guanlan-src/app_backEnd/app_backEnd_guanlan/report_export.py

# -*- coding: utf-8 -*-
"""报告导出 Word / Excel (2026-08-28 用户令: 导出 demo)。

汇报画布选中的块 → .docx / .xlsx。图表块不丢信息: 转成对应的数据表 —
Word 里放位图既不可编辑也不能核对, 上报集团要的是能核的数字。
脱敏跟随受众设置 (上报版沿用服务端 _redact 的规则)。
"""
import io, datetime as dt
import base64

TITLE = '如东海上风电场 · 运行分析报告'


def _fmt(v, n=0):
    try:
        return f'{float(v):,.{n}f}'
    except Exception:
        return str(v)


def blocks_data(fl, blocks, narr=None):
    """把画布块转成 [(标题, 段落[], 表格{head, rows})] — Word/Excel 共用同一份中间表示。"""
    E = (fl.get('control') or {}).get('energy') or {}
    R = fl.get('rel') or {}
    M = ((fl.get('m8') or {}).get('mtbf')) or {}
    F = fl.get('faults') or {}
    FU = fl.get('fus') or {}
    BD = FU.get('链盘') or {}
    C = fl.get('control') or {}
    out = []

    _cur = ['']                      # ★P101 当前块键（供图形使用）

    def add(t, paras=None, head=None, rows=None):
        out.append(dict(k=_cur[0], title=t, paras=paras or [], head=head, rows=rows or []))

