
     parse_button.click(
@@ -273,7 +919,7 @@ with gr.Blocks(title="數據可視化工具") as demo:
     ).then(
         update_columns,
         inputs=[data_state],
-        outputs=[x_column, y_column]
+        outputs=[x_column, y_column, group_column, size_column]
     )
     
     export_button.click(
@@ -282,43 +928,184 @@ with gr.Blocks(title="數據可視化工具") as demo:
         outputs=[export_result, export_status]
     )
     
+    # 處理圖案和顏色設置
+    color_customization.change(
+        parse_custom_colors,
+        inputs=[color_customization],
+        outputs=[custom_colors_state]
+    )
+    
+    pattern1.change(
+        update_patterns,
+        inputs=[pattern1, pattern2, pattern3],
+        outputs=[patterns_state]
+    )
+    
+    pattern2.change(
+        update_patterns,
+        inputs=[pattern1, pattern2, pattern3],
+        outputs=[patterns_state]
+    )
+    
+    pattern3.change(
+        update_patterns,
+        inputs=[pattern1, pattern2, pattern3],
+        outputs=[patterns_state]
+    )
+    
+    # 更新圖表
     update_button.click(
-        create_plot,
-        inputs=[data_state, chart_type, x_column, y_column, color_scheme, 
-                chart_title, chart_width, chart_height, show_grid, show_legend],
+        lambda df, chart_type, x_col, y_col, group_col, size_col, color_scheme, patterns, title, width, height, show_grid, show_legend, agg_func, custom_colors: 
+        create_plot(
+            df, chart_type, x_col, y_col, 
+            None if group_col == "無" else group_col, 
+            None if size_col == "無" else size_col,
+            color_scheme, patterns, title, width, height, 
+            show_grid, show_legend, agg_func, custom_colors
+        ),
+        inputs=[
+            data_state, chart_type, x_column, y_column, 
+            group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height, 
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
         outputs=[chart_output]
     )
     
+    # 導出圖表
     download_button.click(
         download_figure,
-        inputs=[chart_output],
-        outputs=[download_output, gr.Textbox(label="下載狀態")]
-    ).then(
-        lambda: gr.update(visible=True),
-        outputs=[download_output]
+        inputs=[chart_output, export_format],
+        outputs=[export_chart, export_chart_status]
+    )
+    
+    # 圖表類型改變時更新界面元素可見性
+    def update_element_visibility(chart_type):
+        """根據圖表類型更新UI元素的可見性"""
+        # 圓餅圖和環形圖不需要X軸，只需要類別和數值
+        is_pie_chart = chart_type in ["圓餅圖", "環形圖"]
+        
+        # 氣泡圖需要額外的大小控制列
+        needs_size_column = chart_type in ["氣泡圖", "甘特圖", "樹狀圖"]
+        
+        # 需要分組列的圖表類型
+        needs_group_column = chart_type in [
+            "群組長條圖", "堆疊長條圖", "多重折線圖", "堆疊區域圖",
+            "熱力圖", "雷達圖", "散點圖", "氣泡圖"
+        ]
+        
+        return (
+            gr.update(visible=not is_pie_chart, label="類別列" if is_pie_chart else "X軸"),
+            gr.update(visible=True, label="數值列" if is_pie_chart else "Y軸"),
+            gr.update(visible=needs_group_column, label="分組列"),
+            gr.update(visible=needs_size_column, label="大小列")
+        )
+    
+    chart_type.change(
+        update_element_visibility,
+        inputs=[chart_type],
+        outputs=[x_column, y_column, group_column, size_column]
     )
     
-    # 自動顯示圖表預覽
+    # 圖表相關自動更新函數
+    def auto_update_chart(df, chart_type, x_col, y_col, group_col, size_col, color_scheme, patterns_list, title, width, height, show_grid, show_legend, agg_func, custom_colors):
+        """圖表自動更新函數，用於各種輸入改變時"""
+        # 處理可能為"無"的選項
+        group_column_value = None if group_col == "無" else group_col
+        size_column_value = None if size_col == "無" else size_col
+        
+        return create_plot(
+            df, chart_type, x_col, y_col, group_column_value, size_column_value,
+            color_scheme, patterns_list, title, width, height, show_grid, show_legend, agg_func, custom_colors
+        )
+    
+    # 添加自動更新事件
     chart_type.change(
-        create_plot,
-        inputs=[data_state, chart_type, x_column, y_column, color_scheme, 
-                chart_title, chart_width, chart_height, show_grid, show_legend],
+        auto_update_chart,
+        inputs=[
+            data_state, chart_type, x_column, y_column, group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height, 
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
         outputs=[chart_output]
     )
     
     x_column.change(
-        create_plot,
-        inputs=[data_state, chart_type, x_column, y_column, color_scheme, 
-                chart_title, chart_width, chart_height, show_grid, show_legend],
+        auto_update_chart,
+        inputs=[
+            data_state, chart_type, x_column, y_column, group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height,
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
         outputs=[chart_output]
     )
     
     y_column.change(
-        create_plot,
-        inputs=[data_state, chart_type, x_column, y_column, color_scheme, 
-                chart_title, chart_width, chart_height, show_grid, show_legend],
+        auto_update_chart,
+        inputs=[
+            data_state, chart_type, x_column, y_column, group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height,
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
+        outputs=[chart_output]
+    )
+    
+    group_column.change(
+        auto_update_chart,
+        inputs=[
+            data_state, chart_type, x_column, y_column, group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height,
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
+        outputs=[chart_output]
+    )
+    
+    agg_function.change(
+        auto_update_chart,
+        inputs=[
+            data_state, chart_type, x_column, y_column, group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height,
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
         outputs=[chart_output]
     )
+    
+    color_scheme.change(
+        auto_update_chart,
+        inputs=[
+            data_state, chart_type, x_column, y_column, group_column, size_column,
+            color_scheme, patterns_state, chart_title, chart_width, chart_height,
+            show_grid, show_legend, agg_function, custom_colors_state
+        ],
+        outputs=[chart_output]
+    )
+
+# 在啟動應用之前添加使用說明
+with demo:
+    gr.Markdown("""
+    ## 使用說明
+    
+    ### 數據輸入
+    - 上傳CSV或Excel文件，或在文本框中直接輸入數據
+    - 第一行被視為欄位名稱（表頭），不會納入統計
+    - 支持逗號分隔(CSV)或空格分隔的數據格式
+    
+    ### 圖表創建
+    - 選擇圖表類型：長條圖、折線圖、圓餅圖等多種專業圖表
+    - 聚合函數：選擇如何彙總數據（求和、平均值、最大值等）
+    - 分組列：用於創建多系列圖表，如按類別分組的長條圖
+    - 大小列：用於氣泡圖等需要額外數值控制大小的圖表
+    
+    ### 自定義選項
+    - 圖案填充：為圖表元素設置填充圖案，特別適用於黑白印刷
+    - 自定義顏色：為特定類別設置顏色，格式為"類別1:#FF0000,類別2:#00FF00"
+    - 導出格式：支持PNG、SVG、PDF和JPEG格式導出
+    
+    ### 注意事項
+    - 不同圖表類型需要不同的數據組織形式
+    - 圓餅圖和環形圖只需要類別和數值列
+    - 複雜圖表（如熱力圖、雷達圖）需要適當的數據結構
+    """)
 
 # 啟動應用
 demo.launch()