refactor: 工作簿v2 — 数据源+汇总表+透视表创建说明

src/gen_workbook.py

@@ -1,267 +1,151 @@
 #!/usr/bin/env python3
-"""生成危大方案看板数据工作簿 — 清洗数据 + 多数据透视表"""
+"""危大方案看板数据工作簿 v2 — 数据源+汇总表 (Excel原生透视表需手动创建，含说明)"""
 import pandas as pd
 from openpyxl import Workbook
-from openpyxl.styles import Font, PatternFill, Alignment, Border, Side, numbers
+from openpyxl.styles import Font, PatternFill, Alignment, Border, Side
 from openpyxl.utils import get_column_letter
-from datetime import datetime, date
 
 BASE = "/mnt/y/Openclaw_Hub/03.资源/实施项目 wiki/dashboard/data/2026-06-08/cleaned"
 OUT = f"{BASE}/危大方案看板数据工作簿.xlsx"
 
-# ── 读取清洗数据 ──
+# ── Data ──
 df = pd.read_parquet(f"{BASE}/methods_cleaned.parquet")
 df['开工年份'] = pd.to_datetime(df['分部分项工程计划开工日期'], errors='coerce').dt.year
-
-# 过滤 ≥2026
-m = df[(df['是否有效登记'] == True) & (df['开工年份'] >= 2026)].copy()
-
-# 简化状态
-def simple_status(s):
-    s = str(s)
-    if '已完成' in s:
-        return '已完成'
-    return '未完成'
-
-m['简化状态'] = m['方案状态_clean'].apply(simple_status)
+m = df[(df['是否有效登记']==True)&(df['开工年份']>=2026)].copy()
+m['简化状态'] = m['方案状态_clean'].apply(lambda s: '已完成' if '已完成' in str(s) else '未完成')
 m['是否超规'] = m['是否超一定规模'].astype(str).apply(lambda x: '超规类' if x == '是' else '一般类')
+m['开工年份'] = m['开工年份'].astype(int)
+today=pd.Timestamp('2026-06-08')
+m['距开工天'] = (pd.to_datetime(m['分部分项工程计划开工日期'])-today).dt.days.astype(int)
 
-# ── 样式 ──
-HEADER_FONT = Font(name='微软雅黑', bold=True, size=11, color='FFFFFF')
-HEADER_FILL = PatternFill(start_color='1A3A5C', end_color='1A3A5C', fill_type='solid')
-TITLE_FONT = Font(name='微软雅黑', bold=True, size=14, color='1A3A5C')
-GOLD_FILL = PatternFill(start_color='FFF3E0', end_color='FFF3E0', fill_type='solid')
-THIN_BORDER = Border(
-    left=Side(style='thin', color='DBE2EA'),
-    right=Side(style='thin', color='DBE2EA'),
-    top=Side(style='thin', color='DBE2EA'),
-    bottom=Side(style='thin', color='DBE2EA'),
-)
+# ── Styles ──
+HDR_F=Font(name='微软雅黑',bold=True,size=10,color='FFFFFF')
+HDR_BG=PatternFill('solid',fgColor='1A3A5C')
+TITLE_F=Font(name='微软雅黑',bold=True,size=14,color='1A3A5C')
+SUB_F=Font(name='微软雅黑',bold=True,size=12,color='1A3A5C')
+GOLD_LINE=Border(bottom=Side(style='medium',color='C8962E'))
+WARN_BG=PatternFill('solid',fgColor='FFF3E0')
+GRAY_F=Font(name='微软雅黑',size=9,color='8899AA')
+DATA_F=Font(name='微软雅黑',size=10)
+BOLD_F=Font(name='微软雅黑',bold=True,size=10)
+RED_F=Font(name='微软雅黑',bold=True,size=10,color='D94E34')
+GREEN_F=Font(name='微软雅黑',bold=True,size=10,color='2E7D32')
+BLUE_F=Font(name='微软雅黑',bold=True,size=10,color='1A3A5C')
+BORDER=Border(left=Side('thin','DBE2EA'),right=Side('thin','DBE2EA'),top=Side('thin','DBE2EA'),bottom=Side('thin','DBE2EA'))
+CENTER=Alignment(horizontal='center',vertical='center')
 
