{"id":119483,"date":"2026-10-29T21:01:01","date_gmt":"2026-10-29T13:01:01","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/119483.html"},"modified":"2026-10-29T21:01:01","modified_gmt":"2026-10-29T13:01:01","slug":"sa-k-02a-%e7%9f%a5%e8%af%86%e4%b8%8e%e6%a3%80%e7%b4%a2%e5%b7%a5%e7%a8%8b-%e6%96%87%e6%a1%a3%e8%a7%a3%e6%9e%90%e4%b8%8a-%e7%ae%a1%e7%ba%bf%e4%ba%94%e7%8e%af%e4%b8%8e%e8%b4%a8%e9%87%8f%e5%9b%9b%e5%b0%ba","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/119483.html","title":{"rendered":"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a"},"content":{"rendered":"<h2>\u6587\u6863\u89e3\u6790\u4e22\u7684\u4e0d\u662f\u5b57&#xff0c;\u662f\u7ed3\u6784&#xff1a;\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u628a\u5c3a\u5b50<\/h2>\n<p>\u7cfb\u5217\u5bfc\u822a&#xff1a;00 \u7cfb\u5217\u5bfc\u822a \u00b7 01 \u56db\u7c7b\u77e5\u8bc6 \u00b7 \u672c\u7bc7&#xff08;02A&#xff09; \u00b7 02B \u65ad\u88c2\u4fee\u590d \u00b7 03 \u5207\u5206\u6807\u5b9a \u00b7 04 \u67e5\u8be2\u7406\u89e3 \u00b7 05 \u6df7\u5408\u68c0\u7d22\u4e0e\u91cd\u6392 \u00b7 06 \u5f15\u7528\u9a8c\u8bc1 \u00b7 07 \u77e5\u8bc6\u65b0\u9c9c\u5ea6 \u00b7 08 \u591a\u6e90\u4e00\u81f4\u6027 \u00b7 09 \u7ba1\u9053\u95e8\u7981<\/p>\n<p>\u5c0f\u63d0\u4e00\u53e5&#xff1a; \u7531\u4e8e\u5e73\u53f0\u6bcf\u65e5\u53d1\u5e03\u7b14\u8bb0\u6570\u91cf\u9650\u5236&#xff0c;\u76ee\u524d\u7cfb\u5217\u7b14\u8bb0\u8fd8\u6ca1\u5168\u90e8\u4e0a\u4f20\u3002\u540e\u7eed\u5185\u5bb9\u4f1a\u6301\u7eed\u66f4\u65b0&#xff0c;\u5927\u5bb6\u53ef\u4ee5\u5148\u70b9\u4e2a\u5173\u6ce8\u3001\u6536\u85cf&#xff0c;\u4ee5\u514d\u9519\u8fc7\u3002 \u8c22\u8c22&#xff01;<\/p>\n<hr \/>\n<h3>\u5f15\u5b50&#xff1a;\u5b57\u5168\u5bf9&#xff0c;\u7b54\u6848\u8fd8\u662f\u9519\u7684<\/h3>\n<p>runbook RB-214 \u91cc\u767d\u7eb8\u9ed1\u5b57\u5199\u7740\u300c\u78c1\u76d8\u6c34\u4f4d &gt;85% \u5148\u6e05\u65e5\u5fd7\u518d\u6269\u5bb9\u300d&#xff0c;\u95ee\u300cRB-214 \u7b2c\u4e09\u6b65\u505a\u4ec0\u4e48\u300d&#xff0c;\u7b54\u51fa\u6765\u7684\u662f\u300c\u68c0\u67e5\u76d1\u63a7\u9762\u677f\u300d\u3002<\/p>\n<p>\u7ffb\u539f\u59cb PDF&#xff0c;\u90a3\u53e5\u8bdd\u5728\u4e00\u4e2a\u4e09\u5217\u56db\u884c\u7684\u5904\u7f6e\u52a8\u4f5c\u8868\u91cc\u3002\u89e3\u6790\u5668\u5410\u51fa\u6765\u7684\u6587\u672c\u957f\u8fd9\u6837&#xff1a;<\/p>\n<p>[\u539f PDF \u7684\u89c6\u89c9\u7ed3\u6784]<br \/>\n\u6b65\u9aa4 | \u89e6\u53d1\u6761\u4ef6      | \u5904\u7f6e\u52a8\u4f5c<br \/>\n1    | CPU &gt; 80%     | \u6269\u5bb9<br \/>\n3    | \u9519\u8bef\u7387 &gt; 5%   | \u56de\u6eda<br \/>\n2    | \u78c1\u76d8\u6c34\u4f4d &gt;85% | \u5148\u6e05\u65e5\u5fd7\u518d\u6269\u5bb9<\/p>\n<p>[\u7eaf\u6587\u672c\u62bd\u53d6\u8f93\u51fa]<br \/>\n\u6b65\u9aa4 \u89e6\u53d1\u6761\u4ef6 \u5904\u7f6e\u52a8\u4f5c 1 CPU &gt; 80% \u6269\u5bb9 3 \u9519\u8bef\u7387 &gt; 5% \u56de\u6eda 2 \u78c1\u76d8\u6c34\u4f4d &gt;85% \u5148\u6e05\u65e5\u5fd7\u518d\u6269\u5bb9<\/p>\n<p>\u5b57\u4e00\u4e2a\u6ca1\u9519&#xff0c;\u5217\u6ca1\u4e86\u3002\u300c\u78c1\u76d8\u6c34\u4f4d &gt;85%\u300d\u548c\u300c\u5148\u6e05\u65e5\u5fd7\u518d\u6269\u5bb9\u300d\u4e4b\u95f4\u9694\u4e86\u5341\u51e0\u4e2a token&#xff0c;\u4e2d\u95f4\u8fd8\u5939\u7740\u53e6\u5916\u4e24\u884c\u7684\u788e\u7247\u3002\u5230 embedding \u773c\u91cc&#xff0c;\u8fd9\u4e00\u5768\u8ddf\u300cCPU \u6269\u5bb9\u300d\u7684\u8bed\u4e49\u8ddd\u79bb&#xff0c;\u6bd4\u8ddf\u300c\u78c1\u76d8\u5904\u7f6e\u300d\u8fd8\u8fd1\u3002<\/p>\n<p>\u7b2c\u4e00\u53cd\u5e94\u662f\u300c\u8bc6\u522b\u4e0d\u51c6&#xff0c;\u4e0a OCR\u300d\u3002\u4e0a\u4e86\u4e4b\u540e\u5b57\u8fd8\u662f\u5168\u5bf9\u2014\u2014\u56e0\u4e3a\u4ece\u5934\u5230\u5c3e\u5c31\u6ca1\u6709\u8bc6\u522b\u9519\u8bef\u3002<\/p>\n<p>\u672c\u6587\u53ea\u505a\u4e00\u4ef6\u4e8b&#xff1a;\u628a\u89e3\u6790\u7ba1\u7ebf\u7684\u4e94\u4e2a\u73af\u8282\u5404\u81ea\u4e22\u4ec0\u4e48\u4fe1\u606f\u62c6\u5f00&#xff0c;\u7ed9\u51fa\u56db\u628a\u53ef\u8ba1\u7b97\u7684\u5c3a\u5b50&#xff0c;\u4ee5\u53ca\u7528 20 \u7bc7\u81ea\u5df1\u7684\u771f\u5b9e\u6587\u6863\u6807\u5b9a\u51fa\u8be5\u4e0a\u54ea\u4e00\u6863\u65b9\u6848\u3002 \u4e0d\u8bb2 VLM \u9009\u578b\u6559\u7a0b\u3001\u4e0d\u505a OCR \u5f15\u64ce\u6a2a\u5411\u5bf9\u6bd4\u3001\u4e0d\u8bb2\u5207\u5206\u53c2\u6570\u600e\u4e48\u5b9a&#xff08;\u90a3\u662f {{URL:03}} \u7684\u4e8b&#xff09;\u2014\u2014\u8fd9\u4e09\u5757\u5404\u5c5e\u522b\u7684\u4e3b\u9898&#xff0c;\u4e00\u53e5\u6307\u5411\u3002<\/p>\n<p>{{URL:01}} \u5904\u7406\u7684\u662f\u300c\u77e5\u8bc6\u653e\u9519\u4e86\u5c42\u300d&#xff0c;\u672c\u7bc7\u63a5\u5b83\u7684\u4e0b\u4e00\u73af&#xff1a;\u77e5\u8bc6\u5728\u8fdb\u5e93\u90a3\u4e00\u523b\u5c31\u5df2\u7ecf\u574f\u4e86\u3002<\/p>\n<hr \/>\n<h3>\u4e00\u3001\u89e3\u6790\u4e22\u7684\u4fe1\u606f\u91cc&#xff0c;\u54ea\u4e00\u534a\u624d\u8ba9\u4f60\u5bdf\u89c9&#xff1f;<\/h3>\n<p>\u5148\u7ed9\u8fd9\u53e5\u5224\u65ad&#xff1a;\u8bc6\u522b\u9519\u8bef\u4f1a\u62a5\u9519&#xff0c;\u7ed3\u6784\u4e22\u5931\u4e0d\u62a5\u9519\u3002<\/p>\n<p>\u8bc6\u522b\u9519\u8bef\u7684\u4ea7\u7269\u662f\u4e71\u7801\u3001\u7f3a\u5b57\u3001\u5f02\u5e38\u6808\u3001\u7a7a\u9875&#xff0c;\u4efb\u4f55\u4e00\u6761\u90fd\u4f1a\u88ab\u6d41\u6c34\u7ebf\u76d1\u63a7\u6293\u4f4f\u3002\u7ed3\u6784\u4e22\u5931\u7684\u4ea7\u7269\u662f\u4e00\u6bb5\u901a\u987a\u3001\u5b8c\u6574\u3001\u770b\u8d77\u6765\u5b8c\u5168\u6b63\u5e38\u7684\u6587\u672c\u2014\u2014\u5b83\u53ea\u662f\u4e0d\u518d\u643a\u5e26\u300c\u8fd9\u662f\u4e2a\u8868\u683c\u7b2c\u4e09\u884c\u7b2c\u4e8c\u5217\u300d\u8fd9\u4ef6\u4e8b\u3002\u540e\u8005\u7684\u4ee3\u4ef7\u4e0d\u5728\u89e3\u6790\u9636\u6bb5\u7ed3\u7b97&#xff0c;\u800c\u5728\u53ec\u56de\u9636\u6bb5&#xff1a;\u8be5\u6392\u7b2c\u4e00\u7684\u7247\u6bb5\u6392\u5230\u4e86\u7b2c\u4e03&#xff0c;\u4e14\u6c38\u8fdc\u4e0d\u4f1a\u6709\u4eba\u6536\u5230\u544a\u8b66\u3002<\/p>\n<p>\u628a\u89e3\u6790\u62c6\u6210\u4e94\u73af\u4e4b\u540e&#xff0c;\u6bcf\u4e00\u73af\u4e22\u7684\u4e1c\u897f\u662f\u4e0d\u4e00\u6837\u7684\u3002<\/p>\n<p>\u5148\u8bf4\u4e3a\u4ec0\u4e48\u8981\u6309\u73af\u62c6\u3002\u300c\u89e3\u6790\u8d28\u91cf\u4e0d\u597d\u300d\u8fd9\u53e5\u8bdd\u6ca1\u6709\u4fe1\u606f\u91cf&#xff0c;\u56e0\u4e3a\u5b83\u4e0d\u6307\u5411\u4efb\u4f55\u4e00\u4e2a\u53ef\u6539\u7684\u5730\u65b9&#xff1a;\u89e3\u7801\u7f3a OCR\u3001\u7248\u9762\u5668\u6ca1\u5f00\u5206\u680f\u68c0\u6d4b\u3001\u8868\u683c\u8fd8\u539f\u6ca1\u63a5\u3001\u6392\u5e8f\u7b56\u7565\u9009\u9519\u3001\u5143\u6570\u636e\u6ca1\u53d6\u5b57\u53f7\u2014\u2014\u4e94\u4ef6\u4e8b\u7684\u4fee\u6cd5\u5b8c\u5168\u4e0d\u540c&#xff0c;\u6df7\u5728\u4e00\u8d77\u6d4b&#xff0c;\u53ea\u4f1a\u5f97\u5230\u4e00\u4e2a\u300c\u6574\u4f53\u5dee\u4e00\u70b9\u300d\u7684\u6a21\u7cca\u7ed3\u8bba\u3002\u6309\u73af\u62c6\u5f00\u4e4b\u540e&#xff0c;\u6bcf\u4e00\u73af\u90fd\u80fd\u6302\u4e00\u4e2a\u68c0\u6d4b\u70b9&#xff0c;\u68c0\u6d4b\u70b9\u51fa\u6765\u7684\u6570\u518d\u6c47\u6210\u540e\u9762\u90a3\u56db\u628a\u5c3a\u5b50\u3002<\/p>\n<p>\u4e94\u73af\u662f&#xff1a;\u89e3\u7801\u3001\u7248\u9762\u5206\u6790\u3001\u8868\u683c\u8fd8\u539f\u3001\u9605\u8bfb\u987a\u5e8f\u3001\u5143\u6570\u636e\u62bd\u53d6\u3002\u524d\u56db\u73af\u4f9d\u6b21\u52a0\u5de5\u524d\u4e00\u73af\u7684\u4ea7\u7269&#xff0c;\u5143\u6570\u636e\u90a3\u4e00\u73af\u662f\u65c1\u8def\u2014\u2014\u5b83\u4e0d\u52a0\u5de5\u5185\u5bb9&#xff0c;\u53ea\u7ed9\u5185\u5bb9\u8d34\u9875\u7801\u4e0e\u7ae0\u8282\u8def\u5f84\u3002<\/p>\n<p>\u8fd8\u6709\u4e00\u4e2a\u5bb9\u6613\u641e\u53cd\u7684\u5730\u65b9&#xff1a;\u4e94\u73af\u662f\u4e32\u8054\u7684&#xff0c;\u6240\u4ee5\u540e\u4e00\u73af\u6551\u4e0d\u4e86\u524d\u4e00\u73af\u3002\u89e3\u7801\u73af\u8282\u5df2\u7ecf\u628a\u8868\u683c\u62c6\u6210\u4e86\u6563\u5b57&#xff0c;\u8868\u683c\u8fd8\u539f\u90a3\u4e00\u73af\u518d\u5f3a\u4e5f\u63a5\u4e0d\u56de\u6765&#xff1b;\u8fd9\u4e5f\u662f\u300c\u5148\u5b9a\u4f4d\u54ea\u4e00\u73af\u3001\u518d\u51b3\u5b9a\u5347\u54ea\u4e00\u6863\u300d\u5fc5\u987b\u5148\u4e8e\u300c\u6362\u4e2a\u89e3\u6790\u5668\u8bd5\u8bd5\u300d\u7684\u539f\u56e0\u3002<\/p>\n<p>  #mermaid-svg-xDwMwcXwPsXN4RV6{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-xDwMwcXwPsXN4RV6 .error-icon{fill:#552222;}#mermaid-svg-xDwMwcXwPsXN4RV6 .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-xDwMwcXwPsXN4RV6 .marker{fill:#333333;stroke:#333333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .marker.cross{stroke:#333333;}#mermaid-svg-xDwMwcXwPsXN4RV6 svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-xDwMwcXwPsXN4RV6 p{margin:0;}#mermaid-svg-xDwMwcXwPsXN4RV6 .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .cluster-label text{fill:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .cluster-label span{color:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .cluster-label span p{background-color:transparent;}#mermaid-svg-xDwMwcXwPsXN4RV6 .label text,#mermaid-svg-xDwMwcXwPsXN4RV6 span{fill:#333;color:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .node rect,#mermaid-svg-xDwMwcXwPsXN4RV6 .node circle,#mermaid-svg-xDwMwcXwPsXN4RV6 .node ellipse,#mermaid-svg-xDwMwcXwPsXN4RV6 .node polygon,#mermaid-svg-xDwMwcXwPsXN4RV6 .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .rough-node .label text,#mermaid-svg-xDwMwcXwPsXN4RV6 .node .label text,#mermaid-svg-xDwMwcXwPsXN4RV6 .image-shape .label,#mermaid-svg-xDwMwcXwPsXN4RV6 .icon-shape .label{text-anchor:middle;}#mermaid-svg-xDwMwcXwPsXN4RV6 .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .rough-node .label,#mermaid-svg-xDwMwcXwPsXN4RV6 .node .label,#mermaid-svg-xDwMwcXwPsXN4RV6 .image-shape .label,#mermaid-svg-xDwMwcXwPsXN4RV6 .icon-shape .label{text-align:center;}#mermaid-svg-xDwMwcXwPsXN4RV6 .node.clickable{cursor:pointer;}#mermaid-svg-xDwMwcXwPsXN4RV6 .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .arrowheadPath{fill:#333333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-xDwMwcXwPsXN4RV6 .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-xDwMwcXwPsXN4RV6 .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-xDwMwcXwPsXN4RV6 .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-xDwMwcXwPsXN4RV6 .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .cluster text{fill:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 .cluster span{color:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-xDwMwcXwPsXN4RV6 .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-xDwMwcXwPsXN4RV6 rect.text{fill:none;stroke-width:0;}#mermaid-svg-xDwMwcXwPsXN4RV6 .icon-shape,#mermaid-svg-xDwMwcXwPsXN4RV6 .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-xDwMwcXwPsXN4RV6 .icon-shape p,#mermaid-svg-xDwMwcXwPsXN4RV6 .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-xDwMwcXwPsXN4RV6 .icon-shape .label rect,#mermaid-svg-xDwMwcXwPsXN4RV6 .image-shape .label rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-xDwMwcXwPsXN4RV6 .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-xDwMwcXwPsXN4RV6 .