🔬 启动研究
{archetype && ARCHETYPES_CONFIG[archetype] && (
{ARCHETYPES_CONFIG[archetype].icon} {ARCHETYPES_CONFIG[archetype].name} 维度
)}
setTopic(e.target.value)} onKeyDown={e=>e.key==='Enter'&&runQuant()} placeholder="输入策略研究主题，例如: 涨停板块轮动短线策略" className="flex-1 bg-surface-3 border border-border rounded-xl px-4 py-2.5 text-sm text-zinc-200 focus:outline-none focus:border-brand-500/40 transition" />
{presets.map((p,i) => ())}
{(running || optimizing) && ({optimizing ? '策略优化中' : '量化流水线执行中'}... 请耐心等待
)}
{/* Sub-tabs */}
{[{id:'library',label:'因子库',icon:'📦'},{id:'screener',label:'DSL 选股',icon:'🎯'},{id:'analysis',label:'健康分析',icon:'🔬'},{id:'combine_history',label:'融合记录',icon:'🔮'},{id:'exhaustive',label:'穷举融合',icon:'🧪'},{id:'pipeline',label:'自动化 管线',icon:'🔗'},{id:'mine',label:'启动挖掘',icon:'⛏️'},{id:'logs',label:'挖掘日志',icon:'📋',badge:diggerLogs.length},{id:'memory',label:'记忆',icon:'🧠'}].map(t=>(
))}
{/* === 因子库 Sub-tab === */}
{diggerSubTab === 'library' && (<>
{/* Header + Actions */}
📦 因子库
{factorList.length > 0 && (() => {
const gradeCount = {A:0,B:0,C:0,D:0,E:0};
factorList.forEach(f => { const g = getFactorGrade(f); gradeCount[g] = (gradeCount[g]||0)+1; });
return (<>
{gradeCount.A>0&&}
{gradeCount.B>0&&}
{gradeCount.C>0&&}
{gradeCount.D>0&&}
{gradeCount.E>0&&}
);
})()}
{selectedForCombine.length > 0 && (
)}
{/* Stats row */}
{factorStats && (
{/* ABCDE 评级分布行 */}
{(factorStats.grade_dist || factorStats.tier_dist) && (() => {
const gd = factorStats.grade_dist || {};
return (
评级:
{['A','B','C','D','E'].map(g => {
const cnt = gd[g] || 0;
if (!cnt) return null;
return (
);
})}
);
})()}
{/* 统计指标行 */}
{[
{label:'活跃因子',value:factorStats.active_count,color:'text-emerald-400'},
{label:'最佳 Sharpe',value:(factorStats.best_sharpe||0).toFixed(2),color:'text-brand-400'},
{label:'最佳 IR',value:(factorStats.best_ir||0).toFixed(2),color:'text-amber-400'},
{label:'⚡ 向量化',value:factorStats.complexity_dist?.vectorized||0,color:'text-cyan-400'},
{label:'🐌 嵌套循环',value:factorStats.complexity_dist?.nested||0,color:'text-red-400'},
{label:'可融合',value:(factorStats.complexity_dist?.vectorized||0)+(factorStats.complexity_dist?.apply||0),color:factorStats.ready_to_combine?'text-emerald-400':'text-zinc-500'},
].map((m,i)=>(
))}
{Object.entries(factorStats.theme_distribution).sort((a,b)=>b[1]-a[1]).slice(0,12).map(([t,c])=>(
))}
)}
{/* Search + Filter + Sort toolbar */}
setFactorSearch(e.target.value)} placeholder="搜索因子（主题/ID/代码关键词）..."
className="flex-1 min-w-[200px] bg-surface-3 rounded-lg px-3 py-1.5 text-[12px] text-zinc-300 border border-border placeholder-zinc-600 focus:border-brand-500/50 focus:outline-none transition" />
{/* Factor list table */}
{(() => {
const filtered = getFilteredFactors();
const totalPages = Math.max(1, Math.ceil(filtered.length / FACTOR_PAGE_SIZE));
const currentPage = Math.min(factorPage, totalPages - 1);
const pageStart = currentPage * FACTOR_PAGE_SIZE;
const paged = filtered.slice(pageStart, pageStart + FACTOR_PAGE_SIZE);
return filtered.length > 0 ? (
显示 {pageStart+1}-{pageStart+paged.length} / {filtered.length} (共 {factorList.length} 个因子)
{factorFilter === 'grade_ABC' && (() => {
const de = factorList.filter(f=>{ const g=getFactorGrade(f); return g==='D'||g==='E'; }).length;
return de > 0 ?