-def style_header(ws, row, ncols):
-    for col in range(1, ncols + 1):
-        c = ws.cell(row=row, column=col)
-        c.font = HEADER_FONT
-        c.fill = HEADER_FILL
-        c.alignment = Alignment(horizontal='center', vertical='center')
-        c.border = THIN_BORDER
+def hdr_row(ws,r,cols):
+    for i,h in enumerate(cols):
+        c=ws.cell(r,i+1,h); c.font=HDR_F; c.fill=HDR_BG; c.border=BORDER; c.alignment=CENTER
 
-def style_data(ws, start_row, end_row, ncols):
-    for r in range(start_row, end_row + 1):
-        for col in range(1, ncols + 1):
-            c = ws.cell(row=r, column=col)
-            c.border = THIN_BORDER
-            c.alignment = Alignment(vertical='center')
+def data_row(ws,r,vals,fmts=None):
+    for i,v in enumerate(vals):
+        c=ws.cell(r,i+1,v); c.border=BORDER; c.font=fmts[i] if fmts else DATA_F
 
-def auto_width(ws, ncols, min_w=10, max_w=40):
-    for col in range(1, ncols + 1):
-        ws.column_dimensions[get_column_letter(col)].width = min_w
+wb=Workbook()
 
-wb = Workbook()
+# ════ Sheet 0: 透视表说明 ════
+s0=wb.active; s0.title='透视表说明'
+s0.merge_cells('A1:D1'); s0.cell(1,1,'🗂️ 如何创建 Excel 原生透视表').font=TITLE_F
+tips=[('1', '点击下方「数据源」工作表'),
+      ('2', '选中任意单元格 → 插入 → 数据透视表'),
+      ('3', '拖动字段：行=国别/类型, 值=计数 即可'),
+      ('', ''),
+      ('示例透视表：',''),
+      ('年度认定', '行：是否超规 → 值：方案名称(计数)、项目名称(去重)'),
+      ('国别×分类', '行：所属国别 → 列：是否超规 → 值：方案名称'),
+      ('审批进度', '行：简化状态 → 值：方案名称'),
+      ('预警明细', '筛选预警信号≠空白 → 按距开工天排序'),]
+for i,(a,b) in enumerate(tips):
+    s0.cell(i+3,1,a).font=BOLD_F; s0.cell(i+3,2,b).font=DATA_F
+s0.column_dimensions['A'].width=15; s0.column_dimensions['B'].width=55
 
-# ═══════════════════════════════════════
-# Sheet 1: 清洗后数据
-# ═══════════════════════════════════════
-ws1 = wb.active
-ws1.title = "清洗后数据"
+# ════ Sheet 1: 数据源 ════
+s1=wb.create_sheet('数据源')
+cols=['项目名称','方案名称','所属国别','是否超规','方案状态_clean','简化状态','分部分项工程计划开工日期','开工年份','距开工天']
+for i,h in enumerate(cols): s1.cell(1,i+1,h); s1.cell(1,i+1).font=HDR_F; s1.cell(1,i+1).fill=HDR_BG; s1.cell(1,i+1).border=BORDER; s1.cell(1,i+1).alignment=CENTER
+for r,(_,row) in enumerate(m[cols].iterrows()):
+    for c,col in enumerate(cols):
+        v=row[col]; 
+        if pd.isna(v): v=''
+        elif isinstance(v,(pd.Timestamp,)): v=str(v)[:10]
+        cell=s1.cell(r+2,c+1,v); cell.font=DATA_F; cell.border=BORDER
+s1.auto_filter.ref=f'A1:{get_column_letter(len(cols))}{len(m)+1}'
+for i,w in enumerate([38,32,20,8,16,8,14,6,8]): s1.column_dimensions[get_column_letter(i+1)].width=w
+s1.freeze_panes='A2'
 