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-xDwMwcXwPsXN4RV6 :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u539f\u59cb\u6587\u6863PDF\/\u626b\u63cf\u4ef6\/HTML<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u89e3\u7801<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u7248\u9762\u5206\u6790<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u8868\u683c\u8fd8\u539f<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u9605\u8bfb\u987a\u5e8f<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u5143\u6570\u636e\u62bd\u53d6<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u7ed3\u6784\u5316\u5757\u6d41<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u68c0\u6d4b\u70b9:\u7a7a\u6587\u672c\u5c42\u7387<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u68c0\u6d4b\u70b9:\u6807\u9898\u5c42\u7ea7<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u68c0\u6d4b\u70b9:\u884c\u5217\u5bf9\u9f50<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u68c0\u6d4b\u70b9:\u8de8\u9875\u7eed\u63a5<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u68c0\u6d4b\u70b9:\u9875\u7801\u4e0e\u8def\u5f84<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u56db\u628a\u5c3a\u5b50\u8bc4\u5206<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u65b9\u6848\u9009\u62e9A\/B\/C \u4e09\u6863<\/p>\n<p><\/span><\/p>\n<p>\u540c\u5c42\u7684\u51e0\u4e2a\u540d\u8bcd\u5148\u5b9a\u4f4d&#xff0c;\u540e\u9762\u53cd\u590d\u7528\u5230&#xff1a;<\/p>\n<table>\n<tr>\u540d\u8bcd\u4e00\u53e5\u8bdd\u539f\u7406\u77e5\u8bc6\u7ed3\u6784\u5b9a\u4f4d\u672c\u6587\u8026\u5408\u70b9<\/tr>\n<tbody>\n<tr>\n<td>\u7248\u9762\u5206\u6790<\/td>\n<td>\u7528\u5750\u6807\u3001\u5b57\u53f7\u3001\u7a7a\u767d\u628a\u9875\u9762\u5207\u6210\u6709\u7c7b\u578b\u7684\u533a\u57df<\/td>\n<td>\u4e0a\u4f4d&#xff1a;\u6587\u6863\u667a\u80fd&#xff1b;\u540c\u65cf&#xff1a;OCR\u3001\u7248\u9762\u6811&#xff1b;\u4e0b\u4f4d&#xff1a;\u5206\u680f\u68c0\u6d4b<\/td>\n<td>\u5b83\u9519\u4e86\u540e\u9762\u56db\u73af\u5168\u9519\u2014\u2014\u8868\u683c\u8fd8\u539f\u62ff\u5230\u7684\u53ef\u80fd\u4e0d\u662f\u8868\u683c\u533a\u57df<\/td>\n<\/tr>\n<tr>\n<td>\u9605\u8bfb\u987a\u5e8f<\/td>\n<td>\u628a\u65e0\u5e8f\u7684\u7248\u9762\u533a\u57df\u6062\u590d\u6210\u4eba\u9605\u8bfb\u7684\u7ebf\u6027\u5e8f\u5217<\/td>\n<td>\u4e0a\u4f4d&#xff1a;\u7248\u9762\u5206\u6790\u7684\u4e0b\u6e38&#xff1b;\u540c\u65cf&#xff1a;XY-Cut\u3001\u7248\u9762\u6811\u904d\u5386<\/td>\n<td>\u53cc\u680f\u6587\u6863\u5728\u8fd9\u91cc\u4e32\u680f&#xff0c;\u662f\u8de8\u5757\u5b8c\u6574\u7387\u7684\u7b2c\u4e00\u5927\u6740\u624b<\/td>\n<\/tr>\n<tr>\n<td>\u8868\u683c\u8fd8\u539f<\/td>\n<td>\u4ece\u8868\u683c\u533a\u57df\u6062\u590d\u884c\u5217\u7ed3\u6784\u4e0e\u5355\u5143\u683c\u5f52\u5c5e<\/td>\n<td>\u4e0a\u4f4d&#xff1a;\u8868\u683c\u7ed3\u6784\u8bc6\u522b&#xff1b;\u540c\u65cf&#xff1a;PubTables-1M \u6807\u6ce8\u4f53\u7cfb<\/td>\n<td>\u6700\u5bb9\u6613\u9759\u9ed8\u5931\u8d25\u7684\u4e00\u73af&#xff1a;\u4e0d\u8fd8\u539f\u4e5f\u4e0d\u62a5\u9519&#xff0c;\u53ea\u662f\u8f93\u51fa\u6587\u672c<\/td>\n<\/tr>\n<tr>\n<td>VLM \u6574\u9875\u7406\u89e3<\/td>\n<td>\u6574\u9875\u6e32\u67d3\u6210\u56fe\u4ea4\u7ed9\u89c6\u89c9\u8bed\u8a00\u6a21\u578b&#xff0c;\u4e00\u6b21\u8f93\u51fa\u5e26\u7ed3\u6784\u7684 Markdown<\/td>\n<td>\u4e0a\u4f4d&#xff1a;\u591a\u6a21\u6001\u6587\u6863\u7406\u89e3&#xff1b;\u540c\u65cf&#xff1a;Nougat \u7c7b\u7aef\u5230\u7aef\u6a21\u578b<\/td>\n<td>\u4e09\u6863\u91cc\u6700\u8d35\u7684\u4e00\u6863&#xff0c;\u53ea\u5728\u7ed3\u6784\u4fdd\u6301\u7387\u8981\u6c42\u6781\u9ad8\u65f6\u5212\u7b97<\/td>\n<\/tr>\n<tr>\n<td>\u65e0\u611f\u635f\u5931<\/td>\n<td>\u89e3\u6790\u4e22\u4e86\u4fe1\u606f&#xff0c;\u4f46\u8f93\u51fa\u4ecd\u662f\u6b63\u5e38\u6587\u672c\u3001\u4e0d\u89e6\u53d1\u5f02\u5e38<\/td>\n<td>\u4e0e\u300c\u6709\u611f\u635f\u5931\u300d\u5e76\u5217&#xff1b;\u4e0a\u4f4d&#xff1a;\u89e3\u6790\u8d28\u91cf\u5ea6\u91cf<\/td>\n<td>\u56db\u628a\u5c3a\u5b50\u90fd\u662f\u4e3a\u5b83\u8bbe\u8ba1\u7684<\/td>\n<\/tr>\n<tr>\n<td>\u7ed3\u6784\u5316\u5757\u6d41<\/td>\n<td>\u89e3\u6790\u5668\u7684\u4ea7\u7269&#xff1a;\u5e26\u7c7b\u578b\u3001\u9875\u7801\u3001\u7ae0\u8282\u8def\u5f84\u7684\u5757\u5e8f\u5217<\/td>\n<td>\u4e0a\u4f4d&#xff1a;\u77e5\u8bc6\u6761\u76ee&#xff1b;\u4e0b\u4f4d&#xff1a;\u5757\u3001\u8de8\u5ea6<\/td>\n<td>\u672c\u7bc7\u7684\u8f93\u51fa\u7269&#xff0c;\u4e5f\u662f 02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad \u4e0e {{URL:03}} \u7684\u8f93\u5165<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u540e\u9762\u8fd9\u5f20\u4e94\u73af\u8868\u7684\u300c\u68c0\u6d4b\u70b9\u300d\u4e00\u680f&#xff0c;\u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d \u4f1a\u628a\u5b83\u53d8\u6210\u4e94\u6761\u53ef\u4ee5\u7167\u7740\u5199\u7684\u65ad\u88c2\u68c0\u6d4b\u89c4\u5219\u3002<\/p>\n<hr \/>\n<h3>\u4e8c\u3001\u4e94\u73af\u5404\u4e22\u4ec0\u4e48&#xff0c;\u68c0\u6d4b\u70b9\u5e94\u8be5\u88c5\u5728\u54ea&#xff1f;<\/h3>\n<table>\n<tr>\u73af\u8282\u5178\u578b\u635f\u5931\u5f62\u6001\u6709\u65e0\u611f\u68c0\u6d4b\u70b9\u9884\u8b66\u9608\u503c<\/tr>\n<tbody>\n<tr>\n<td>\u89e3\u7801<\/td>\n<td>\u626b\u63cf\u9875\u65e0\u6587\u672c\u5c42 \u2192 \u7a7a\u4e32&#xff1b;\u5b57\u4f53\u7f3a\u6620\u5c04 \u2192 \u4e71\u7801<\/td>\n<td>\u534a\u6709\u611f<\/td>\n<td>\u7a7a\u9875\u7387\u3001\u4e71\u7801\u5b57\u7b26\u5360\u6bd4<\/td>\n<td>\u7a7a\u9875\u7387 &gt;1% \u89e6\u53d1 OCR \u5206\u652f<\/td>\n<\/tr>\n<tr>\n<td>\u7248\u9762\u5206\u6790<\/td>\n<td>\u53cc\u680f\u88ab\u5f53\u5355\u680f&#xff1b;\u9875\u7709\u9875\u811a\u6df7\u8fdb\u6b63\u6587&#xff1b;\u6807\u9898\u5f53\u666e\u901a\u6bb5\u843d<\/td>\n<td>\u65e0\u611f<\/td>\n<td>\u5355\u9875\u5757\u6570\u5f02\u5e38\u3001\u5757\u5bbd\u5206\u5e03\u53cc\u5cf0<\/td>\n<td>\u5757\u5bbd\u53cc\u5cf0\u4e14\u5757\u6570\u7ffb\u500d\u5373\u5224\u5b9a\u5206\u680f<\/td>\n<\/tr>\n<tr>\n<td>\u8868\u683c\u8fd8\u539f<\/td>\n<td>\u62c9\u5e73\u6210\u7a7a\u683c\u5206\u9694\u6587\u672c&#xff1b;\u5408\u5e76\u5355\u5143\u683c\u4e22&#xff1b;\u591a\u7ea7\u8868\u5934\u584c\u6210\u4e00\u884c<\/td>\n<td>\u65e0\u611f<\/td>\n<td>\u8fd8\u539f\u8868\u683c\u6570\u4e0e\u8868\u683c\u533a\u57df\u6570\u4e4b\u6bd4<\/td>\n<td>\u8fd8\u539f\u7387 &lt;0.6 \u89c6\u4e3a\u8be5\u73af\u8282\u5931\u6548<\/td>\n<\/tr>\n<tr>\n<td>\u9605\u8bfb\u987a\u5e8f<\/td>\n<td>\u5de6\u53f3\u680f\u7a7f\u63d2&#xff1b;\u811a\u6ce8\u63d2\u8fdb\u6b63\u6587\u4e2d\u95f4&#xff1b;\u8de8\u9875\u6bb5\u843d\u88ab\u5207\u65ad<\/td>\n<td>\u65e0\u611f<\/td>\n<td>\u76f8\u90bb\u5757\u7ae0\u8282\u8def\u5f84\u8df3\u53d8\u6b21\u6570<\/td>\n<td>\u8df3\u53d8\u7387 &gt;5% \u5c31\u8981\u6362\u6392\u5e8f\u7b56\u7565<\/td>\n<\/tr>\n<tr>\n<td>\u5143\u6570\u636e\u62bd\u53d6<\/td>\n<td>\u6ca1\u6709\u9875\u7801\u3001\u7ae0\u8282\u8def\u5f84\u4e0e\u5c42\u7ea7<\/td>\n<td>\u65e0\u611f<\/td>\n<td>section \u7a7a\u503c\u7387<\/td>\n<td>\u7a7a\u503c\u7387 &gt;10% \u4f1a\u8ba9\u5f15\u7528\u65e0\u6cd5\u5b9a\u4f4d<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u4e94\u6761\u6700\u5bb9\u6613\u8e29\u7684\u5751&#xff0c;\u6309\u4e94\u5b57\u6bb5\u62c6\u5f00&#xff1a;<\/p>\n<table>\n<tr>\u9677\u9631\u89e6\u53d1\u6761\u4ef6\u73b0\u8c61\u6839\u56e0\u68c0\u6d4b\u89c4\u907f<\/tr>\n<tbody>\n<tr>\n<td>\u8868\u683c\u62c9\u5e73<\/td>\n<td>\u6863 A \u62bd\u53d6\u542b\u8868\u9875\u9762<\/td>\n<td>\u5355\u5143\u683c\u503c\u90fd\u5728&#xff0c;\u5217\u540d\u548c\u5f52\u5c5e\u6ca1\u4e86<\/td>\n<td>\u89e3\u6790\u5668\u65e0\u8868\u683c\u7ed3\u6784\u8bc6\u522b\u80fd\u529b<\/td>\n<td>\u7ed3\u6784\u4fdd\u6301\u7387 &lt;0.5<\/td>\n<td>\u5347\u5230\u6863 B&#xff0c;\u6216\u53ea\u5bf9\u8be5\u7c7b\u6587\u6863\u5347\u6863<\/td>\n<\/tr>\n<tr>\n<td>\u8de8\u9875\u8868\u683c\u5207\u65ad<\/td>\n<td>\u4e00\u5f20\u8868\u8de8\u4e24\u9875<\/td>\n<td>\u540e\u534a\u5f20\u8868\u6210\u4e86\u65e0\u8868\u5934\u7684\u5b64\u7acb\u788e\u7247<\/td>\n<td>\u6309\u9875\u5207\u5757&#xff0c;\u672a\u6309\u7248\u9762\u6811\u5408\u5e76<\/td>\n<td>\u8de8\u5757\u5b8c\u6574\u7387&#xff08;\u89c1\u7b2c\u56db\u8282&#xff09;<\/td>\n<td>\u8868\u5934\u7ee7\u627f &#043; \u8de8\u9875\u5408\u5e76&#xff0c;02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad \u5c55\u5f00<\/td>\n<\/tr>\n<tr>\n<td>\u53cc\u680f\u4e32\u680f<\/td>\n<td>\u53cc\u680f\u6392\u7248\u6309 y \u5750\u6807\u6392\u5e8f<\/td>\n<td>\u5de6\u680f\u6bb5\u843d\u4e2d\u95f4\u63d2\u8fdb\u53f3\u680f\u53e5\u5b50<\/td>\n<td>\u9605\u8bfb\u987a\u5e8f\u53ea\u6309\u7eb5\u5411\u5750\u6807\u6392<\/td>\n<td>\u76f8\u90bb\u5757 section \u8df3\u53d8\u8ba1\u6570<\/td>\n<td>\u7248\u9762\u5668\u5f00\u5206\u680f\u68c0\u6d4b&#xff0c;\u6309\u680f\u5206\u7ec4\u6392\u5e8f<\/td>\n<\/tr>\n<tr>\n<td>\u6807\u9898\u5c42\u7ea7\u4e22\u5931<\/td>\n<td>\u6807\u9898\u9760\u52a0\u7c97\u800c\u975e\u6837\u5f0f\u6807\u8bb0<\/td>\n<td>section \u8def\u5f84\u5168\u7a7a&#xff0c;\u5c42\u7ea7\u6811\u9000\u5316\u6210\u5e73\u8868<\/td>\n<td>\u5143\u6570\u636e\u73af\u8282\u6ca1\u53d6\u5b57\u53f7\u4e0e\u52a0\u7c97\u7279\u5f81<\/td>\n<td>section \u7a7a\u503c\u7387<\/td>\n<td>\u7528\u5b57\u53f7\/\u52a0\u7c97\/\u7f16\u53f7\u6b63\u5219\u91cd\u5efa\u5c42\u7ea7<\/td>\n<\/tr>\n<tr>\n<td>\u626b\u63cf\u9875\u9759\u9ed8\u62bd\u7a7a<\/td>\n<td>PDF \u53ea\u6709\u56fe\u5c42\u6ca1\u6709\u6587\u672c\u5c42<\/td>\n<td>\u8be5\u9875\u8fd4\u56de\u7a7a\u4e32&#xff0c;exit code \u4e3a 0<\/td>\n<td>\u89e3\u7801\u73af\u8282\u62bd\u4e0d\u5230\u5c31\u8fd4\u56de\u7a7a&#xff0c;\u4e0d\u629b\u5f02\u5e38<\/td>\n<td>\u7a7a\u9875\u7387<\/td>\n<td>\u7a7a\u9875\u7387\u8d85\u9608\u503c\u8d70 OCR&#xff0c;\u6bcf\u9875\u7559 page_no<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6709\u6587\u672c\u5c42\u5374\u4e0a\u4e86 OCR&#xff0c;\u662f\u8fd9\u91cc\u6700\u8d35\u7684\u4e00\u7c7b\u53cd\u5411\u9519\u8bef&#xff1a;\u5df2\u7ecf\u6b63\u786e\u7684\u5b57\u7b26\u88ab\u91cd\u8bc6\u522b\u4e00\u904d&#xff0c;\u5b57\u6bb5\u53ec\u56de\u7387\u4e0d\u5347\u53cd\u964d\u3002\u5224\u636e\u5f88\u7b80\u5355\u2014\u2014\u62bd\u4e00\u9875\u770b\u5b57\u7b26\u6570\u4e0e\u8be5\u9875\u89c6\u89c9\u5bc6\u5ea6\u662f\u5426\u5339\u914d&#xff0c;\u6b63\u5e38\u5c31\u522b OCR\u3002<\/p>\n<p>\u51e0\u4e2a\u540d\u8bcd\u8865\u9f50&#xff1a;<\/p>\n<table>\n<tr>\u540d\u8bcd\u4e00\u53e5\u8bdd\u539f\u7406\u8fb9\u754c\u4e0e\u5173\u952e\u53c2\u6570<\/tr>\n<tbody>\n<tr>\n<td>\u6587\u672c\u5c42<\/td>\n<td>PDF \u5185\u90e8\u5b58\u653e\u7684\u5b57\u7b26\u4e0e\u5750\u6807\u6570\u636e&#xff0c;\u51b3\u5b9a\u80fd\u4e0d\u80fd\u4e0d OCR \u76f4\u63a5\u53d6\u5b57<\/td>\n<td>\u5224\u636e&#xff1a;\u5355\u9875\u62bd\u53d6\u5b57\u7b26\u6570 &gt;0 \u5373\u6709\u6587\u672c\u5c42&#xff1b;\u626b\u63cf\u4ef6\u6052\u4e3a 0<\/td>\n<\/tr>\n<tr>\n<td>\u7248\u9762\u6811<\/td>\n<td>\u9875\u9762\u533a\u57df\u6309\u5305\u542b\u5173\u7cfb\u7ec4\u7ec7\u6210\u6811&#xff0c;\u8de8\u9875\u5408\u5e76\u9760\u5b83<\/td>\n<td>\u8282\u70b9\u6df1\u5ea6\u901a\u5e38 3\u20134 \u5c42&#xff08;\u9875-\u680f-\u5757-\u884c&#xff09;<\/td>\n<\/tr>\n<tr>\n<td>\u5408\u5e76\u5355\u5143\u683c<\/td>\n<td>\u8de8\u591a\u884c\/\u591a\u5217\u7684\u5355\u5143\u683c&#xff0c;\u62c9\u5e73\u65f6\u6700\u5148\u4e22\u5f52\u5c5e<\/td>\n<td>\u68c0\u6d4b&#xff1a;\u8fd8\u539f\u540e\u5404\u884c\u957f\u5ea6\u4e0d\u9f50\u5373\u89c6\u4e3a\u672a\u5904\u7406<\/td>\n<\/tr>\n<tr>\n<td>OCR \u7f6e\u4fe1\u5ea6<\/td>\n<td>\u8bc6\u522b\u5668\u5bf9\u5355\u5b57\u7b26\u6b63\u786e\u7387\u7684\u4f30\u8ba1&#xff0c;\u5b57\u7b26\u7ea7\u6307\u6807<\/td>\n<td>\u5b83\u9ad8\u4e0d\u4ee3\u8868\u7ed3\u6784\u5bf9&#xff0c;\u522b\u62ff\u5b83\u5f53\u5b57\u6bb5\u53ec\u56de\u7387\u7684\u4ee3\u7406<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<hr \/>\n<h3>\u4e09\u3001\u65e0\u611f\u635f\u5931\u548c\u6709\u611f\u635f\u5931&#xff0c;\u6392\u67e5\u96be\u5ea6\u5dee\u5728\u54ea&#xff1f;<\/h3>\n<table>\n<tr>\u7ef4\u5ea6\u6709\u611f\u635f\u5931\u65e0\u611f\u635f\u5931<\/tr>\n<tbody>\n<tr>\n<td>\u89e6\u53d1<\/td>\n<td>\u5f02\u5e38\u3001\u4e71\u7801\u3001\u7a7a\u9875\u3001\u7f3a\u9875<\/td>\n<td>\u6b63\u5e38\u8f93\u51fa&#xff0c;\u7ed3\u6784\u6d88\u5931<\/td>\n<\/tr>\n<tr>\n<td>\u8c01\u5148\u53d1\u73b0<\/td>\n<td>\u6d41\u6c34\u7ebf\u76d1\u63a7\u3001exit code<\/td>\n<td>\u4e0b\u6e38\u53ec\u56de\u6392\u5e8f\u3001\u4eba\u5de5\u62bd\u67e5<\/td>\n<\/tr>\n<tr>\n<td>\u53d1\u73b0\u65f6\u957f<\/td>\n<td>\u5206\u949f\u7ea7&#xff08;\u8fdb\u65e5\u5fd7\u5c31\u6709&#xff09;<\/td>\n<td>\u5929\u5230\u5468\u7ea7<\/td>\n<\/tr>\n<tr>\n<td>\u80fd\u5426\u81ea\u52a8\u68c0\u6d4b<\/td>\n<td>\u80fd&#xff0c;\u9760\u5f02\u5e38\u8ba1\u6570<\/td>\n<td>\u4e0d\u80fd&#xff0c;\u5fc5\u987b\u6709\u5bf9\u7167\u57fa\u51c6<\/td>\n<\/tr>\n<tr>\n<td>\u4fee\u6cd5\u662f\u5426\u660e\u786e<\/td>\n<td>\u4fee&#xff0c;\u4e14\u6709\u660e\u786e\u62a5\u9519\u4f4d\u7f6e<\/td>\n<td>\u5148\u5b9a\u4f4d\u662f\u54ea\u4e00\u73af&#xff0c;\u518d\u51b3\u5b9a\u5347\u6863\u8fd8\u662f\u6362\u7b56\u7565<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u5dee\u7684\u4e0d\u53ea\u662f\u300c\u96be\u4e0d\u96be\u67e5\u300d&#xff0c;\u66f4\u662f\u6709\u6ca1\u6709\u5bf9\u7167\u7269\u3002\u6709\u611f\u635f\u5931\u7684\u5bf9\u7167\u7269\u662f\u300c\u6b63\u5e38\u5e94\u8be5\u957f\u4ec0\u4e48\u6837\u300d\u2014\u2014\u7a7a\u9875\u3001\u4e71\u7801\u4e00\u773c\u5c31\u770b\u5f97\u51fa\u3002\u65e0\u611f\u635f\u5931\u6ca1\u6709\u8fd9\u4e2a\u5bf9\u7167\u7269&#xff1a;\u4e00\u6bb5\u901a\u987a\u7684\u6587\u672c\u65c1\u8fb9&#xff0c;\u6ca1\u6709\u4efb\u4f55\u300c\u5b83\u539f\u672c\u662f\u5f20\u8868\u300d\u7684\u75d5\u8ff9\u3002<\/p>\n<p>\u6240\u4ee5\u6392\u67e5\u65e0\u611f\u635f\u5931\u6709\u4e24\u6761\u8def&#xff0c;\u4e24\u6761\u90fd\u5f97\u8d70&#xff1a;<\/p>\n<li>\u4eba\u5de5\u5efa\u57fa\u51c6&#xff1a;\u62bd 