    for k in blocks:
        _cur[0] = k
        if k == 'kpi':
            add('一、核心指标', head=['指标', '数值', '口径'], rows=[
                ['上网电量', f"{_fmt(E.get('act'))} MWh", f"理论可发 {_fmt(E.get('theo'))} − 损失 {_fmt(E.get('loss'))}"],
                ['等效满发小时', f"{_fmt(E.get('eflh'))} h", '上网电量 ÷ (台数×额定容量)'],
                ['时间可用率', f"{C.get('avail')}%", '分母已剔调度令与低风待机'],
                ['损失率', f"{E.get('loss_pct')}%", '损失 ÷ 理论可发'],
                ['平均停机间隔', f"{M.get('MTBF_h')} h", f"停机 {M.get('停机事件数')} 起, 单次均 {M.get('MDT_h')} h; 分母=窗内实测台时"],
                ['需行动机组', f"{len(BD.get('rows') or [])} 台", f"四源融合判级; 定论 {(FU.get('kpi') or {}).get('定论')} · 预警 {(FU.get('kpi') or {}).get('预警')} · 候选 {(FU.get('kpi') or {}).get('候选观察')}"],
            ])
        elif k == 'water':
            rows = [['理论可发', _fmt(E.get('theo')), '100%']]
            for x in (E.get('items') or []):
                rows.append([x['k'], '−' + _fmt(x['loss']), f"{x['loss']/max(E.get('theo') or 1,1)*100:.1f}%"])
            rows.append(['实际上网', _fmt(E.get('act')), f"{(E.get('act') or 0)/max(E.get('theo') or 1,1)*100:.1f}%"])
            add('二、发电完成情况', [f"理论可发 {_fmt(E.get('theo'))} MWh = 上网 {_fmt(E.get('act'))} + 损失 {_fmt(E.get('loss'))} MWh, 损失率 {E.get('loss_pct')}%。"],
                ['项目', 'MWh', '占理论可发'], rows)
        elif k == 'emon':
            add('逐月发电', head=['月份', '上网 MWh', '损失 MWh', '损失率', '等效满发 h', '数据覆盖'],
                rows=[[x['m'], _fmt(x['act']), _fmt(x['loss']), f"{x['loss_pct']}%", _fmt(x['eflh']),
                       f"{x['cov']*100:.0f}%" + ('(不满月)' if x['cov'] < 0.5 else '')] for x in (E.get('monthly') or [])])
        elif k == 'blame':
            cat = lambda kk: '交易与调度侧' if ('限电' in kk or '调度' in kk) else ('设备侧' if kk == '停机' else '运行侧')
            g = {}
            for x in (E.get('items') or []):
                g[cat(x['k'])] = g.get(cat(x['k']), 0) + x['loss']
            tot = sum(g.values()) or 1
            add('损失责任分账', ['进设备考核的仅"故障与检修停机"; 限电属交易与调度侧。'],
                ['责任方', '损失 MWh', '占比'], [[a, _fmt(b), f'{b/tot*100:.0f}%'] for a, b in sorted(g.items(), key=lambda x: -x[1])])
        elif k in ('ram', 'parts') and (R.get('表')):
            add('三、设备可靠性' if k == 'ram' else '部件停机明细',
                [f"总停机 {_fmt(R.get('总停机h'))} h = 设备类 {_fmt(R.get('设备类停机h'))} + 外部类 {_fmt(R.get('外部类停机h'))} (电网/风况/运维操作不计设备可靠性)。"],
                ['部件', '停机 h', '事件数', 'MTBF h', 'MDT h', '影响台数'],
                [[x['部件'] + ('(外部)' if x['外部'] else ''), _fmt(x['停机时长h']), x['事件数'],
                  x['MTBF_h'] or '—', x['MDT_h'], x['影响台数']] for x in (R.get('表') or [])])
        elif k == 'attrib' and R.get('表'):
            nr = next((x['停机时长h'] for x in R['表'] if x['部件'] == '无记录停机'), 0)
            add('归因覆盖', [f"无记录停机 {_fmt(nr)} h, 占总停机 "
                             f"{nr/max(R.get('总停机h') or 1,1)*100:.1f}% — 属归因缺口, 非无故障; 成因待定。"])