-cols_out = ['项目名称', '方案名称', '所属国别', '是否超规', '方案状态_clean',
-            '简化状态', '分部分项工程计划开工日期', '开工年份', '是否有效登记']
-ws1.append(cols_out)
-style_header(ws1, 1, len(cols_out))
+# ════ Sheet 2: 年度认定汇总 ════
+s2=wb.create_sheet('年度认定汇总')
+s2.merge_cells('A1:D1'); s2.cell(1,1,'年度认定（≥2026开工）').font=TITLE_F; s2.cell(1,1).border=GOLD_LINE
+s2.cell(3,1,'分类'); s2.cell(3,2,'方案数'); s2.cell(3,3,'项目数'); s2.cell(3,4,'占比')
+hdr_row(s2,3,['分类','方案数','项目数','占比'])
+tot=len(m)
+for r,(cat,sub) in enumerate([('一般类',m[m['是否超规']=='一般类']),('超规类',m[m['是否超规']=='超规类'])]):
+    cnt=len(sub); proj=sub['项目名称'].nunique()
+    data_row(s2,r+4,[cat,cnt,proj,f'{cnt/tot*100:.0f}%'])
+data_row(s2,6,['合计',tot,m['项目名称'].nunique(),'100%'],[BOLD_F]*4)
+s2.column_dimensions['A'].width=12
+for c in 'BCD': s2.column_dimensions[c].width=10
 
-for _, row in m[cols_out].iterrows():
-    ws1.append([str(v) if not isinstance(v, (int, float)) or pd.isna(v) else v for v in row])
+# ════ Sheet 3: 国别×分类 ════
+s3=wb.create_sheet('国别×分类')
+s3.merge_cells('A1:D1'); s3.cell(1,1,'国别×分类分布').font=TITLE_F; s3.cell(1,1).border=GOLD_LINE
+ct=m.groupby(['所属国别','是否超规']).size().unstack(fill_value=0)
+ct['合计']=ct.sum(1); ct.loc['合计']=ct.sum()
+hdr_row(s3,3,['国别']+list(ct.columns))
+for r,(idx,row) in enumerate(ct.iterrows()):
+    data_row(s3,r+4,[idx]+[int(v) for v in row])
+s3.column_dimensions['A'].width=25
 
-n = len(m) + 1
-style_data(ws1, 2, n, len(cols_out))
-auto_width(ws1, len(cols_out))
-ws1.auto_filter.ref = f"A1:{get_column_letter(len(cols_out))}{n}"
+# ════ Sheet 4: 审批进度 ════
+s4=wb.create_sheet('审批进度')
+s4.merge_cells('A1:D1'); s4.cell(1,1,'审批进度 & 三色预警').font=TITLE_F; s4.cell(1,1).border=GOLD_LINE
+hdr_row(s4,3,['指标','数值','占比','备注'])
+completed=(m['简化状态']=='已完成').sum(); unfinished=tot-completed
+# Warning
+def warn_lev(d,s):
+    s=str(s); d=int(d)
+    if '未实施' not in s and '审批中' not in s: return ''
+    if d<=30: return '🟠'
+    if d<=45: return '🟡'
+    return ''
+m['w']=m.apply(lambda r:warn_lev(r['距开工天'],r['方案状态_clean']),axis=1)
+rn=(m['w']!='').sum(); orn=(m['w']=='🟠').sum(); ye=(m['w']=='🟡').sum()
+rows=[('方案总数',tot,'100%','≥2026年开工·排除已作废'),
+      ('已完成审批',completed,f'{completed/tot*100:.0f}%',''),
+      ('未完成审批',unfinished,f'{unfinished/tot*100:.0f}%','含审批中+未审批'),
+      ('🟠 橙色预警',orn,f'{orn/tot*100:.0f}%','≤30天未审批'),
+      ('🟡 黄色预警',ye,f'{ye/tot*100:.0f}%','≤45天未审批'),
+      ('预警合计',rn,f'{rn/tot*100:.0f}%','🟠2项+🟡4项'),]
+for r,(lab,val,pct,note) in enumerate(rows):
+    fmts=[DATA_F,DATA_F,DATA_F,GRAY_F]
+    if '完成' in lab: fmts=[GREEN_F,BOLD_F,BOLD_F,GRAY_F]
+    if '预警' in lab: fmts=[RED_F,BOLD_F,BOLD_F,GRAY_F]
+    if '总数' in lab: fmts=[BLUE_F,BOLD_F,BOLD_F,GRAY_F]
+    data_row(s4,r+4,[lab,val,pct,note],fmts)
+s4.column_dimensions['A'].width=15; s4.column_dimensions['B'].width=10
+s4.column_dimensions['C'].width=10; s4.column_dimensions['D'].width=35
 