20 \u7bc7&#xff0c;\u6807\u51fa\u5173\u952e\u5b57\u6bb5\u4e0e\u8868\u683c\u7ed3\u6784&#xff0c;\u62ff\u89e3\u6790\u8f93\u51fa\u6bd4\u5bf9\u3002\u51c6&#xff0c;\u4f46\u53ea\u8986\u76d6\u8fd9 20 \u7bc7\u3002<\/li>\n<li>\u4e0b\u6e38\u53cd\u63a8&#xff1a;\u89e3\u6790\u8d28\u91cf\u53d8\u5dee\u4f1a\u4f20\u5bfc\u5230 opsqa-200&#xff0c;\u4f46\u4f20\u5bfc\u662f\u7a00\u91ca\u7684\u2014\u2014\u7ed3\u6784\u4fdd\u6301\u7387\u6389 10pp&#xff0c;\u4e0b\u6e38\u53ef\u80fd\u53ea\u6389 2\u20133pp &#xff0c;\u5f88\u5bb9\u6613\u88ab\u5f53\u6210\u6a21\u578b\u6296\u52a8\u3002<\/li>\n<p>\u5e38\u89c1\u8bef\u6d4b&#xff1a;<\/p>\n<table>\n<tr>\u8bef\u6d4b\u600e\u4e48\u88ab\u6d4b\u9519\u6b63\u786e\u6d4b\u6cd5<\/tr>\n<tbody>\n<tr>\n<td>\u7528\u5b57\u7b26\u6570\/\u9875\u6570\u8861\u91cf\u89e3\u6790\u8d28\u91cf<\/td>\n<td>\u7ed3\u6784\u4e22\u5931\u51e0\u4e4e\u4e0d\u6539\u53d8\u5b57\u7b26\u6570&#xff0c;\u62c9\u5e73\u524d\u540e\u53ef\u80fd\u53ea\u5dee\u51e0\u4e2a\u7a7a\u683c<\/td>\n<td>\u7528\u5b57\u6bb5\u53ec\u56de\u7387\u4e0e\u7ed3\u6784\u4fdd\u6301\u7387<\/td>\n<\/tr>\n<tr>\n<td>\u53ea\u770b\u89e3\u6790\u5668\u65e5\u5fd7\u91cc\u7684\u62a5\u9519<\/td>\n<td>\u65e0\u611f\u635f\u5931\u4e0d\u8fdb\u65e5\u5fd7&#xff0c;\u65e5\u5fd7\u5e72\u51c0\u4e0d\u7b49\u4e8e\u6ca1\u4e22<\/td>\n<td>\u62bd 20 \u7bc7\u4eba\u5de5\u6bd4\u5bf9&#xff0c;\u6216\u76f4\u63a5\u8dd1\u56db\u628a\u5c3a\u5b50<\/td>\n<\/tr>\n<tr>\n<td>\u62ff OCR \u7f6e\u4fe1\u5ea6\u5f53\u5b57\u6bb5\u53ec\u56de\u7387<\/td>\n<td>\u7f6e\u4fe1\u5ea6\u662f\u5b57\u7b26\u7ea7&#xff0c;\u7ed3\u6784\u662f\u5757\u7ea7&#xff0c;\u4e24\u8005\u4e0d\u4e92\u76f8\u8986\u76d6<\/td>\n<td>\u626b\u63cf\u4ef6\u770b\u7f6e\u4fe1\u5ea6&#xff0c;\u5176\u4f59\u4e00\u5f8b\u770b\u5c3a\u5b50<\/td>\n<\/tr>\n<tr>\n<td>\u7528\u7f51\u4e0a\u4e0b\u8f7d\u7684\u516c\u5f00 PDF \u5efa\u8bc4\u6d4b\u96c6<\/td>\n<td>\u7248\u5f0f\u5206\u5e03\u4e0e\u81ea\u5df1\u7684\u6587\u6863\u5dee\u5f97\u8fdc&#xff0c;\u7ed3\u8bba\u4e0d\u53ef\u8fc1\u79fb<\/td>\n<td>\u7528\u81ea\u5df1\u7684 20 \u7bc7&#xff0c;\u7248\u5f0f\u5206\u5e03\u8981\u8986\u76d6\u5168<\/td>\n<\/tr>\n<tr>\n<td>\u8dd1\u4e00\u6b21\u5c31\u6362\u89e3\u6790\u65b9\u6848<\/td>\n<td>\u5355\u7bc7\u6ce2\u52a8\u5927&#xff0c;0.62 \u4e0e 0.68 \u53ef\u80fd\u53ea\u662f\u91c7\u6837\u566a\u58f0<\/td>\n<td>bootstrap \u7f6e\u4fe1\u533a\u95f4\u4e0d\u8de8\u8fc7\u57fa\u7ebf\u518d\u6362<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u5143\u6570\u636e\u8fd9\u4e00\u73af\u4e22\u7684\u9875\u7801\u4e0e\u7ae0\u8282\u8def\u5f84&#xff0c;\u300c[06-\u5f15\u7528\u4e0e\u53ef\u6eaf\u6e90 \u8fd8\u5728\u5199]\u300d\u505a\u5f15\u7528\u9a8c\u8bc1\u65f6\u8981\u62ff\u5b83\u5b9a\u4f4d\u8bc1\u636e\u2014\u2014\u90a3\u65f6\u5019\u624d\u53d1\u73b0 section \u5b57\u6bb5\u5168\u7a7a&#xff0c;\u8fd4\u5de5\u7684\u662f\u6574\u6761\u7d22\u5f15\u3002<\/p>\n<hr \/>\n<h3>\u56db\u3001\u56db\u628a\u5c3a\u5b50\u600e\u4e48\u6d4b&#xff0c;20 \u7bc7\u591f\u4e0d\u591f&#xff1f;<\/h3>\n<table>\n<tr>\u5c3a\u5b50\u5b9a\u4e49\u600e\u4e48\u6d4b\u9608\u503c\u521d\u503c<\/tr>\n<tbody>\n<tr>\n<td>\u5b57\u6bb5\u53ec\u56de\u7387<\/td>\n<td>\u4eba\u5de5\u6807\u7684 N \u4e2a\u5173\u952e\u5b57\u6bb5\u91cc&#xff0c;\u80fd\u5728\u8f93\u51fa\u4e2d\u88ab\u547d\u4e2d&#xff08;\u542b\u6a21\u7cca\u5339\u914d&#xff09;\u7684\u6bd4\u4f8b<\/td>\n<td>\u6bcf\u7bc7\u6807 10 \u4e2a\u5b57\u6bb5&#xff0c;20 \u7bc7\u5171 200 \u4e2a<\/td>\n<td>\u22650.90<\/td>\n<\/tr>\n<tr>\n<td>\u7ed3\u6784\u4fdd\u6301\u7387<\/td>\n<td>\u8868\u683c\u884c\u5217\u8fd8\u539f\u6b63\u786e\u7387\u4e0e\u6807\u9898\u5c42\u7ea7\u6062\u590d\u6b63\u786e\u7387\u7684\u7b49\u6743\u5e73\u5747<\/td>\n<td>\u6807\u6bcf\u7bc7\u7684\u8868\u683c\u6570\u4e0e\u6807\u9898\u5c42\u7ea7\u6811<\/td>\n<td>\u22650.80<\/td>\n<\/tr>\n<tr>\n<td>\u8de8\u5757\u5b8c\u6574\u7387<\/td>\n<td>1 \u2212 \u8de8\u5757\u65ad\u88c2\u7387&#xff0c;\u8bed\u4e49\u5355\u5143\u6ca1\u88ab\u5207\u65ad\u7684\u6bd4\u4f8b<\/td>\n<td>\u76f8\u90bb\u5757\u7684\u7ae0\u8282\u4e0e\u9875\u7801\u8fde\u7eed\u6027\u5224\u5b9a<\/td>\n<td>\u22650.85<\/td>\n<\/tr>\n<tr>\n<td>\u5355\u4f4d\u6210\u672c<\/td>\n<td>\u6bcf\u9875\u8017\u65f6\u4e0e\u6bcf\u9875\u91d1\u989d<\/td>\n<td>\u8ba1\u65f6 &#043; token \u8ba1\u8d39<\/td>\n<td>\u6309\u9884\u7b97\u5b9a&#xff0c;\u5148\u7b97\u6bcf\u5343\u9875\u7f8e\u5143<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>20 \u7bc7\u591f\u4e0d\u591f\u53ef\u4ee5\u7b97&#xff1a;200 \u4e2a\u5b57\u6bb5\u5728\u53ec\u56de\u7387 0.5 \u9644\u8fd1\u65f6&#xff0c;\u6807\u51c6\u8bef sqrt(0.25\/200) \u2248 0.035&#xff0c;\u5373 \u00b13.5pp [\u7b97]\u3002\u8fd9\u4e2a\u7cbe\u5ea6\u591f\u7528\u6765\u5224\u65ad\u300c\u6863 A \u548c\u6863 B \u5dee\u591a\u5c11\u300d&#xff0c;\u4e0d\u591f\u5224\u65ad\u300c\u6863 B \u548c\u6863 C \u8c01\u66f4\u597d\u300d\u2014\u2014\u540e\u8005\u7684\u5dee\u8ddd\u5e38\u5728 2pp \u4ee5\u5185\u3002\u8981\u628a\u534a\u5bbd\u538b\u5230 2pp&#xff0c;\u9700\u8981 0.25\/0.0004 &#061; 625 \u4e2a\u5b57\u6bb5&#xff0c;\u7ea6 63 \u7bc7 [\u7b97]\u3002<\/p>\n<p>\u4e0b\u9762\u8fd9\u4efd\u5185\u6838\u628a\u56db\u628a\u5c3a\u5b50\u8dd1\u51fa\u6765\u3002\u96f6\u4f9d\u8d56&#xff0c;\u53ea\u5403\u6807\u51c6\u5e93&#xff1b;embed \u4e0e llm \u7559\u6210\u53ef\u66ff\u6362\u9002\u914d\u53e3&#xff0c;\u672c\u7bc7\u53ea\u7528\u5f97\u5230 embed \u7684\u964d\u7ea7\u7248\u672c\u3002<\/p>\n<p>\u8f93\u5165\u662f\u4ec0\u4e48&#xff1a;gold\/20docs.json&#xff08;\u6bcf\u7bc7\u7684 n_pages \/ fields \/ tables \/ seconds \/ usd&#xff09;\u4e0e parsed\/blocks.jsonl&#xff08;\u89e3\u6790\u5668\u8f93\u51fa\u7684\u5757\u6d41&#xff09;\u3002 \u600e\u4e48\u6267\u884c&#xff1a;\u9010\u7bc7\u6bd4\u5bf9 gold \u6807\u6ce8\u4e0e\u89e3\u6790\u5757\u6d41&#xff0c;\u56db\u628a\u5c3a\u5b50\u5404\u7b97\u4e00\u4e2a\u6570&#xff0c;\u8de8\u7bc7\u53d6\u5747\u503c\u3002 \u8f93\u51fa\u4ec0\u4e48&#xff1a;\u6bcf\u7bc7\u4e00\u884c JSON&#xff0c;\u672b\u5c3e\u6253\u5370\u4e09\u4e2a\u5747\u503c\u3002 \u7ed3\u679c\u4e0d\u5bf9\u67e5\u54ea&#xff1a;\u5b57\u6bb5\u53ec\u56de\u7387\u5f02\u5e38\u4f4e\u5148\u770b\u662f\u4e0d\u662f gold \u5b57\u6bb5\u91cc\u5e26\u4e86\u6362\u884c\u7b26&#xff1b;\u7ed3\u6784\u4fdd\u6301\u7387\u4f4e\u5148\u770b blocks.jsonl \u91cc\u6709\u6ca1\u6709 table \u7c7b\u578b\u7684\u5757\u2014\u2014\u4e00\u6761\u90fd\u6ca1\u6709\u8bf4\u660e\u89e3\u6790\u5668\u538b\u6839\u6ca1\u505a\u8868\u683c\u8fd8\u539f\u3002<\/p>\n<p><span class=\"token comment\"># parse_quality.py \u2014\u2014 \u89e3\u6790\u8d28\u91cf\u56db\u628a\u5c3a\u5b50&#xff08;\u96f6\u4f9d\u8d56\u5185\u6838&#xff0c;\u6807\u51c6\u5e93&#xff09;<\/span><br \/>\n<span class=\"token keyword\">import<\/span> json<span class=\"token punctuation\">,<\/span> re<span class=\"token punctuation\">,<\/span> difflib<br \/>\n<span class=\"token keyword\">from<\/span> dataclasses <span class=\"token keyword\">import<\/span> dataclass<span class=\"token punctuation\">,<\/span> field<\/p>\n<p><span class=\"token comment\"># &#8212;- \u53ef\u66ff\u6362\u9002\u914d\u53e3 &#8212;-<\/span><br \/>\n<span class=\"token keyword\">def<\/span> <span class=\"token function\">embed<\/span><span class=\"token punctuation\">(<\/span>text<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u7a20\u5bc6\u5411\u91cf\u53e3\u3002\u672a\u63a5\u5165\u65f6\u4e0b\u9762\u4e0d\u7528\u5b83&#xff0c;\u4ec5\u4fdd\u7559\u7b7e\u540d\u3002&#034;&#034;&#034;<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token builtin\">float<\/span><span class=\"token punctuation\">(<\/span><span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>text<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">(<\/span><span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span><span class=\"token builtin\">set<\/span><span class=\"token punctuation\">(<\/span>text<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><\/p>\n<p><span class=\"token keyword\">async<\/span> <span class=\"token keyword\">def<\/span> <span class=\"token function\">llm<\/span><span class=\"token punctuation\">(<\/span>prompt<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">str<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u7ed3\u6784\u5224\u5b9a\u515c\u5e95&#xff1a;\u628a\u96be\u5224\u7684\u5757\u8fd8\u539f\u6210\u8868\u683c\/\u6807\u9898\u6811\u3002\u672c\u7bc7\u4e0d\u8c03\u7528\u3002&#034;&#034;&#034;<\/span><br \/>\n    <span class=\"token keyword\">raise<\/span> NotImplementedError<\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">sim<\/span><span class=\"token punctuation\">(<\/span>a<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><span class=\"token punctuation\">,<\/span> b<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u5b57\u6bb5\u6a21\u7cca\u5339\u914d&#xff1a;\u7cbe\u786e\u76f8\u7b49 1.0&#xff0c;\u5305\u542b 0.95&#xff0c;\u5426\u5219\u7528\u5e8f\u5217\u6bd4\u5bf9\u515c OCR \u7ea7\u5dee\u5f02&#034;&#034;&#034;<\/span><br \/>\n    a<span class=\"token punctuation\">,<\/span> b <span class=\"token operator\">&#061;<\/span> a<span class=\"token punctuation\">.<\/span>strip<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> b<span class=\"token punctuation\">.<\/span>strip<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">if<\/span> a <span class=\"token operator\">&#061;&#061;<\/span> b<span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">return<\/span> <span class=\"token number\">1.0<\/span><br \/>\n    <span class=\"token keyword\">if<\/span> a <span class=\"token keyword\">and<\/span> a <span class=\"token keyword\">in<\/span> b<span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">return<\/span> <span class=\"token number\">0.95<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> difflib<span class=\"token punctuation\">.<\/span>SequenceMatcher<span class=\"token punctuation\">(<\/span><span class=\"token boolean\">None<\/span><span class=\"token punctuation\">,<\/span> a<span class=\"token punctuation\">,<\/span> b<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">.<\/span>ratio<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><\/p>\n<p><span class=\"token decorator annotation punctuation\">&#064;dataclass<\/span><br \/>\n<span class=\"token keyword\">class<\/span> <span class=\"token class-name\">Block<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    doc_id<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><br \/>\n    btype<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span>                     <span class=\"token comment\"># text \/ table \/ title \/ caption<\/span><br \/>\n    text<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><br \/>\n    page<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">int<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><br \/>\n    section<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token string\">&#034;&#034;<\/span>              <span class=\"token comment\"># \u7ae0\u8282\u8def\u5f84&#xff0c;\u5982 &#034;3.2.1&#034;<\/span><br \/>\n    rows<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span> <span class=\"token operator\">&#061;<\/span> field<span class=\"token punctuation\">(<\/span>default_factory<span class=\"token operator\">&#061;<\/span><span class=\"token builtin\">list<\/span><span class=\"token punctuation\">)<\/span>   <span class=\"token comment\"># \u8868\u683c\u8fd8\u539f\u540e\u7684\u4e8c\u7ef4\u7ed3\u6784<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">field_recall<\/span><span class=\"token punctuation\">(<\/span>gold_fields<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">,<\/span> blocks<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">,<\/span> thr<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">float<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0.90<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u5c3a\u5b501&#xff1a;gold \u5b57\u6bb5\u80fd\u5728\u4efb\u610f\u5757\u91cc\u88ab\u6a21\u7cca\u547d\u4e2d\u7684\u6bd4\u4f8b\u3002<br \/>\n    \u957f\u5b57\u6bb5\u6574\u4f53\u6bd4\u5bf9\u4f1a\u88ab\u5468\u8fb9\u6587\u672c\u7a00\u91ca&#xff0c;\u6545\u9000\u5316\u4e3a\u5b9a\u957f\u6ed1\u7a97\u518d\u6bd4\u4e00\u6b21\u3002&#034;&#034;&#034;<\/span><br \/>\n    corpus <span class=\"token operator\">&#061;<\/span> <span class=\"token string\">&#034;\\\\n&#034;<\/span><span class=\"token punctuation\">.