: null;
})()}
{(factorFilter==='grade_D'||factorFilter==='grade_E'||factorFilter==='low_pool') && (
)}
{purgePreview && !purgePreview.dry_run === false && (
)}
{purgePreview && purgePreview.dry_run && (
预览: {purgePreview.to_retire} / {purgePreview.total_active} 个因子不达标
{(purgePreview.preview||[]).map((p,i) => (
{(p.id||'').slice(-6)}
{p.theme}
{p.reasons.join(' / ')}
))}
)}
{totalPages > 1 && (
{Array.from({length: Math.min(totalPages, 10)}, (_, i) => {
const p = totalPages <= 10
? i
: Math.max(0, Math.min(currentPage - 4, totalPages - 10)) + i;
return (
);
})}
共 {totalPages} 页
)}
) : !factorListLoading ? (
{factorList.length > 0 ? '没有匹配的因子，试试调整搜索条件' : '暂无因子数据，点击上方「刷新」加载'}
) : null; })()}
{/* Factor detail panel */}
{selectedFactor && (
{/* Header */}
{selectedFactor.sub_theme || selectedFactor.theme}
{selectedFactor.id}
{(() => { const t = FACTOR_TIER[getFactorTier(selectedFactor)]; return t ? {t.icon} {t.label} : null; })()}
{selectedFactor.complexity==='vectorized'?'⚡ 向量化':selectedFactor.complexity==='apply'?'⚠️ apply':'🐌 嵌套循环'}
{selectedFactor.code_lines && {selectedFactor.code_lines} 行}
创建: {selectedFactor.created_at?new Date(selectedFactor.created_at*1000).toLocaleString():'N/A'}
{selectedFactor.theme && selectedFactor.theme !== selectedFactor.sub_theme && 分类: {selectedFactor.theme}}
{/* Tabs */}
{[
{id:'metrics',label:'📊 指标'},
{id:'code',label:'💻 代码'},
{id:'usage',label:'📖 使用方法'},
{id:'narrative',label:'💡 叙述'}, // F4
{id:'neighbors',label:'🔗 邻近因子'}, // F4
].map(t=>(
))}
{/* Metrics tab */}
{factorDetailTab === 'metrics' && (
{[
{label:'Sharpe Ratio',value:(selectedFactor.sharpe||0).toFixed(3),desc:'风险调整收益，>1 优秀',good:selectedFactor.sharpe>=1},
{label:'IR (Sortino)',value:(selectedFactor.ir||0).toFixed(3),desc:'下行风险调整收益，>1 优秀',good:Math.abs(selectedFactor.ir)>=1},
{label:'Mean IC',value:(selectedFactor.ic_mean||0).toFixed(4),desc:'因子预测能力，|IC|>0.03 有效',good:Math.abs(selectedFactor.ic_mean)>=0.03},
{label:'Win Rate',value:((selectedFactor.win_rate||0)*100).toFixed(1)+'%',desc:'盈利交易占比，>50% 为佳',good:selectedFactor.win_rate>=0.5},
{label:'交易次数',value:selectedFactor.trades||0,desc:'回测期总交易数，>100 样本充足',good:selectedFactor.trades>=100},
{label:'最大回撤',value:((selectedFactor.max_drawdown||0)*100).toFixed(2)+'%',desc:'最大权益回撤，<20% 为佳',good:selectedFactor.max_drawdown<0.2},
{label:'换手率',value:(selectedFactor.turnover||0).toFixed(4),desc:'日均换手率',good:true},
{label:'单调性',value:(selectedFactor.monotonicity||0).toFixed(4),desc:'分组收益单调性，越接近1越好',good:selectedFactor.monotonicity>=0.5},
].map((m,i)=>(
{m.label}
{m.value}
{m.desc}
))}
generate_factor 源码 ({selectedFactor.code_lines||'?'} 行)
{(selectedFactor.code||'(代码不可用)').split('\n').map((line,i)=>(
{i+1}{line}
))}
)}
{/* Usage tab */}
{factorDetailTab === 'usage' && (
如何使用此因子
1
直接回测
将 generate_factor(matrices) 代码复制到回测系统。函数接收 OHLCV 矩阵字典，返回同 shape 的 DataFrame —— 值越高表示越看好该股票（做多信号越强）。
2
策略化研发
点击「转入策略研发」，系统会基于此因子自动生成完整交易策略（含入场/出场/仓位管理），并在量化研发模块中迭代优化。因子 → 策略的转化由 LLM 自动完成。
3
因子融合
勾选多个因子后点击「融合」，系统将多因子等权平均组合成复合因子，在 后端沙箱重新回测。融合通常能提升 Sharpe 和稳定性。注意：仅 ⚡ 向量化因子可参与融合（🐌 嵌套循环因子会导致沙箱超时）。
指标速查
{[
{metric:'Sharpe > 1',meaning:'优秀因子，风险调整后收益显著'},
{metric:'|IC| > 0.03',meaning:'因子对未来收益有预测能力'},
{metric:'IR > 1',meaning:'信息比率高，因子收益稳定'},
{metric:'Win Rate > 50%',meaning:'多数交易盈利'},
{metric:'Trades > 100',meaning:'回测样本充足，统计可靠'},
{metric:'Max DD < 20%',meaning:'风险可控，回撤较小'},
].map((r,i)=>(
{r.metric}
{r.meaning}
))}
调用示例 (Python)
{`# 因子 ID: ${selectedFactor.id}
# ${selectedFactor.sub_theme || selectedFactor.theme}
import pandas as pd
# 准备数据：每个 key 是一个 DataFrame (index=日期, columns=股票代码)
matrices = {
'open': open_df, # shape: (120天, 6300只)
'high': high_df,
'low': low_df,
'close': close_df,
'volume': volume_df,
}
# 调用因子
factor = generate_factor(matrices)
# factor 是 DataFrame, 同 shape
# 每日每股一个因子值，值越高 = 越看好
top_stocks = factor.iloc[-1].nlargest(20) # 今日 Top 20`}
)}
{/* F4 — narrative tab */}
{factorDetailTab === 'narrative' && (
)}
{/* F4 — neighbors tab */}
{factorDetailTab === 'neighbors' && (
{
const next = factorList.find(f => f.id === fid);
if (next) setSelectedFactor(next);
}} />
)}
{/* Action buttons */}
{selectedFactor.status==='active' && (<>
)}
)}
)}
{/* === 健康分析 Sub-tab === */}
{diggerSubTab === 'screener' && (
DSL 因子选股
输入 generate_factor 代码，直接调用后端 screener 运行选股。
因子代码
（或从因子库选择）
{selectedFactor && (
)}