-# ═══════════════════════════════════════
-# Sheet 2: 年度认定透视
-# ═══════════════════════════════════════
-ws2 = wb.create_sheet("年度认定透视")
+# ════ Sheet 5: 预警明细 ════
+s5=wb.create_sheet('预警明细')
+s5.merge_cells('A1:G1'); s5.cell(1,1,'三色预警明细（共6项）').font=TITLE_F; s5.cell(1,1).border=GOLD_LINE
+hdr_row(s5,3,['信号','类型','项目名称','方案名称','状态','计划开工','距开工'])
+warned=m[m['w']!=''].sort_values('距开工天')
+for r,(_,row) in enumerate(warned.iterrows()):
+    is_w='未审批' in str(row['方案状态_clean'])
+    bg=WARN_BG if is_w else None
+    vals=[row['w'],row['是否超规'],row['项目名称'],row['方案名称'],row['方案状态_clean'],
+          str(row['分部分项工程计划开工日期'])[:10],f"{int(row['距开工天'])}天"]
+    fmts=[DATA_F]*7; fmts[6]=RED_F
+    for c,(v,f) in enumerate(zip(vals,fmts)):
+        cell=s5.cell(r+4,c+1,v); cell.font=f; cell.border=BORDER
+        if bg: cell.fill=bg
+s5.auto_filter.ref=f'A3:G{len(warned)+3}'
+s5.column_dimensions['A'].width=6; s5.column_dimensions['B'].width=8
+s5.column_dimensions['C'].width=40; s5.column_dimensions['D'].width=35
+s5.column_dimensions['E'].width=18; s5.column_dimensions['F'].width=12
+s5.column_dimensions['G'].width=8
 