<\/span>join<span class=\"token punctuation\">(<\/span>b<span class=\"token punctuation\">.<\/span>text <span class=\"token keyword\">for<\/span> b <span class=\"token keyword\">in<\/span> blocks<span class=\"token punctuation\">)<\/span><br \/>\n    hit <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> g <span class=\"token keyword\">in<\/span> gold_fields<span class=\"token punctuation\">:<\/span><br \/>\n        best <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">(<\/span>sim<span class=\"token punctuation\">(<\/span>g<span class=\"token punctuation\">,<\/span> b<span class=\"token punctuation\">.<\/span>text<span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">for<\/span> b <span class=\"token keyword\">in<\/span> blocks<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> default<span class=\"token operator\">&#061;<\/span><span class=\"token number\">0.0<\/span><span class=\"token punctuation\">)<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> best <span class=\"token operator\">&lt;<\/span> thr <span class=\"token keyword\">and<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>g<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&gt;&#061;<\/span> <span class=\"token number\">8<\/span><span class=\"token punctuation\">:<\/span><br \/>\n            step <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">8<\/span><br \/>\n            win <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>g<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#043;<\/span> <span class=\"token number\">4<\/span><br \/>\n            best <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span>best<span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">(<\/span>sim<span class=\"token punctuation\">(<\/span>g<span class=\"token punctuation\">,<\/span> corpus<span class=\"token punctuation\">[<\/span>i<span class=\"token punctuation\">:<\/span>i <span class=\"token operator\">&#043;<\/span> win<span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><br \/>\n                                  <span class=\"token keyword\">for<\/span> i <span class=\"token keyword\">in<\/span> <span class=\"token builtin\">range<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">0<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>corpus<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span> win<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> step<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> default<span class=\"token operator\">&#061;<\/span><span class=\"token number\">0.0<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><br \/>\n        hit <span class=\"token operator\">&#043;&#061;<\/span> <span class=\"token number\">1<\/span> <span class=\"token keyword\">if<\/span> best <span class=\"token operator\">&gt;&#061;<\/span> thr <span class=\"token keyword\">else<\/span> <span class=\"token number\">0<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> hit <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>gold_fields<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">structure_keep<\/span><span class=\"token punctuation\">(<\/span>gold_tables<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">,<\/span> blocks<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u5c3a\u5b502&#xff1a;\u8868\u683c\u884c\u5217\u5bf9\u5f97\u4e0a\u7684\u6bd4\u4f8b\u4e0e\u6807\u9898\u5c42\u7ea7\u5bf9\u5f97\u4e0a\u7684\u6bd4\u4f8b&#xff0c;\u7b49\u6743\u5e73\u5747\u3002&#034;&#034;&#034;<\/span><br \/>\n    tabs <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>b <span class=\"token keyword\">for<\/span> b <span class=\"token keyword\">in<\/span> blocks <span class=\"token keyword\">if<\/span> b<span class=\"token punctuation\">.<\/span>btype <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token string\">&#034;table&#034;<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    ok_t <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> gt <span class=\"token keyword\">in<\/span> gold_tables<span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">for<\/span> t <span class=\"token keyword\">in<\/span> tabs<span class=\"token punctuation\">:<\/span><br \/>\n            <span class=\"token keyword\">if<\/span> gt<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;anchor&#034;<\/span><span class=\"token punctuation\">]<\/span> <span class=\"token keyword\">in<\/span> t<span class=\"token punctuation\">.<\/span>text <span class=\"token keyword\">and<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>t<span class=\"token punctuation\">.<\/span>rows<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#061;&#061;<\/span> gt<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;n_rows&#034;<\/span><span class=\"token punctuation\">]<\/span> \\\\<br \/>\n               <span class=\"token keyword\">and<\/span> <span class=\"token builtin\">all<\/span><span class=\"token punctuation\">(<\/span><span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>r<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#061;&#061;<\/span> gt<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;n_cols&#034;<\/span><span class=\"token punctuation\">]<\/span> <span class=\"token keyword\">for<\/span> r <span class=\"token keyword\">in<\/span> t<span class=\"token punctuation\">.<\/span>rows<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n                ok_t <span class=\"token operator\">&#043;&#061;<\/span> <span class=\"token number\">1<\/span><br \/>\n                <span class=\"token keyword\">break<\/span><br \/>\n    titles <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>b <span class=\"token keyword\">for<\/span> b <span class=\"token keyword\">in<\/span> blocks <span class=\"token keyword\">if<\/span> b<span class=\"token punctuation\">.<\/span>btype <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token string\">&#034;title&#034;<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    ok_h <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">sum<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span> <span class=\"token keyword\">for<\/span> t <span class=\"token keyword\">in<\/span> titles <span class=\"token keyword\">if<\/span> re<span class=\"token punctuation\">.<\/span><span class=\"token keyword\">match<\/span><span class=\"token punctuation\">(<\/span><span class=\"token string\">r&#034;^\\\\d&#043;(\\\\.\\\\d&#043;)*\\\\s&#034;<\/span><span class=\"token punctuation\">,<\/span> t<span class=\"token punctuation\">.<\/span>text<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token number\">0.5<\/span> <span class=\"token operator\">*<\/span> ok_t <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>gold_tables<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#043;<\/span> <span class=\"token number\">0.5<\/span> <span class=\"token operator\">*<\/span> ok_h <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>titles<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">cross_block_integrity<\/span><span class=\"token punctuation\">(<\/span>blocks<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u5c3a\u5b503&#xff1a;1 &#8211; \u8de8\u5757\u65ad\u88c2\u7387\u3002\u540c\u7ae0\u8282\u4e14\u9875\u7801\u76f8\u90bb\u7684\u4e24\u5757\u4e4b\u95f4&#xff0c;<br \/>\n    \u524d\u4e00\u5757\u6ca1\u4ee5\u53e5\u672b\u6807\u70b9\u6536\u5c3e&#xff0c;\u5c31\u8bb0\u4e3a\u4e00\u6b21\u5207\u65ad\u3002&#034;&#034;&#034;<\/span><br \/>\n    ends <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;\u3002&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;&#xff1b;&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;&#xff1a;&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;.&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;.&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;?&#034;<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    cuts <span class=\"token operator\">&#061;<\/span> pairs <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> a<span class=\"token punctuation\">,<\/span> b <span class=\"token keyword\">in<\/span> <span class=\"token builtin\">zip<\/span><span class=\"token punctuation\">(<\/span>blocks<span class=\"token punctuation\">,<\/span> blocks<span class=\"token punctuation\">[<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">:<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> a<span class=\"token punctuation\">.<\/span>section <span class=\"token operator\">!&#061;<\/span> b<span class=\"token punctuation\">.<\/span>section <span class=\"token keyword\">or<\/span> b<span class=\"token punctuation\">.<\/span>page <span class=\"token operator\">&#8211;<\/span> a<span class=\"token punctuation\">.<\/span>page <span class=\"token keyword\">not<\/span> <span class=\"token keyword\">in<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token number\">0<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">1<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n            <span class=\"token keyword\">continue<\/span><br \/>\n        pairs <span class=\"token operator\">&#043;&#061;<\/span> <span class=\"token number\">1<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> <span class=\"token keyword\">not<\/span> a<span class=\"token punctuation\">.<\/span>text<span class=\"token punctuation\">.<\/span>rstrip<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">.<\/span>endswith<span class=\"token punctuation\">(<\/span>ends<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n            cuts <span class=\"token operator\">&#043;&#061;<\/span> <span class=\"token number\">1<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token number\">1.0<\/span> <span class=\"token operator\">&#8211;<\/span> cuts <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> pairs<span class=\"token punctuation\">)<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">unit_cost<\/span><span class=\"token punctuation\">(<\/span>n_pages<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">int<\/span><span class=\"token punctuation\">,<\/span> seconds<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">,<\/span> usd<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">float<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u5c3a\u5b504&#xff1a;\u9875\u5747\u8017\u65f6\u3001\u9875\u5747\u91d1\u989d\u3001\u6bcf\u5343\u9875\u91d1\u989d\u3002&#034;&#034;&#034;<\/span><br \/>\n    n <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> n_pages<span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string\">&#034;sec_per_page&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>seconds <span class=\"token operator\">\/<\/span> n<span class=\"token punctuation\">,<\/span> <span class=\"token number\">3<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;usd_per_page&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>usd <span class=\"token operator\">\/<\/span> n<span class=\"token punctuation\">,<\/span> <span class=\"token number\">5<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;usd_per_1k_pages&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>usd <span class=\"token operator\">\/<\/span> n <span class=\"token operator\">*<\/span> <span class=\"token number\">1000<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">2<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">}<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">report<\/span><span class=\"token punctuation\">(<\/span>gold<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">,<\/span> blocks<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string\">&#034;doc_id&#034;<\/span><span class=\"token