        elif k == 'pareto':
            add('四、故障统计', [f"窗内 {_fmt(F.get('total'))} 条报警 / {F.get('n_codes')} 个码。"],
                ['报警码与名称', '条数'], [[x['k'], _fmt(x['v'])] for x in (F.get('pareto_n') or [])])
        elif k == 'quad':
            add('故障模式分型', ['条数高不等于问题严重: 单次时长在分钟级以下的属重复触发, 不应按条数排定优先级。'],
                ['码', '名称', '条数', '单次中位 s', '影响台数'],
                [[x['code'], x['k'], _fmt(x['n']), x['med'], x['nt']] for x in (F.get('quad') or [])[:12]])
        elif k == 'frate':
            mo = F.get('monthly') or {}
            add('月度故障率', ['按天归一, 数据不满月单独标注。'], ['月份', '条数', '条/日', '数据覆盖'],
                [[m, _fmt(v), (mo.get('per_day') or [None])[i] if i < len(mo.get('per_day') or []) else '—',
                  f"{(mo.get('cov') or [1])[i]*100:.0f}%" if i < len(mo.get('cov') or []) else '—']
                 for i, (m, v) in enumerate(zip(mo.get('months') or [], mo.get('vals') or []))])
        elif k == 'units':
            add('五、重点机组与处置',
                [f"需跟踪 {len(BD.get('rows') or [])} 台 / 38。瓶颈: " +
                 ', '.join(f"{v} 台处于「{kk}」环节" for kk, v in (BD.get('stuck') or {}).items()) + '。'],
                ['机组', '设备状态', '部件', '证据源', '六步进度', '当前环节'],
                [[r['t'].replace('WTG', '') + '#', r.get('ostate') or '—',
                  (r.get('part') or '')[:22], '/'.join(r.get('srcs') or []) or '—',
                  f"{r['done']}/6", r.get('stuck') or '全通'] for r in (BD.get('rows') or [])])
        elif k == 'chain':
            add('决策链进度', ['六步: 证据 → 机制 → 判级 → 排期 → 动作 → 验收。'],
                ['环节', '停滞机组数'], [[kk, v] for kk, v in (BD.get('stuck') or {}).items()])
        elif k == 'loop':
            add('六、检修执行与闭环', ['检出 → 换件 → 复测回落的历史例证。'],
                ['机组', '部件', '换件日期', '换前', '换后', '比值'],
                [[x['turbine'].replace('WTG', '') + '#', x['component'], x['replace_date'], x['pre'], x['post'], f"{x['ratio']}×"]
                 for x in (BD.get('closed') or [])])
        elif k == 'plan':
            add('七、下期计划与建议', [
                f"1. 补机制定性 — {(BD.get('stuck') or {}).get('机制', 0)} 台已判级但根因未定, 排期与动作缺乏依据, 建议按预警级优先安排现场检查与取样。",
                '2. 回收 2025~2026 工单 — 台账止于 2024-11, 本期检修执行与闭环无法自动判定。',
                '3. 油液采样 — 目前多数机组末样时效偏旧, 仅可作回测锚, 建议纳入定期采样。',
                '4. 限电侧 — 损失主要发生在交易与调度侧, 设备侧改善空间有限。'])
        elif k == 'caliber':
            part = F.get('partial') or []
            add('八、口径与数据边界', [
                f"统计期 {fl.get('win')}" + (f"; 其中 {'/'.join(part)} 数据不满月, 月度对比已按天归一, 环比不纳入残月。" if part else '。'),
                '时间可用率分母已剔除调度令与低风待机; 损失率分母为理论可发, 两者分母不同, 不可直接互推。',
                f"停机口径: 可靠性按 ≥1h 段计 {_fmt(R.get('总停机h'))} h, 电量账按逐条工况态计; 两者差额为短于 1h 的停机段, 不可相减。",
                '工单台账 2020-01 ~ 2024-11, 现场未提供 2025~2026, 本报告不对本期检修执行下结论。',
                '判级由振动/温度/润滑/油液四源融合得出; 行动等级由综合报告出具。'])
        elif k == 'narr' and narr:
            add('运行总结', [narr])
    return out