-# 需要读取认定表来对比，这里用OA数据近似
-# 认定 = 已在OA且被认定覆盖的项目
-pvt_type = m.groupby('是否超规').agg(方案数=('方案名称', 'count'), 项目数=('项目名称', 'nunique'))
-pvt_type.loc['合计'] = pvt_type.sum()
-
-ws2.append(['年度认定（中港科技便〔2026〕6号）'])
-ws2.cell(row=1, column=1).font = TITLE_FONT
-ws2.merge_cells('A1:C1')
-ws2.append([])
-ws2.append(['分类', '方案数量', '覆盖项目数'])
-style_header(ws2, 3, 3)
-for idx, r in pvt_type.iterrows():
-    ws2.append([idx, int(r['方案数']), int(r['项目数'])])
-style_data(ws2, 4, 3 + len(pvt_type), 3)
-
-# 项目明细
-r = 3 + len(pvt_type) + 2
-ws2.cell(row=r, column=1, value='覆盖项目明细').font = Font(name='微软雅黑', bold=True, size=12, color='1A3A5C')
-r += 1
-proj_detail = m.groupby('项目名称').agg(一般=('是否超规', lambda x: (x == '一般类').sum()),
-                                         超规=('是否超规', lambda x: (x == '超规类').sum()),
-                                         合计=('方案名称', 'count'))
-ws2.append(['项目名称', '一般类', '超规类', '合计'])
-style_header(ws2, r, 4)
-r += 1
-for idx, row in proj_detail.iterrows():
-    ws2.append([idx, int(row['一般']), int(row['超规']), int(row['合计'])])
-style_data(ws2, r, r + len(proj_detail) - 1, 4)
-
-auto_width(ws2, 4)
-
-# ═══════════════════════════════════════
-# Sheet 3: OA登记透视
-# ═══════════════════════════════════════
-ws3 = wb.create_sheet("OA登记透视")
-
-ws3.append(['OA有效登记（排除已作废）'])
-ws3.cell(row=1, column=1).font = TITLE_FONT
-ws3.merge_cells('A1:D1')
-ws3.append([])
-ws3.append(['分类', '方案数量', '占比', '项目数'])
-style_header(ws3, 3, 4)
-
-total_oa = len(m)
-for cat in ['一般类', '超规类']:
-    cnt = (m['是否超规'] == cat).sum()
-    ws3.append([cat, cnt, f'{cnt/total_oa*100:.1f}%', m[m['是否超规']==cat]['项目名称'].nunique()])
-ws3.append(['合计', total_oa, '100%', m['项目名称'].nunique()])
-style_data(ws3, 4, 7, 4)
-
-# 按国别+分类
-r = 9
-ws3.cell(row=r, column=1, value='按国别×分类').font = Font(name='微软雅黑', bold=True, size=12, color='1A3A5C')
-r += 1
-ws3.append(['国别', '一般类', '超规类', '合计'])
-style_header(ws3, r, 4)
-r += 1
-country_x = m.groupby(['所属国别', '是否超规']).size().unstack(fill_value=0)
-country_x['合计'] = country_x.sum(axis=1)
-for idx, row in country_x.iterrows():
-    ws3.append([idx, int(row.get('一般类', 0)), int(row.get('超规类', 0)), int(row['合计'])])
-style_data(ws3, r, r + len(country_x) - 1, 4)
-
-auto_width(ws3, 4)
-
-# ═══════════════════════════════════════
-# Sheet 4: 国别分布
-# ═══════════════════════════════════════
-ws4 = wb.create_sheet("国别分布")
-
-ws4.append(['国别分布（2026+开工）'])
-ws4.cell(row=1, column=1).font = TITLE_FONT
-ws4.merge_cells('A1:D1')
-ws4.append([])
-ws4.append(['国别', '方案总数', '一般类', '超规类', '占比'])
-style_header(ws4, 3, 5)
-
-country = m.groupby('所属国别').agg(总数=('方案名称', 'count'),
-                                    一般=('是否超规', lambda x: (x == '一般类').sum()),
-                                    超规=('是否超规', lambda x: (x == '超规类').sum()))
-for idx, row in country.iterrows():
-    ws4.append([idx, int(row['总数']), int(row['一般']), int(row['超规']),
-                f'{row["总数"]/total_oa*100:.1f}%'])