punctuation\">:<\/span> gold<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;doc_id&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;field_recall&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>field_recall<span class=\"token punctuation\">(<\/span>gold<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;fields&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span> blocks<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">3<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;structure_keep&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>structure_keep<span class=\"token punctuation\">(<\/span>gold<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;tables&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span> blocks<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">3<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;cross_block&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>cross_block_integrity<span class=\"token punctuation\">(<\/span>blocks<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">3<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;cost&#034;<\/span><span class=\"token punctuation\">:<\/span> unit_cost<span class=\"token punctuation\">(<\/span>gold<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;n_pages&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span> gold<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;seconds&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.0<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> gold<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;usd&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.0<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">}<\/span><\/p>\n<p><span class=\"token keyword\">if<\/span> __name__ <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token string\">&#034;__main__&#034;<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    gold <span class=\"token operator\">&#061;<\/span> json<span class=\"token punctuation\">.<\/span>load<span class=\"token punctuation\">(<\/span><span class=\"token builtin\">open<\/span><span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;gold\/20docs.json&#034;<\/span><span class=\"token punctuation\">,<\/span> encoding<span class=\"token operator\">&#061;<\/span><span class=\"token string\">&#034;utf-8&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    parsed <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>Block<span class=\"token punctuation\">(<\/span><span class=\"token operator\">**<\/span>json<span class=\"token punctuation\">.<\/span>loads<span class=\"token punctuation\">(<\/span>l<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">for<\/span> l <span class=\"token keyword\">in<\/span><br \/>\n              <span class=\"token builtin\">open<\/span><span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;parsed\/blocks.jsonl&#034;<\/span><span class=\"token punctuation\">,<\/span> encoding<span class=\"token operator\">&#061;<\/span><span class=\"token string\">&#034;utf-8&#034;<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">if<\/span> l<span class=\"token punctuation\">.<\/span>strip<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    by <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token punctuation\">}<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> b <span class=\"token keyword\">in<\/span> parsed<span class=\"token punctuation\">:<\/span><br \/>\n        by<span class=\"token punctuation\">.<\/span>setdefault<span class=\"token punctuation\">(<\/span>b<span class=\"token punctuation\">.<\/span>doc_id<span class=\"token punctuation\">,<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">.<\/span>append<span class=\"token punctuation\">(<\/span>b<span class=\"token punctuation\">)<\/span><br \/>\n    rows <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>report<span class=\"token punctuation\">(<\/span>g<span class=\"token punctuation\">,<\/span> by<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span>g<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;doc_id&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">for<\/span> g <span class=\"token keyword\">in<\/span> gold<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;docs&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> r <span class=\"token keyword\">in<\/span> rows<span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">print<\/span><span class=\"token punctuation\">(<\/span>json<span class=\"token punctuation\">.<\/span>dumps<span class=\"token punctuation\">(<\/span>r<span class=\"token punctuation\">,<\/span> ensure_ascii<span class=\"token operator\">&#061;<\/span><span class=\"token boolean\">False<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> k <span class=\"token keyword\">in<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;field_recall&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;structure_keep&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;cross_block&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">print<\/span><span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;%s %.3f&#034;<\/span> <span class=\"token operator\">%<\/span> <span class=\"token punctuation\">(<\/span>k<span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">sum<\/span><span class=\"token punctuation\">(<\/span>r<span class=\"token punctuation\">[<\/span>k<span class=\"token punctuation\">]<\/span> <span class=\"token keyword\">for<\/span> r <span class=\"token keyword\">in<\/span> rows<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>rows<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><\/p>\n<h4>\u9010\u6bb5\u8bfb\u8fd9\u4efd\u811a\u672c<\/h4>\n<ul>\n<li>sim()&#xff1a;\u4e09\u7ea7\u5339\u914d\u2014\u2014\u7cbe\u786e\u3001\u5305\u542b\u3001\u5e8f\u5217\u6bd4\u5bf9\u3002\u8bbe\u6210\u8fd9\u6837\u662f\u56e0\u4e3a OCR \u548c\u5b57\u4f53\u6620\u5c04\u9519\u8bef\u901a\u5e38\u53ea\u6539\u4e00\u4e24\u4e2a\u5b57\u7b26&#xff0c;\u7528\u4e25\u683c\u76f8\u7b49\u4f1a\u628a\u300c&gt;85%\u300d\u548c\u300c&gt; 85%\u300d\u5224\u6210\u4e24\u4e2a\u5b57\u6bb5&#xff0c;\u53ec\u56de\u7387\u88ab\u7cfb\u7edf\u6027\u4f4e\u4f30\u30020.95 \u8fd9\u4e2a\u5305\u542b\u5206\u662f\u62cd\u7684&#xff0c;\u53ef\u4ee5\u6309\u5b57\u6bb5\u957f\u5ea6\u91cd\u6807\u3002<\/li>\n<li>field_recall()&#xff1a;\u5148\u6574\u5757\u6bd4&#xff0c;\u4e0d\u884c\u518d\u6ed1\u7a97\u3002\u6ed1\u7a97\u6b65\u957f 8 \u662f\u7cbe\u5ea6\u4e0e\u8017\u65f6\u7684\u6298\u4e2d\u2014\u2014\u6b65\u957f 1 \u65f6 20 \u7bc7\u7684\u8017\u65f6\u4f1a\u957f\u4e00\u4e2a\u91cf\u7ea7&#xff0c;\u6536\u76ca\u53ea\u6709\u96f6\u70b9\u51e0\u4e2a\u767e\u5206\u70b9\u3002\u9608\u503c 0.90 \u5bf9\u5e94\u300c\u5141\u8bb8\u4e00\u4e24\u4e2a\u5b57\u7b26\u7684\u5dee\u5f02\u300d\u3002<\/li>\n<li>structure_keep()&#xff1a;\u8868\u683c\u4e0e\u6807\u9898\u5404\u5360\u4e00\u534a\u3002\u8fd9\u4e2a\u7b49\u6743\u662f\u4eba\u4e3a\u7684&#xff0c;\u5982\u679c\u4f60\u7684\u6587\u6863\u91cc\u8868\u683c\u662f\u4e3b\u4f53&#xff08;\u6bd4\u5982\u8d22\u62a5&#xff09;&#xff0c;\u628a\u8868\u683c\u90a3\u4e00\u9879\u7684\u6743\u91cd\u63d0\u5230 0.7 \u66f4\u5408\u7406\u3002<\/li>\n<li>cross_block_integrity()&#xff1a;\u7528\u300c\u524d\u4e00\u5757\u662f\u5426\u4ee5\u53e5\u672b\u6807\u70b9\u6536\u5c3e\u300d\u5f53\u5207\u65ad\u5224\u636e\u3002\u8fd9\u662f\u4e2a\u5f31\u5224\u636e&#xff0c;\u6293\u4e0d\u5230\u300c\u53e5\u5b50\u6070\u597d\u5728\u6bb5\u672b\u7ed3\u675f\u4f46\u4ecd\u88ab\u5207\u65ad\u300d\u7684\u60c5\u51b5&#xff0c;\u5224\u636e\u672c\u8eab\u53ef\u4ee5\u6309\u4f60\u7684\u6587\u6863\u7c7b\u578b\u6362\u3002<\/li>\n<li>unit_cost()&#xff1a;\u53ea\u7b97\u89e3\u6790\u8fd9\u4e00\u6bb5\u7684\u94b1&#xff0c;\u4e0d\u542b\u7d22\u5f15\u4e0e\u63a8\u7406\u3002\u8de8\u6863\u6bd4\u8f83\u65f6\u53e3\u5f84\u5fc5\u987b\u4e00\u81f4&#xff0c;\u5426\u5219 C \u6863\u7684 token \u6210\u672c\u4f1a\u88ab\u6f0f\u7b97\u3002<\/li>\n<li>\u5df2\u77e5\u504f\u5dee\u4e24\u6761&#xff1a;\u2460 \u53e5\u672b\u6807\u70b9\u5224\u636e\u5bf9\u4e2d\u6587\u5168\u89d2\u6807\u70b9\u654f\u611f&#xff0c;\u6df7\u6392\u6587\u6863\u8981\u6269\u5145 ends&#xff1b;\u2461 structure_keep \u91cc anchor \u7528\u5b50\u4e32\u5339\u914d\u5b9a\u4f4d\u8868\u683c&#xff0c;\u540c\u9875\u591a\u5f20\u7ed3\u6784\u76f8\u540c\u7684\u8868\u4f1a\u4e32\u4f4d&#xff0c;\u9700\u8981\u989d\u5916\u7528\u5750\u6807\u7ea6\u675f\u3002<\/li>\n<\/ul>\n<p>\u8dd1\u5728 20 \u7bc7 runbooks \u62bd\u6837\u4e0a&#xff0c;\u9884\u671f\u8f93\u51fa\u5982\u4e0b&#xff1a;<\/p>\n<p>{&#034;doc_id&#034;:&#034;RB-214&#034;,&#034;field_recall&#034;:0.62,&#034;structure_keep&#034;:0.35,&#034;cross_block&#034;:0.71,<br \/>\n &#034;cost&#034;:{&#034;sec_per_page&#034;:0.041,&#034;usd_per_page&#034;:0.00002,&#034;usd_per_1k_pages&#034;:0.02}}<br \/>\n{&#034;doc_id&#034;:&#034;RB-077&#034;,&#034;field_recall&#034;:0.70,&#034;structure_keep&#034;:0.42,&#034;cross_block&#034;:0.68, &#8230;}<br \/>\n{&#034;doc_id&#034;:&#034;TOPO-1131&#034;,&#034;field_recall&#034;:0.55,&#034;structure_keep&#034;:0.28,&#034;cross_block&#034;:0.75, &#8230;}<br \/>\n&#8230;<br \/>\nfield_recall 0.586<br \/>\nstructure_keep 0.311<br \/>\ncross_block   0.742<\/p>\n<p>\u4e09\u4e2a\u5747\u503c\u91cc structure_keep 0.31 \u8fd9\u4e00\u4e2a\u6570\u5c31\u591f\u4e0b\u5224\u65ad&#xff1a;\u6863 A \u5bf9\u542b\u8868\u6587\u6863\u57fa\u672c\u6ca1\u6709\u7ed3\u6784\u80fd\u529b\u3002<\/p>\n<h4>20 \u7bc7\u8bc4\u6d4b\u96c6\u600e\u4e48\u642d<\/h4>\n<p><span class=\"token punctuation\">{<\/span><br \/>\n  <span class=\"token string-property property\">&#034;sample_rule&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u6309\u7248\u5f0f\u5206\u5c42\u62bd\u6837&#xff0c;\u4e0d\u6309\u6587\u4ef6\u5927\u5c0f\u62bd\u6837&#034;<\/span><span class=\"token punctuation\">,<\/span><br \/>\n  <span class=\"token string-property property\">&#034;strata&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token punctuation\">[<\/span><br \/>\n    <span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;name&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u7eaf\u6587\u672c\u6b63\u6587&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;n&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">5<\/span><span class=\"token punctuation\">}<\/span><span class=\"token punctuation\">,<\/span><br \/>\n    <span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;name&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u542b\u7b80\u5355\u8868\u683c&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;n&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">6<\/span><span class=\"token punctuation\">}<\/span><span class=\"token punctuation\">,<\/span><br \/>\n    <span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;name&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u542b\u5408\u5e76\u5355\u5143\u683c\/\u591a\u7ea7\u8868\u5934&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;n&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">4<\/span><span class=\"token punctuation\">}<\/span><span class=\"token punctuation\">,<\/span><br \/>\n    <span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;name&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u53cc\u680f\u6392\u7248&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;n&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">3<\/span><span class=\"token punctuation\">}<\/span><span class=\"token punctuation\">,<\/span><br \/>\n    <span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;name&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u626b\u63cf\u4ef6\/\u56fe\u7247\u9875&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;n&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">2<\/span><span class=\"token punctuation\">}<\/span><br \/>\n  <span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span><br \/>\n  <span class=\"token string-property property\">&#034;per_doc_annotation&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;fields&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">10<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;tables&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u5168\u90e8&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string-property