# ── ★P101 报告配图（用户令 2026-10-06：导出 word 要有图形）────────────────────────────
# 说明：页面上的图是前端 SVG 渲染的；Word 里改用 matplotlib 画**同类图**（同数据、同语义），
#      不经前端字节，保证离线可复现。图类型与页面章节一一对应。
_CHART_KINDS = {'water', 'blame', 'ram', 'attrib', 'pareto', 'quad', 'frate'}


def _figure_png(k, fl):
    """按块键画一张 PNG（bytes）；不支持的块返回 None。"""
    import matplotlib
    matplotlib.use('Agg')
    import matplotlib.pyplot as plt
    from matplotlib import font_manager  # noqa: F401

    E = (fl.get('control') or {}).get('energy') or {}
    Rl = fl.get('rel') or {}
    F = fl.get('faults') or {}
    try:
        plt.rcParams['font.sans-serif'] = ['Microsoft YaHei', 'SimHei', 'DejaVu Sans']
        plt.rcParams['axes.unicode_minus'] = False
    except Exception:
        pass
    fig = None
    try:
        if k == 'water':
            items = list(E.get('items') or [])
            labels = ['理论可发'] + [str(x.get('k')) for x in items] + ['实际上网']
            vals = [float(E.get('theo') or 0)] + [-float(x.get('loss') or 0) for x in items] + [float(E.get('act') or 0)]
            fig, ax = plt.subplots(figsize=(7.2, 3.2), dpi=170)
            colors = ['#1E7FC2'] + ['#C8302B'] * len(items) + ['#2E8B57']
            ax.bar(range(len(vals)), vals, color=colors)
            ax.set_xticks(range(len(vals)))
            ax.set_xticklabels(labels, fontsize=8, rotation=18)
            ax.axhline(0, color='#888', lw=.8)
            ax.set_title('发电完成（瀑布）', fontsize=10)
            ax.grid(axis='y', ls=':', alpha=.35)
        elif k in ('blame', 'attrib'):
            if k == 'blame':
                cat = lambda kk: '交易与调度侧' if ('限电' in kk or '调度' in kk) else ('设备侧' if kk == '停机' else '运行侧')
                g = {}
                for x in (E.get('items') or []):
                    g[cat(x.get('k'))] = g.get(cat(x.get('k')), 0) + float(x.get('loss') or 0)
                title = '损失归因'
            else:
                g = {str(x.get('部件')): float(x.get('停机时长h') or 0) for x in (Rl.get('表') or [])}
                title = '停机时长归因'
            items = sorted(g.items(), key=lambda kv: -kv[1])[:10]
            fig, ax = plt.subplots(figsize=(7.2, 3.0), dpi=170)
            ax.barh([a for a, _ in items][::-1], [b for _, b in items][::-1], color='#C8302B')
            ax.set_title(title, fontsize=10)
            ax.grid(axis='x', ls=':', alpha=.35)
        elif k == 'ram':
            rows = (Rl.get('表') or [])[:10]
            fig, ax = plt.subplots(figsize=(7.2, 3.0), dpi=170)
            ax.bar([str(x.get('部件'))[:10] for x in rows], [float(x.get('停机时长h') or 0) for x in rows], color='#D99A2B')
            ax.set_title('可靠性（按部件停机时长）', fontsize=10)
            ax.tick_params(axis='x', labelrotation=20, labelsize=8)
            ax.grid(axis='y', ls=':', alpha=.35)
        elif k == 'pareto':
            rows = sorted(((str(x.get('k')), float(x.get('v') or 0)) for x in (F.get('pareto_n') or [])),
                          key=lambda kv: -kv[1])[:12]
            fig, ax = plt.subplots(figsize=(7.2, 3.0), dpi=170)
            ax.bar([a[:12] for a, _ in rows], [b for _, b in rows], color='#1E7FC2')
            tot = sum(b for _, b in rows) or 1
            cum, acc = [], 0.0
            for _, b in rows:
                acc += b
                cum.append(acc / tot * 100)
            ax2 = ax.twinx()
            ax2.plot(range(len(cum)), cum, color='#C8302B', marker='o', ms=3)
            ax2.set_ylim(0, 105)
            ax.set_title('报警帕累托', fontsize=10)
            ax.tick_params(axis='x', labelrotation=20, labelsize=8)
            ax.grid(axis='y', ls=':', alpha=.35)
        elif k == 'quad':
            pts = [(float(x.get('x') or 0), float(x.get('y') or 0), str(x.get('k') or '')) for x in (F.get('quad') or [])]
            fig, ax = plt.subplots(figsize=(5.6, 4.0), dpi=170)
            if pts:
                ax.scatter([p[0] for p in pts], [p[1] for p in pts], s=26, color='#1E7FC2')
                for x, y, lab in pts[:20]:
                    ax.annotate(lab[:8], (x, y), fontsize=6, xytext=(2, 2), textcoords='offset points')
            ax.set_title('故障四象限', fontsize=10)
            ax.grid(ls=':', alpha=.35)
        elif k == 'frate':
            mo = (F.get('monthly') or {})
            months, vals = list(mo.get('months') or []), list(mo.get('vals') or [])
            fig, ax = plt.subplots(figsize=(7.2, 2.8), dpi=170)
            ax.plot(range(len(vals)), vals, color='#C8302B', marker='o', ms=3)
            ax.set_xticks(range(len(months)))
            ax.set_xticklabels([str(m)[-2:] for m in months], fontsize=8)
            ax.set_title('故障率（月）', fontsize=10)
            ax.grid(ls=':', alpha=.35)
        if fig is None:
            return None
        buf = io.BytesIO()
        fig.tight_layout()
        fig.savefig(buf, format='png')
        plt.close(fig)
        return buf.getvalue()
    except Exception:
        try:
            if fig is not None:
                plt.close(fig)
        except Exception:
            pass
        return None