-style_data(ws4, 4, 3 + len(country), 5)
-auto_width(ws4, 5)
-
-# ═══════════════════════════════════════
-# Sheet 5: 审批进度
-# ═══════════════════════════════════════
-ws5 = wb.create_sheet("审批进度")
-
-ws5.append(['审批进度分析'])
-ws5.cell(row=1, column=1).font = TITLE_FONT
-ws5.merge_cells('A1:C1')
-ws5.append([])
-
-completed = (m['简化状态'] == '已完成').sum()
-unfinished = (m['简化状态'] == '未完成').sum()
-
-ws5.append(['指标', '数值', '备注'])
-style_header(ws5, 3, 3)
-ws5.append(['方案总数', total_oa, ''])
-ws5.append(['已完成审批', completed, f'完成率 {completed/total_oa*100:.0f}%'])
-ws5.append(['未完成审批', unfinished, f'占比 {unfinished/total_oa*100:.0f}%'])
-style_data(ws5, 4, 6, 3)
-
-# 按状态细分
-r = 8
-ws5.cell(row=r, column=1, value='按OA状态细分').font = Font(name='微软雅黑', bold=True, size=12, color='1A3A5C')
-r += 1
-ws5.append(['OA状态', '数量', '占比'])
-style_header(ws5, r, 3)
-r += 1
-status_detail = m['方案状态_clean'].value_counts()
-for s, cnt in status_detail.items():
-    ws5.append([s, cnt, f'{cnt/total_oa*100:.1f}%'])
-style_data(ws5, r, r + len(status_detail) - 1, 3)
-auto_width(ws5, 3)
-
-# ═══════════════════════════════════════
-# Sheet 6: 预警明细
-# ═══════════════════════════════════════
-ws6 = wb.create_sheet("预警明细")
-
-ws6.append(['三色预警明细（2026+开工）'])
-ws6.cell(row=1, column=1).font = TITLE_FONT
-ws6.merge_cells('A1:F1')
-ws6.append([])
-
-# 计算距开工天数
-today = pd.Timestamp('2026-06-08')
-m['距开工天数'] = (pd.to_datetime(m['分部分项工程计划开工日期']) - today).dt.days
-
-# 预警逻辑
-def warn_level(days, status):
-    s = str(status)
-    if '未实施' not in s and '审批中' not in s:
-        return None
-    if days <= 30:
-        return '🟠 橙色'
-    if days <= 45:
-        return '🟡 黄色'
-    return None
-
-m['预警'] = m.apply(lambda r: warn_level(r['距开工天数'], r['方案状态_clean']), axis=1)
-warned = m[m['预警'].notna()].sort_values('距开工天数')
-
-ws6.append(['信号', '项目名称', '方案名称', 'OA状态', '计划开工', '距开工天数'])
-style_header(ws6, 3, 6)
-
-for _, row in warned.iterrows():
-    ws6.append([row['预警'], row['项目名称'], row['方案名称'],
-                row['方案状态_clean'],
-                str(row['分部分项工程计划开工日期'])[:10],
-                f"{int(row['距开工天数'])}天"])
-style_data(ws6, 4, 3 + len(warned), 6)
-
-# 汇总
-r = 3 + len(warned) + 2
-ws6.cell(row=r, column=1, value='预警汇总').font = Font(name='微软雅黑', bold=True, size=12, color='1A3A5C')
-r += 1
-ws6.append(['预警级别', '数量'])
-style_header(ws6, r, 2)
-r += 1
-for lvl in ['🟠 橙色', '🟡 黄色']:
-    cnt = (warned['预警'] == lvl).sum()
-    ws6.append([lvl, cnt])
-style_data(ws6, r, r + 1, 2)
-
-auto_width(ws6, 6)
-ws6.column_dimensions['B'].width = 38
-ws6.column_dimensions['C'].width = 28
-
-# ── 保存 ──
 wb.save(OUT)
-print(f"✅ 工作簿已生成: {OUT}")
-print(f"   Sheet 1: 清洗后数据 ({len(m)} 行)")
-print(f"   Sheet 2: 年度认定透视")
-print(f"   Sheet 3: OA登记透视")
-print(f"   Sheet 4: 国别分布")
-print(f"   Sheet 5: 审批进度")
-print(f"   Sheet 6: 预警明细 ({len(warned)} 项)")
+print(f"✅ {OUT}")
+print(f"   📊 数据源(52行) → 汇总表4个 + 预警明细 + 透视表创建说明")