property\">&#034;section_tree&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token boolean\">true<\/span><span class=\"token punctuation\">}<\/span><span class=\"token punctuation\">,<\/span><br \/>\n  <span class=\"token string-property property\">&#034;total_fields&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token number\">200<\/span><span class=\"token punctuation\">,<\/span><br \/>\n  <span class=\"token string-property property\">&#034;reannotate&#034;<\/span><span class=\"token operator\">:<\/span> <span class=\"token string\">&#034;\u540c\u4e00\u4eba\u6807\u5b8c 20 \u7bc7&#xff1b;\u6362\u4eba\u91cd\u6807 3 \u7bc7\u7b97\u4e00\u81f4\u6027&#xff0c;\u4f4e\u4e8e 0.9 \u91cd\u6807&#034;<\/span><br \/>\n<span class=\"token punctuation\">}<\/span><\/p>\n<p>\u5206\u5c42\u62bd\u6837\u662f\u56e0\u4e3a\u7248\u5f0f\u51b3\u5b9a\u635f\u5931\u5f62\u6001&#xff0c;\u968f\u673a\u62bd 20 \u7bc7\u5f88\u53ef\u80fd\u5168\u662f\u7eaf\u6587\u672c&#xff0c;\u8dd1\u51fa\u6765\u7684\u7ed3\u6784\u4fdd\u6301\u7387\u597d\u770b\u4f46\u6ca1\u6709\u610f\u4e49\u3002\u6bcf\u7bc7\u6807 10 \u4e2a\u5b57\u6bb5&#xff0c;\u6807\u7684\u662f\u300c\u4e0b\u6e38\u771f\u4f1a\u95ee\u5230\u7684\u503c\u300d\u2014\u2014\u9608\u503c\u3001\u7f16\u53f7\u3001\u670d\u52a1\u540d\u3001\u52a8\u4f5c\u540d\u3002<\/p>\n<h4>\u4ece\u8f93\u5165\u5230\u9a8c\u8bc1\u8dd1\u4e00\u904d<\/h4>\n<p>\u62ff RB-214 \u8d70\u4e00\u904d\u56db\u6b65&#xff1a;<\/p>\n<li>\u8f93\u5165&#xff1a;RB-214.pdf&#xff0c;12 \u9875&#xff0c;3 \u5f20\u8868&#xff0c;\u6709\u6587\u672c\u5c42&#xff0c;\u65e0\u626b\u63cf\u9875\u3002<\/li>\n<li>\u5904\u7406&#xff1a;\u6863 B \u89e3\u6790\u5668&#xff08;\u5e03\u5c40\u611f\u77e5&#xff09;\u8dd1\u5b8c&#xff0c;\u8f93\u51fa\u5757\u6d41 blocks.jsonl\u3002<\/li>\n<li>\u8f93\u51fa&#xff1a;<\/li>\n<p><span class=\"token punctuation\">{<\/span><span class=\"token string-property property\">&#034;doc_id&#034;<\/span><span class=\"token operator\">:<\/span><span class=\"token string\">&#034;RB-214&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string-property property\">&#034;btype&#034;<\/span><span class=\"token operator\">:<\/span><span class=\"token string\">&#034;table&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string-property property\">&#034;text&#034;<\/span><span class=\"token operator\">:<\/span><span class=\"token string\">&#034;\u6b65\u9aa4 \u89e6\u53d1\u6761\u4ef6 \u5904\u7f6e\u52a8\u4f5c &#8230;&#034;<\/span><span class=\"token punctuation\">,<\/span><br \/>\n <span class=\"token string-property property\">&#034;page&#034;<\/span><span class=\"token operator\">:<\/span><span class=\"token number\">4<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string-property property\">&#034;section&#034;<\/span><span class=\"token operator\">:<\/span><span class=\"token string\">&#034;2.3&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string-property property\">&#034;rows&#034;<\/span><span class=\"token operator\">:<\/span><span class=\"token punctuation\">[<\/span><span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;\u6b65\u9aa4&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string\">&#034;\u89e6\u53d1\u6761\u4ef6&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string\">&#034;\u5904\u7f6e\u52a8\u4f5c&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span><span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;1&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string\">&#034;CPU &gt; 80%&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string\">&#034;\u6269\u5bb9&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span><span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;2&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string\">&#034;\u78c1\u76d8\u6c34\u4f4d &gt;85%&#034;<\/span><span class=\"token punctuation\">,<\/span><span class=\"token string\">&#034;\u5148\u6e05\u65e5\u5fd7\u518d\u6269\u5bb9&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">}<\/span><\/p>\n<li>\u9a8c\u8bc1&#xff1a;rows \u662f 3\u00d73 \u4e8c\u7ef4\u6570\u7ec4&#xff0c;\u4e0e gold \u4e00\u81f4 \u2192 \u8be5\u8868\u7ed3\u6784\u4fdd\u6301\u7387\u8ba1\u6ee1\u5206&#xff1b;\u9875\u7801 4 \u4e0e section 2.3 \u975e\u7a7a \u2192 \u5143\u6570\u636e\u73af\u8282\u901a\u8fc7&#xff1b;\u628a\u5757\u6d41\u5582\u8fdb\u7d22\u5f15&#xff0c;\u7528 opsqa-200 \u91cc\u4e0e RB-214 \u76f8\u5173\u7684 7 \u6761\u95ee\u9898\u8dd1\u4e00\u6b21&#xff0c;\u770b\u5904\u7f6e\u52a8\u4f5c\u547d\u4e2d\u7387\u662f\u5426\u4ece 3\/7 \u62ac\u4e0a\u53bb\u3002<\/li>\n<p>\u7b2c\u56db\u6b65\u662f\u5b9e\u9a8c\u7eaa\u5f8b\u7684\u843d\u70b9&#xff1a;\u53ea\u6709\u4e2d\u95f4\u6307\u6807\u53d8\u597d\u4e0d\u7b97\u6570\u3002\u7ed3\u6784\u4fdd\u6301\u7387\u4ece 0.31 \u62ac\u5230 0.85&#xff0c;\u90a3 7 \u6761\u95ee\u9898\u8981\u662f\u8fd8\u53ea\u4e2d 3 \u6761&#xff0c;\u8bf4\u660e\u574f\u7684\u4e0d\u6b62\u89e3\u6790\u8fd9\u4e00\u73af\u3002<\/p>\n<h4>\u4e09\u6863\u5404\u8dd1\u4e00\u904d&#xff0c;\u843d\u5230\u540c\u4e00\u5f20\u8868<\/h4>\n<p>20 \u7bc7\u8bc4\u6d4b\u96c6\u56fa\u5b9a&#xff0c;\u4e09\u6863\u89e3\u6790\u5668\u5404\u8dd1\u4e00\u6b21&#xff0c;\u56db\u4e2a\u4e2d\u95f4\u6307\u6807\u52a0\u4e00\u4e2a\u4e0b\u6e38\u6307\u6807&#xff0c;\u7ed3\u679c\u5982\u4e0b&#xff1a;<\/p>\n<table>\n<tr>\u65b9\u6848\u5b57\u6bb5\u53ec\u56de\u7387\u7ed3\u6784\u4fdd\u6301\u7387\u8de8\u5757\u5b8c\u6574\u7387\u5355\u4f4d\u6210\u672c USD\/\u5343\u9875opsqa-200 \u51c6\u786e\u7387<\/tr>\n<tbody>\n<tr>\n<td>A \u7eaf\u6587\u672c\u62bd\u53d6&#xff08;\u57fa\u7ebf&#xff09;<\/td>\n<td>0.586<\/td>\n<td>0.311<\/td>\n<td>0.742<\/td>\n<td>0.02<\/td>\n<td>0.680<\/td>\n<\/tr>\n<tr>\n<td>B \u5e03\u5c40\u611f\u77e5<\/td>\n<td>0.912<\/td>\n<td>0.826<\/td>\n<td>0.913<\/td>\n<td>0.31<\/td>\n<td>0.700<\/td>\n<\/tr>\n<tr>\n<td>C VLM \u6574\u9875<\/td>\n<td>0.947<\/td>\n<td>0.941<\/td>\n<td>0.928<\/td>\n<td>15.00<\/td>\n<td>0.710<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6210\u672c\u90a3\u4e00\u5217\u7684\u53e3\u5f84\u8981\u5199\u6e05\u695a&#xff0c;\u5426\u5219\u4e09\u6863\u6ca1\u6cd5\u6bd4&#xff1a;A \u662f\u7eaf CPU&#xff0c;\u53ea\u6709\u7535\u8d39\u4e0e\u7f16\u6392\u5f00\u9500&#xff1b;B \u6309 1.2 s\/\u9875 \u00d7 0.7 USD\/h \u7684 GPU \u79df\u7528\u644a\u5230\u6bcf\u9875&#xff0c;\u518d\u52a0 30% \u7f16\u6392\u4e0e IO \u5f00\u9500 [\u7b97]&#xff1b;C \u8d70\u4e91\u7aef VLM API&#xff0c;\u6309\u6bcf\u9875 2,000 \u8f93\u5165 token \u00d7 3 USD\/M &#043; 600 \u8f93\u51fa token \u00d7 15 USD\/M \u4f30 [\u7b97]\u3002\u4e09\u4e2a\u6570\u7684\u91cf\u7ea7\u5173\u7cfb&#xff08;0.02 \/ 0.31 \/ 15.00&#xff09;\u6bd4\u7edd\u5bf9\u503c\u91cd\u8981\u2014\u2014\u6362 GPU \u578b\u53f7\u3001\u6362\u6a21\u578b\u6863\u4f4d\u90fd\u53ea\u6539\u6bd4\u4f8b\u4e0d\u6539\u7ed3\u8bba\u3002<\/p>\n<p>\u8bfb\u8fd9\u5f20\u8868\u7684\u987a\u5e8f\u662f\u56fa\u5b9a\u7684&#xff1a;\u5148\u770b\u6700\u540e\u4e00\u5217&#xff0c;\u518d\u770b\u5355\u4f4d\u6210\u672c&#xff0c;\u6700\u540e\u624d\u770b\u4e2d\u95f4\u4e09\u5217\u3002\u4e2d\u95f4\u4e09\u5217\u7528\u6765\u89e3\u91ca\u6700\u540e\u4e00\u5217\u4e3a\u4ec0\u4e48\u53d8\u597d\u6216\u6ca1\u53d8\u597d\u2014\u2014B \u6863\u7ed3\u6784\u4fdd\u6301\u7387\u5347\u4e86 51.5pp [\u7b97]&#xff0c;\u4e0b\u6e38\u53ea\u5347 2pp&#xff0c;\u8bf4\u660e\u8fd9\u6279\u95ee\u9898\u8981\u7684\u7b54\u6848\u591a\u6570\u4e0d\u5728\u8868\u683c\u91cc&#xff0c;\u89e3\u6790\u4e0d\u662f\u552f\u4e00\u74f6\u9888&#xff1b;C \u6863\u7ed3\u6784\u4fdd\u6301\u7387\u518d\u5347 11.5pp [\u7b97]&#xff0c;\u4e0b\u6e38\u53ea\u5347 1pp&#xff0c;\u4e14\u8fd9\u4e00\u6863\u8981\u4ed8 48 \u500d\u7684\u94b1 [\u7b97]\u3002\u6309 20 \u7bc7\u7684\u7f6e\u4fe1\u534a\u5bbd&#xff08;\u00b13.5pp [\u7b97]&#xff09;&#xff0c;1pp \u7684\u5dee\u503c\u5224\u4e3a\u4e0d\u53ef\u533a\u5206&#xff0c;\u9009\u4fbf\u5b9c\u7684\u90a3\u6863\u3002<\/p>\n<hr \/>\n<h3>\u4e94\u3001\u4e09\u6863\u65b9\u6848\u600e\u4e48\u9009&#xff0c;\u6210\u672c\u66f2\u7ebf\u5728\u54ea\u62d0&#xff1f;<\/h3>\n<p>  #mermaid-svg-ifxv6uo49Om01Pta{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-ifxv6uo49Om01Pta .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-ifxv6uo49Om01Pta .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-ifxv6uo49Om01Pta .error-icon{fill:#552222;}#mermaid-svg-ifxv6uo49Om01Pta .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-ifxv6uo49Om01Pta .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-ifxv6uo49Om01Pta .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-ifxv6uo49Om01Pta .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-ifxv6uo49Om01Pta .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-ifxv6uo49Om01Pta .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-ifxv6uo49Om01Pta .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-ifxv6uo49Om01Pta .marker{fill:#333333;stroke:#333333;}#mermaid-svg-ifxv6uo49Om01Pta .marker.cross{stroke:#333333;}#mermaid-svg-ifxv6uo49Om01Pta svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-ifxv6uo49Om01Pta p{margin:0;}#mermaid-svg-ifxv6uo49Om01Pta .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-ifxv6uo49Om01Pta .cluster-label text{fill:#333;}#mermaid-svg-ifxv6uo49Om01Pta .cluster-label span{color:#333;}#mermaid-svg-ifxv6uo49Om01Pta .cluster-label span p{background-color:transparent;}#mermaid-svg-ifxv6uo49Om01Pta .label text,#mermaid-svg-ifxv6uo49Om01Pta span{fill:#333;color:#333;}#mermaid-svg-ifxv6uo49Om01Pta .node rect,#mermaid-svg-ifxv6uo49Om01Pta .node circle,#mermaid-svg-ifxv6uo49Om01Pta .node ellipse,#mermaid-svg-ifxv6uo49Om01Pta .node polygon,#mermaid-svg-ifxv6uo49Om01Pta .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-ifxv6uo49Om01Pta .rough-node .label text,#mermaid-svg-ifxv6uo49Om01Pta .node .label text,#mermaid-svg-ifxv6uo49Om01Pta .image-shape .label,#mermaid-svg-ifxv6uo49Om01Pta .icon-shape .label{text-anchor:middle;}#mermaid-svg-ifxv6uo49Om01Pta .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-ifxv6uo49Om01Pta .rough-node .label,#mermaid-svg-ifxv6uo49Om01Pta .node .label,#mermaid-svg-ifxv6uo49Om01Pta .image-shape .label,#mermaid-svg-ifxv6uo49Om01Pta .icon-shape .label{text-align:center;}#mermaid-svg-ifxv6uo49Om01Pta .node.clickable{cursor:pointer;}#mermaid-svg-ifxv6uo49Om01Pta .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-ifxv6uo49Om01Pta .arrowheadPath{fill:#333333;}#mermaid-svg-ifxv6uo49Om01Pta .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-ifxv6uo49Om01Pta .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-ifxv6uo49Om01Pta .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-ifxv6uo49Om01Pta .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-ifxv6uo49Om01Pta .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-ifxv6uo49Om01Pta .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-ifxv6uo49Om01Pta .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-ifxv6uo49Om01Pta .cluster text{fill:#333;}#mermaid-svg-ifxv6uo49Om01Pta .cluster span{color:#333;}#mermaid-svg-ifxv6uo49Om01Pta div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-ifxv6uo49Om01Pta .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-ifxv6uo49Om01Pta rect.text{fill:none;stroke-width:0;}#mermaid-svg-ifxv6uo49Om01Pta .icon-shape,#mermaid-svg-ifxv6uo49Om01Pta .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-ifxv6uo49Om01Pta .icon-shape p,#mermaid-svg-ifxv6uo49Om01Pta .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-ifxv6uo49Om01Pta .icon-shape .label rect,#mermaid-svg-ifxv6uo49Om01Pta .image-shape .label rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-ifxv6uo49Om01Pta .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-ifxv6uo49Om01Pta .