def _decode_png(v):
    """把前端传来的 dataURL 或裸 base64 解成 PNG 字节；失败返回 None。"""
    if not v:
        return None
    if isinstance(v, (bytes, bytearray)):
        return bytes(v)
    s = str(v)
    if s.startswith('data:'):
        s = s.split(',', 1)[-1]
    try:
        return base64.b64decode(s)
    except Exception:
        return None


def to_docx(fl, blocks, win, aud='in', narr=None, images=None):
    """★P102 images：{块键: dataURL|base64} —— 前端按页面原样栅格化的图，**优先原样嵌入**，
    缺失的块才回落到 matplotlib 重绘（保证 Word 与页面完全一致）。"""
    _user_images = images or {}
    from docx import Document
    from docx.shared import Pt
    from docx.enum.text import WD_ALIGN_PARAGRAPH
    doc = Document()
    st = doc.styles['Normal']
    st.font.name = 'Microsoft YaHei'
    st.font.size = Pt(10.5)
    h = doc.add_heading(TITLE, level=0)
    h.alignment = WD_ALIGN_PARAGRAPH.CENTER
    sub = doc.add_paragraph(f"统计期 {win} ｜ 38 × SWT-4.0-130" + ('  ｜ 上报版(判据门槛与特征频率已脱敏)' if aud == 'up' else ''))
    sub.alignment = WD_ALIGN_PARAGRAPH.CENTER
    for b in blocks_data(fl, blocks, narr):
        doc.add_heading(b['title'], level=1)
        _png = _decode_png(_user_images.get(b.get('k')))
        if _png is None and b.get('k') in _CHART_KINDS:
            _png = _figure_png(b.get('k'), fl)
        if _png:
            try:
                from docx.shared import Inches
                doc.add_picture(io.BytesIO(_png), width=Inches(6.4))
            except Exception:
                pass
        for p in b['paras']:
            doc.add_paragraph(p)
        if b['head'] and b['rows']:
            t = doc.add_table(rows=1, cols=len(b['head']))
            t.style = 'Light Grid Accent 1'
            for i, c in enumerate(b['head']):
                t.rows[0].cells[i].text = str(c)
            for r in b['rows']:
                cells = t.add_row().cells
                for i, c in enumerate(r[:len(b['head'])]):
                    cells[i].text = str(c)
    doc.add_paragraph()
    f = doc.add_paragraph(f"数据源 SCADA 10min · 报警事件 · 工单台账 · CMS 振动 · 油液化验　｜　出件 {dt.datetime.now():%Y-%m-%d}")
    f.runs[0].font.size = Pt(8.5)
    buf = io.BytesIO(); doc.save(buf); return buf.getvalue()


def to_xlsx(fl, blocks, win, aud='in', narr=None):
    from openpyxl import Workbook
    from openpyxl.styles import Font, PatternFill
    wb = Workbook(); wb.remove(wb.active)
    hdr_f = Font(bold=True, color='FFFFFF')
    hdr_b = PatternFill('solid', fgColor='1E4D6B')
    for b in blocks_data(fl, blocks, narr):
        name = b['title'].split('、')[-1][:28] or 'sheet'
        ws = wb.create_sheet(name[:31])
        r = 1
        for p in b['paras']:
            ws.cell(r, 1, p); r += 1
        if b['paras']:
            r += 1
        if b['head']:
            for i, c in enumerate(b['head'], 1):
                cell = ws.cell(r, i, c); cell.font = hdr_f; cell.fill = hdr_b
            r += 1
            for row in b['rows']:
                for i, c in enumerate(row[:len(b['head'])], 1):
                    ws.cell(r, i, c)
                r += 1
            for i, c in enumerate(b['head'], 1):
                w = max([len(str(c))] + [len(str(x[i-1])) for x in b['rows'] if len(x) >= i]) + 2
                ws.column_dimensions[ws.cell(1, i).column_letter].width = min(w * 1.6, 52)
    if not wb.sheetnames:
        wb.create_sheet('空')
    buf = io.BytesIO(); wb.save(buf); return buf.getvalue()