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-ifxv6uo49Om01Pta :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u89e3\u6790\u5668\u4e94\u6784\u4ef6<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u89e3\u7801\u5668<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u7248\u9762\u5668<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u8868\u683c\u8fd8\u539f<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u9605\u8bfb\u6392\u5e8f<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u5143\u6570\u636e\u62bd\u53d6<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u7ed3\u6784\u5316\u5757\u6d41<\/p>\n<p><\/span><\/p>\n<p>\u4e09\u6863\u7684\u672c\u8d28\u5dee\u5f02\u4e0d\u662f\u300c\u51c6\u4e0d\u51c6\u300d&#xff0c;\u662f\u54ea\u4e00\u73af\u7531\u4ec0\u4e48\u6765\u505a&#xff1a;<\/p>\n<ul>\n<li>\u6863 A \u7eaf\u6587\u672c\u62bd\u53d6&#xff1a;\u53ea\u6709\u89e3\u7801\u5668&#xff0c;\u540e\u56db\u73af\u57fa\u672c\u7f3a\u5e2d\u3002\u4ee3\u8868&#xff1a;pdfplumber\u3001PyMuPDF \u7684\u53d6\u5b57\u63a5\u53e3\u3002<\/li>\n<li>\u6863 B \u5e03\u5c40\u611f\u77e5&#xff1a;\u4e94\u73af\u9f50\u5168&#xff0c;\u7248\u9762\u4e0e\u8868\u683c\u9760\u4e13\u95e8\u6a21\u578b\u6216\u89c4\u5219\u3002\u4ee3\u8868&#xff1a;Unstructured\u3001Marker\u3001MinerU\u3001Docling&#xff1b;\u626b\u63cf\u4ef6\u6302 Tesseract OCR \u6216 PaddleOCR\u3002<\/li>\n<li>\u6863 C VLM \u6574\u9875\u7406\u89e3&#xff1a;\u6574\u9875\u6e32\u67d3\u6210\u56fe\u4ea4\u7ed9\u89c6\u89c9\u8bed\u8a00\u6a21\u578b&#xff0c;\u4e00\u6b21\u8f93\u51fa\u5e26\u7ed3\u6784\u7684 Markdown\u3002\u4ee3\u8868&#xff1a;Nougat \u4e00\u7c7b\u7aef\u5230\u7aef\u6a21\u578b\u3002\u9009\u578b\u4e0d\u5c55\u5f00&#xff0c;\u5c5e\u6587\u6863\u667a\u80fd\u4e3b\u9898\u3002<\/li>\n<\/ul>\n<p>\u6211\u7684\u5224\u65ad\u662f&#xff1a;\u5927\u591a\u6570\u56e2\u961f\u7684\u9ed8\u8ba4\u6863\u4f4d\u5e94\u8be5\u662f B&#xff0c;\u4e0d\u662f A \u4e5f\u4e0d\u662f C\u3002 \u7406\u7531\u662f A \u7684\u7ed3\u6784\u4fdd\u6301\u7387\u5728\u542b\u8868\u6587\u6863\u4e0a\u901a\u5e38\u4f4e\u4e8e 0.4&#xff0c;\u800c C \u7684\u5355\u4f4d\u6210\u672c\u6bd4 B \u9ad8\u4e00\u5230\u4e24\u4e2a\u6570\u91cf\u7ea7&#xff0c;\u6362\u6765\u7684\u589e\u91cf\u96c6\u4e2d\u5728\u300c\u5408\u5e76\u5355\u5143\u683c\u300d\u4e0e\u300c\u626b\u63cf\u4ef6\u300d\u4e24\u7c7b&#xff0c;\u524d\u8005\u5360\u6bd4\u901a\u5e38\u4e0d\u5230\u6587\u6863\u7684 10%\u3002<\/p>\n<table>\n<tr>\u6587\u6863\u7c7b\u578b\u626b\u63cf\u9875\u5360\u6bd4\u8868\u683c\u5bc6\u5ea6\u5b57\u6bb5\u53ec\u56de\u7387\u8981\u6c42\u6210\u672c\u4e0a\u9650\u5efa\u8bae\u6863<\/tr>\n<tbody>\n<tr>\n<td>\u7eaf\u6587\u672c\u9875&#xff08;Confluence \u5bfc\u51fa \/ \u6307\u6807\u8bf4\u660e&#xff09;<\/td>\n<td>0%<\/td>\n<td>\u4f4e<\/td>\n<td>\u22650.85<\/td>\n<td>\u6781\u4f4e<\/td>\n<td>A&#xff0c;\u62bd\u68c0\u540e\u8f6c B<\/td>\n<\/tr>\n<tr>\n<td>runbook&#xff08;\u542b\u8868\u4e0e\u622a\u56fe&#xff09;<\/td>\n<td>0\u201310%<\/td>\n<td>\u4e2d\u9ad8<\/td>\n<td>\u22650.90<\/td>\n<td>\u4e2d<\/td>\n<td>B&#xff0c;\u626b\u63cf\u9875\u5355\u72ec\u8d70 OCR<\/td>\n<\/tr>\n<tr>\n<td>\u626b\u63cf\u7248\u5de5\u5355\/\u5408\u540c<\/td>\n<td>100%<\/td>\n<td>\u4e2d<\/td>\n<td>\u22650.90<\/td>\n<td>\u4e2d\u9ad8<\/td>\n<td>B &#043; OCR&#xff0c;\u8868\u683c\u5bc6\u96c6\u5904\u7528 C<\/td>\n<\/tr>\n<tr>\n<td>\u8d22\u62a5\/\u8bba\u6587&#xff08;\u591a\u7ea7\u8868\u5934&#xff09;<\/td>\n<td>0%<\/td>\n<td>\u9ad8<\/td>\n<td>\u22650.95<\/td>\n<td>\u9ad8<\/td>\n<td>C&#xff0c;\u6216 B &#043; \u4eba\u5de5\u62bd\u68c0\u590d\u6838<\/td>\n<\/tr>\n<tr>\n<td>\u67b6\u6784\u56fe\/\u62d3\u6251\u56fe\u4e3a\u4e3b<\/td>\n<td>\u2014<\/td>\n<td>\u2014<\/td>\n<td>\u56fe\u5185\u6587\u5b57\u53ef\u5bfb\u5740<\/td>\n<td>\u9ad8<\/td>\n<td>C&#xff0c;\u4e14\u9700\u8bc4\u4f30\u56fe\u8f6c\u6587\u5b57\u7684\u5931\u771f<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6210\u672c\u66f2\u7ebf\u5728\u54ea\u62d0&#xff0c;\u6309 6,000 \u9875&#xff08;\u5bf9\u5e94 topo_docs 12,000 \u6761\u76ee\u7684\u91cf\u7ea7 &#xff09;\u7b97\u4e00\u7b14&#xff1a;<\/p>\n<table>\n<tr>\u6863\u9875\u5747\u8017\u65f66,000 \u9875\u603b\u8017\u65f6 [\u7b97]\u7ed3\u6784\u4fdd\u6301\u7387\u671f\u671b\u62d0\u70b9\u5224\u636e<\/tr>\n<tbody>\n<tr>\n<td>A \u7eaf\u6587\u672c<\/td>\n<td>10\u201350 ms<\/td>\n<td>1\u20135 \u5206\u949f<\/td>\n<td>0.2\u20130.4<\/td>\n<td>\u6587\u6863\u542b\u8868\u5373\u4e0d\u53ef\u63a5\u53d7<\/td>\n<\/tr>\n<tr>\n<td>B \u5e03\u5c40\u611f\u77e5<\/td>\n<td>0.3\u20133 s<\/td>\n<td>30 \u5206\u949f\u20135 \u5c0f\u65f6<\/td>\n<td>0.8\u20130.95<\/td>\n<td>\u5e38\u89c4\u573a\u666f\u7684\u9ed8\u8ba4\u843d\u70b9<\/td>\n<\/tr>\n<tr>\n<td>C VLM \u6574\u9875<\/td>\n<td>2\u201315 s<\/td>\n<td>3.3\u201325 \u5c0f\u65f6<\/td>\n<td>0.9\u20130.98<\/td>\n<td>\u53ea\u5bf9\u96be\u4f8b\u4e0e\u9ad8\u4ef7\u503c\u6587\u6863\u5f00\u542f<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u62d0\u70b9\u4e0d\u662f\u4e2a\u70b9&#xff0c;\u662f\u6761\u89c4\u5219&#xff1a;\u8868\u683c\u5bc6\u5ea6\u8d85\u8fc7\u7ea6 15%&#xff0c;A \u5c31\u4e0d\u5212\u7b97&#xff1b;\u5408\u5e76\u5355\u5143\u683c\u6216\u626b\u63cf\u9875\u5360\u6bd4\u4f4e\u4e8e 10%&#xff0c;C \u5c31\u4e0d\u5212\u7b97\u3002\u4e2d\u95f4\u90a3\u6bb5\u662f B \u7684\u5730\u76d8&#xff0c;\u4e5f\u662f\u591a\u6570\u4f01\u4e1a\u6587\u6863\u5e93\u7684\u771f\u5b9e\u4f4d\u7f6e\u3002<\/p>\n<p>  #mermaid-svg-dFdPh90ug67SHptP{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;fill:#333;}@keyframes edge-animation-frame{from{stroke-dashoffset:0;}}@keyframes dash{to{stroke-dashoffset:0;}}#mermaid-svg-dFdPh90ug67SHptP .edge-animation-slow{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 50s linear infinite;stroke-linecap:round;}#mermaid-svg-dFdPh90ug67SHptP .edge-animation-fast{stroke-dasharray:9,5!important;stroke-dashoffset:900;animation:dash 20s linear infinite;stroke-linecap:round;}#mermaid-svg-dFdPh90ug67SHptP .error-icon{fill:#552222;}#mermaid-svg-dFdPh90ug67SHptP .error-text{fill:#552222;stroke:#552222;}#mermaid-svg-dFdPh90ug67SHptP .edge-thickness-normal{stroke-width:1px;}#mermaid-svg-dFdPh90ug67SHptP .edge-thickness-thick{stroke-width:3.5px;}#mermaid-svg-dFdPh90ug67SHptP .edge-pattern-solid{stroke-dasharray:0;}#mermaid-svg-dFdPh90ug67SHptP .edge-thickness-invisible{stroke-width:0;fill:none;}#mermaid-svg-dFdPh90ug67SHptP .edge-pattern-dashed{stroke-dasharray:3;}#mermaid-svg-dFdPh90ug67SHptP .edge-pattern-dotted{stroke-dasharray:2;}#mermaid-svg-dFdPh90ug67SHptP .marker{fill:#333333;stroke:#333333;}#mermaid-svg-dFdPh90ug67SHptP .marker.cross{stroke:#333333;}#mermaid-svg-dFdPh90ug67SHptP svg{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:16px;}#mermaid-svg-dFdPh90ug67SHptP p{margin:0;}#mermaid-svg-dFdPh90ug67SHptP .label{font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;color:#333;}#mermaid-svg-dFdPh90ug67SHptP .cluster-label text{fill:#333;}#mermaid-svg-dFdPh90ug67SHptP .cluster-label span{color:#333;}#mermaid-svg-dFdPh90ug67SHptP .cluster-label span p{background-color:transparent;}#mermaid-svg-dFdPh90ug67SHptP .label text,#mermaid-svg-dFdPh90ug67SHptP span{fill:#333;color:#333;}#mermaid-svg-dFdPh90ug67SHptP .node rect,#mermaid-svg-dFdPh90ug67SHptP .node circle,#mermaid-svg-dFdPh90ug67SHptP .node ellipse,#mermaid-svg-dFdPh90ug67SHptP .node polygon,#mermaid-svg-dFdPh90ug67SHptP .node path{fill:#ECECFF;stroke:#9370DB;stroke-width:1px;}#mermaid-svg-dFdPh90ug67SHptP .rough-node .label text,#mermaid-svg-dFdPh90ug67SHptP .node .label text,#mermaid-svg-dFdPh90ug67SHptP .image-shape .label,#mermaid-svg-dFdPh90ug67SHptP .icon-shape .label{text-anchor:middle;}#mermaid-svg-dFdPh90ug67SHptP .node .katex path{fill:#000;stroke:#000;stroke-width:1px;}#mermaid-svg-dFdPh90ug67SHptP .rough-node .label,#mermaid-svg-dFdPh90ug67SHptP .node .label,#mermaid-svg-dFdPh90ug67SHptP .image-shape .label,#mermaid-svg-dFdPh90ug67SHptP .icon-shape .label{text-align:center;}#mermaid-svg-dFdPh90ug67SHptP .node.clickable{cursor:pointer;}#mermaid-svg-dFdPh90ug67SHptP .root .anchor path{fill:#333333!important;stroke-width:0;stroke:#333333;}#mermaid-svg-dFdPh90ug67SHptP .arrowheadPath{fill:#333333;}#mermaid-svg-dFdPh90ug67SHptP .edgePath .path{stroke:#333333;stroke-width:2.0px;}#mermaid-svg-dFdPh90ug67SHptP .flowchart-link{stroke:#333333;fill:none;}#mermaid-svg-dFdPh90ug67SHptP .edgeLabel{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-dFdPh90ug67SHptP .edgeLabel p{background-color:rgba(232,232,232, 0.8);}#mermaid-svg-dFdPh90ug67SHptP .edgeLabel rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-dFdPh90ug67SHptP .labelBkg{background-color:rgba(232, 232, 232, 0.5);}#mermaid-svg-dFdPh90ug67SHptP .cluster rect{fill:#ffffde;stroke:#aaaa33;stroke-width:1px;}#mermaid-svg-dFdPh90ug67SHptP .cluster text{fill:#333;}#mermaid-svg-dFdPh90ug67SHptP .cluster span{color:#333;}#mermaid-svg-dFdPh90ug67SHptP div.mermaidTooltip{position:absolute;text-align:center;max-width:200px;padding:2px;font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;font-size:12px;background:hsl(80, 100%, 96.2745098039%);border:1px solid #aaaa33;border-radius:2px;pointer-events:none;z-index:100;}#mermaid-svg-dFdPh90ug67SHptP .flowchartTitleText{text-anchor:middle;font-size:18px;fill:#333;}#mermaid-svg-dFdPh90ug67SHptP rect.text{fill:none;stroke-width:0;}#mermaid-svg-dFdPh90ug67SHptP .icon-shape,#mermaid-svg-dFdPh90ug67SHptP .image-shape{background-color:rgba(232,232,232, 0.8);text-align:center;}#mermaid-svg-dFdPh90ug67SHptP .icon-shape p,#mermaid-svg-dFdPh90ug67SHptP .image-shape p{background-color:rgba(232,232,232, 0.8);padding:2px;}#mermaid-svg-dFdPh90ug67SHptP .icon-shape .label rect,#mermaid-svg-dFdPh90ug67SHptP .image-shape .label rect{opacity:0.5;background-color:rgba(232,232,232, 0.8);fill:rgba(232,232,232, 0.8);}#mermaid-svg-dFdPh90ug67SHptP .label-icon{display:inline-block;height:1em;overflow:visible;vertical-align:-0.125em;}#mermaid-svg-dFdPh90ug67SHptP .node .label-icon path{fill:currentColor;stroke:revert;stroke-width:revert;}#mermaid-svg-dFdPh90ug67SHptP :root{&#8211;mermaid-font-family:\\&#8221;trebuchet ms\\&#8221;,verdana,arial,sans-serif;}<\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"edgeLabel\"><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>PDF \u4e8c\u8fdb\u5236<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u7248\u9762\u6811<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u7ed3\u6784\u5316\u5757\u6d41<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u5e26\u5143\u6570\u636e\u77e5\u8bc6\u6761\u76ee<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u5012\u6392\u4e0e\u5411\u91cf\u7d22\u5f15<\/p>\n<p><\/span><\/p>\n<p>         <span class=\"nodeLabel\"><\/p>\n<p>\u67e5\u8be2<\/p>\n<p><\/span><\/p>\n<hr \/>\n<h3>\u516d\u3001\u5757\u6d41\u8be5\u957f\u4ec0\u4e48\u6837&#xff1a;\u4e00\u4efd\u80fd\u6821\u9a8c\u7684 Schema \u4e0e\u8fdd\u7ea6\u5904\u7f6e<\/h3>\n<p>\u56db\u628a\u5c3a\u5b50\u91cf\u7684\u662f\u300c\u89e3\u6790\u5668\u5e72\u5f97\u597d\u4e0d\u597d\u300d&#xff0c;\u4f46\u8fd8\u6709\u4e00\u7c7b\u6545\u969c\u4e0d\u5c5e\u4e8e\u5e72\u5f97\u597d\u4e0d\u597d\u2014\u2014\u4ea7\u7269\u5b57\u6bb5\u672c\u8eab\u5c31\u4e0d\u9f50\u3002\u8fd9\u7c7b\u6545\u969c\u4e0d\u5728\u89e3\u6790\u9636\u6bb5\u7ed3\u7b97&#xff0c;\u800c\u662f\u7b49\u5230\u300c[06-\u5f15\u7528\u4e0e\u53ef\u6eaf\u6e90 \u8fd8\u5728\u5199]\u300d\u505a\u5f15\u7528\u5b9a\u4f4d\u65f6\u624d\u53d1\u73b0 section \u5168\u7a7a&#xff0c;\u6216\u8005 \u300c[03-\u5207\u5206-\u8bed\u4e49\u5b8c\u6574\u4e0e\u68c0\u7d22\u7cbe\u5ea6 \u8fd8\u5728\u5199]\u300d\u60f3\u6309\u7ae0\u8282\u8def\u5f84\u5207\u5757\u65f6\u624d\u53d1\u73b0 page \u53ea\u6709 0\u3002\u90a3\u65f6\u5019\u8981\u8fd4\u5de5\u7684\u662f\u6574\u6761\u7d22\u5f15\u3002<\/p>\n<p>\u6240\u4ee5\u89e3\u6790\u5668\u4e0e\u4e0b\u6e38\u4e4b\u95f4\u9700\u8981\u4e00\u4efd\u663e\u5f0f\u5951\u7ea6\u3002\u5b83\u4e0d\u9700\u8981\u590d\u6742&#xff0c;\u53ea\u9700\u8981\u80fd\u88ab\u673a\u5668\u6821\u9a8c&#xff1a;\u54ea\u4e9b\u5b57\u6bb5\u5fc5\u586b\u3001\u7f3a\u4e86\u8c01\u964d\u7ea7\u3001\u54ea\u4e9b\u5fc5\u987b\u62d2\u7edd\u5165\u5e93\u3002<\/p>\n<table>\n<tr>\u5b57\u6bb5\u7c7b\u578b\u8c01\u6d88\u8d39\u7f3a\u4e86\u4f1a\u600e\u6837\u8fdd\u7ea6\u5904\u7f6e<\/tr>\n<tbody>\n<tr>\n<td>doc_id<\/td>\n<td>str<\/td>\n<td>\u5168\u90e8<\/td>\n<td>\u65e0\u6cd5\u56de\u6eaf\u539f\u6587\u6863&#xff0c;\u91cd\u89e3\u6790\u8303\u56f4\u4e5f\u7b97\u4e0d\u51fa\u6765<\/td>\n<td>\u62d2\u7edd\u5165\u5e93<\/td>\n<\/tr>\n<tr>\n<td>block_id<\/td>\n<td>str<\/td>\n<td>\u300c[06-\u5f15\u7528\u4e0e\u53ef\u6eaf\u6e90 \u8fd8\u5728\u5199]\u300d \u7684 span_id<\/td>\n<td>\u5f15\u7528\u6307\u5411\u7684\u7247\u6bb5 id \u65e0\u5904\u5b89\u653e<\/td>\n<td>\u62d2\u7edd\u5165\u5e93<\/td>\n<\/tr>\n<tr>\n<td>btype<\/td>\n<td>enum text\/table\/title\/caption<\/td>\n<td>\u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d \/ \u300c[03-\u5207\u5206-\u8bed\u4e49\u5b8c\u6574\u4e0e\u68c0\u7d22\u7cbe\u5ea6 \u8fd8\u5728\u5199]\u300d<\/td>\n<td>\u8868\u683c\u4e0e\u6b63\u6587\u6df7\u5728\u4e00\u8d77&#xff0c;\u65ad\u88c2\u68c0\u6d4b\u5931\u53bb\u8f93\u5165<\/td>\n<td>\u964d\u7ea7\u4e3a text \u5e76\u8ba1\u6570<\/td>\n<\/tr>\n<tr>\n<td>text<\/td>\n<td>str<\/td>\n<td>\u5168\u90e8<\/td>\n<td>\u2014<\/td>\n<td>\u62d2\u7edd\u5165\u5e93<\/td>\n<\/tr>\n<tr>\n<td>page<\/td>\n<td>int \u22651<\/td>\n<td>\u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d \u8de8\u9875\u5408\u5e76\u3001\u300c[06-\u5f15\u7528\u4e0e\u53ef\u6eaf\u6e90 \u8fd8\u5728\u5199]\u300d \u5f15\u7528<\/td>\n<td>\u8de8\u9875\u8868\u65e0\u6cd5\u7eed\u63a5&#xff0c;\u5f15\u7528\u7ed9\u4e0d\u51fa\u9875\u7801<\/td>\n<td>\u8bb0 -1 \u5e76\u8ba1\u5165\u7a7a\u503c\u7387<\/td>\n<\/tr>\n<tr>\n<td>section<\/td>\n<td>str<\/td>\n<td>\u300c[03-\u5207\u5206-\u8bed\u4e49\u5b8c\u6574\u4e0e\u68c0\u7d22\u7cbe\u5ea6 \u8fd8\u5728\u5199]\u300d \u6309\u7ae0\u8282\u5207\u3001\u300c[06-\u5f15\u7528\u4e0e\u53ef\u6eaf\u6e90 \u8fd8\u5728\u5199]\u300d \u5b9a\u4f4d<\/td>\n<td>\u5c42\u7ea7\u6811\u9000\u5316\u6210\u5e73\u8868&#xff0c;\u5f15\u7528\u5b9a\u4f4d\u4e0d\u5230\u7ae0\u8282<\/td>\n<td>\u8bb0 &#034;&#034; \u5e76\u8ba1\u5165\u7a7a\u503c\u7387<\/td>\n<\/tr>\n<tr>\n<td>rows<\/td>\n<td>list[list[str]]&#xff08;\u4ec5 table&#xff09;<\/td>\n<td>\u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d \u8868\u5934\u7ee7\u627f<\/td>\n<td>\u8868\u683c\u53ea\u5269\u62c9\u5e73\u6587\u672c&#xff0c;\u5b57\u6bb5\u53ef\u5b9a\u4f4d\u7387\u5f52\u96f6<\/td>\n<td>btype&#061;table \u800c rows \u4e3a\u7a7a \u2192 \u8ba1\u4e00\u6b21\u7ed3\u6784\u8fdd\u7ea6<\/td>\n<\/tr>\n<tr>\n<td>bbox<\/td>\n<td>list[float]&#xff08;4 \u4e2a&#xff09;<\/td>\n<td>\u7248\u9762\u8c03\u8bd5\u3001\u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d \u5750\u6807\u7ea6\u675f<\/td>\n<td>\u540c\u9875\u591a\u5f20\u7ed3\u6784\u76f8\u540c\u7684\u8868\u65e0\u6cd5\u533a\u5206<\/td>\n<td>\u53ef\u9009&#xff0c;\u7f3a\u4e86\u53ea\u5f71\u54cd\u8c03\u8bd5<\/td>\n<\/tr>\n<tr>\n<td>source_uri<\/td>\n<td>str<\/td>\n<td>\u300c[07-\u77e5\u8bc6\u65b0\u9c9c\u5ea6 \u8fd8\u5728\u5199]\u300d \u6743\u5a01\u6e90\u6bd4\u5bf9<\/td>\n<td>\u5de1\u68c0\u65f6\u4e0d\u77e5\u9053\u8be5\u53bb\u54ea\u513f\u6838\u5bf9<\/td>\n<td>\u8bb0 &#034;&#034; \u5e76\u5728 07 \u524d\u8865<\/td>\n<\/tr>\n<tr>\n<td>parser_version<\/td>\n<td>str<\/td>\n<td>\u300c[09-\u68c0\u7d22\u7ba1\u9053\u95e8\u7981 \u8fd8\u5728\u5199]\u300d \u53d8\u66f4\u7559\u75d5<\/td>\n<td>\u7d22\u5f15\u91cd\u5efa\u540e\u8bf4\u4e0d\u6e05\u7528\u7684\u662f\u54ea\u4e00\u7248\u89e3\u6790\u5668<\/td>\n<td>\u62d2\u7edd\u5165\u5e93<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6700\u540e\u4e24\u4e2a\u5b57\u6bb5\u662f\u8fd9\u4e00\u8f6e\u8865\u7684&#xff1a;source_uri \u770b\u7740\u8ddf\u89e3\u6790\u65e0\u5173&#xff0c;\u4f46\u300c[07-\u77e5\u8bc6\u65b0\u9c9c\u5ea6 \u8fd8\u5728\u5199]\u300d \u7684\u4e09\u8def\u5de1\u68c0\u91cc\u300c\u6743\u5a01\u6e90\u6bd4\u5bf9\u300d\u8fd9\u4e00\u8def\u552f\u4e00\u9700\u8981\u7684\u8f93\u5165\u5c31\u662f\u5b83&#xff1b;\u7b49 07 \u518d\u56de\u5934\u8865&#xff0c;\u6210\u672c\u662f\u91cd\u65b0\u89e3\u6790\u6574\u4e2a\u5e93\u3002parser_version \u540c\u7406\u2014\u2014\u300c09-\u68c0\u7d22\u7ba1\u9053\u95e8\u7981 \u8fd8\u5728\u5199]\u300d \u7684\u56de\u6eda\u4e09\u6b65\u91cc&#xff0c;\u7b2c\u4e00\u6b65\u5c31\u8981\u56de\u7b54\u300c\u4e0a\u6b21\u662f\u7528\u54ea\u4e00\u7248\u89e3\u6790\u5668\u5efa\u7684\u7d22\u5f15\u300d\u3002<\/p>\n<p>\u6821\u9a8c\u8fd9\u4efd\u5951\u7ea6\u7684\u811a\u672c&#xff0c;\u96f6\u4f9d\u8d56&#xff1a;<\/p>\n<p>\u8f93\u5165\u662f\u4ec0\u4e48&#xff1a;parsed\/blocks.jsonl&#xff08;\u4e0a\u4e00\u8282\u90a3\u4efd\u5757\u6d41&#xff09;\u3002 \u600e\u4e48\u6267\u884c&#xff1a;\u9010\u5757\u68c0\u67e5\u5fc5\u586b\u5b57\u6bb5\u4e0e\u53d6\u503c\u5408\u6cd5\u6027&#xff0c;\u8fdd\u7ea6\u5206\u300c\u62d2\u7edd \/ \u964d\u7ea7 \/ \u8ba1\u6570\u300d\u4e09\u6863\u5904\u7f6e&#xff0c;\u8de8\u5757\u6c47\u603b\u6210\u4e00\u4efd\u8fdd\u7ea6\u62a5\u544a\u3002 \u8f93\u51fa\u4ec0\u4e48&#xff1a;\u6bcf\u4e2a\u5b57\u6bb5\u7684\u8fdd\u7ea6\u6570\u4e0e\u5904\u7f6e\u6863\u4f4d&#xff0c;\u5916\u52a0\u4e00\u53e5\u603b\u4f53\u5224\u5b9a\u3002 \u7ed3\u679c\u4e0d\u5bf9\u67e5\u54ea&#xff1a;\u5982\u679c btype \u8fdd\u7ea6\u6570\u662f\u603b\u5757\u6570\u7684\u5168\u90e8&#xff0c;\u8bf4\u660e\u89e3\u6790\u5668\u538b\u6839\u6ca1\u8f93\u51fa\u5757\u7c7b\u578b\u5b57\u6bb5&#xff0c;\u5148\u53bb\u770b\u89e3\u6790\u5668\u7684\u8f93\u51fa\u914d\u7f6e&#xff1b;\u5982\u679c rows \u8fdd\u7ea6\u6570\u7b49\u4e8e\u8868\u683c\u5757\u6570&#xff0c;\u90a3\u662f\u6ca1\u505a\u8868\u683c\u8fd8\u539f&#xff0c;\u4e0d\u662f\u683c\u5f0f\u95ee\u9898\u3002<\/p>\n<p><span class=\"token comment\"># block_contract.py \u2014\u2014 \u5757\u6d41\u5951\u7ea6\u6821\u9a8c&#xff08;\u96f6\u4f9d\u8d56\u5185\u6838&#xff0c;\u6807\u51c6\u5e93&#xff09;<\/span><br \/>\n<span class=\"token keyword\">import<\/span> json<br \/>\n<span class=\"token keyword\">from<\/span> collections <span class=\"token keyword\">import<\/span> Counter<\/p>\n<p>REQUIRED <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;doc_id&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;block_id&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;text&#034;<\/span><span class=\"token punctuation\">)<\/span>          <span class=\"token comment\"># \u7f3a\u5373\u62d2\u7edd<\/span><br \/>\nOPTIONAL_COUNT <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;page&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;section&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;source_uri&#034;<\/span><span class=\"token punctuation\">)<\/span>  <span class=\"token comment\"># \u7f3a\u5373\u8ba1\u6570<\/span><br \/>\nTYPES <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string\">&#034;text&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;table&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;title&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;caption&#034;<\/span><span class=\"token punctuation\">}<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">check_block<\/span><span class=\"token punctuation\">(<\/span>b<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;\u5355\u5757\u6821\u9a8c&#xff0c;\u8fd4\u56de {reject:bool, downgrade:list, missing:list}&#034;&#034;&#034;<\/span><br \/>\n    reject <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>k <span class=\"token keyword\">for<\/span> k <span class=\"token keyword\">in<\/span> REQUIRED <span class=\"token keyword\">if<\/span> <span class=\"token keyword\">not<\/span> b<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span>k<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    missing <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>k <span class=\"token keyword\">for<\/span> k <span class=\"token keyword\">in<\/span> OPTIONAL_COUNT <span class=\"token keyword\">if<\/span> b<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span>k<span class=\"token punctuation\">)<\/span> <span class=\"token keyword\">in<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token boolean\">None<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;&#034;<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    down <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    btype <span class=\"token operator\">&#061;<\/span> b<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;btype&#034;<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">if<\/span> btype <span class=\"token keyword\">not<\/span> <span class=\"token keyword\">in<\/span> TYPES<span class=\"token punctuation\">:<\/span><br \/>\n        down<span class=\"token punctuation\">.<\/span>append<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;btype-&gt;text&#034;<\/span><span class=\"token punctuation\">)<\/span>       <span class=\"token comment\"># \u7c7b\u578b\u4e0d\u8ba4\u8bc6&#xff0c;\u964d\u7ea7\u4e3a\u6b63\u6587<\/span><br \/>\n        btype <span class=\"token operator\">&#061;<\/span> <span class=\"token string\">&#034;text&#034;<\/span><br \/>\n    <span class=\"token keyword\">if<\/span> btype <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token string\">&#034;table&#034;<\/span> <span class=\"token keyword\">and<\/span> <span class=\"token keyword\">not<\/span> b<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;rows&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n        down<span class=\"token punctuation\">.<\/span>append<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;table_without_rows&#034;<\/span><span class=\"token punctuation\">)<\/span>  <span class=\"token comment\"># \u7ed3\u6784\u8fdd\u7ea6&#xff1a;\u58f0\u79f0\u662f\u8868\u683c\u5374\u6ca1\u6709\u884c\u5217<\/span><br \/>\n    <span class=\"token keyword\">if<\/span> <span class=\"token keyword\">not<\/span> b<span class=\"token punctuation\">.<\/span>get<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;parser_version&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n        reject<span class=\"token punctuation\">.<\/span>append<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;parser_version&#034;<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string\">&#034;reject&#034;<\/span><span class=\"token punctuation\">:<\/span> reject<span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;downgrade&#034;<\/span><span class=\"token punctuation\">:<\/span> down<span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;missing&#034;<\/span><span class=\"token punctuation\">:<\/span> missing<span class=\"token punctuation\">}<\/span><\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">report<\/span><span class=\"token punctuation\">(<\/span>path<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">str<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    n <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><br \/>\n    rejected<span class=\"token punctuation\">,<\/span> downgrades<span class=\"token punctuation\">,<\/span> miss <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">0<\/span><span class=\"token punctuation\">,<\/span> Counter<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> Counter<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> line <span class=\"token keyword\">in<\/span> <span class=\"token builtin\">open<\/span><span class=\"token punctuation\">(<\/span>path<span class=\"token punctuation\">,<\/span> encoding<span class=\"token operator\">&#061;<\/span><span class=\"token string\">&#034;utf-8&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> <span class=\"token keyword\">not<\/span> line<span class=\"token punctuation\">.<\/span>strip<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n            <span class=\"token keyword\">continue<\/span><br \/>\n        n <span class=\"token operator\">&#043;&#061;<\/span> <span class=\"token number\">1<\/span><br \/>\n        r <span class=\"token operator\">&#061;<\/span> check_block<span class=\"token punctuation\">(<\/span>json<span class=\"token punctuation\">.<\/span>loads<span class=\"token punctuation\">(<\/span>line<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><br \/>\n        <span class=\"token keyword\">if<\/span> r<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;reject&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">:<\/span><br \/>\n            rejected <span class=\"token operator\">&#043;&#061;<\/span> <span class=\"token number\">1<\/span><br \/>\n        downgrades<span class=\"token punctuation\">.<\/span>update<span class=\"token punctuation\">(<\/span>r<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;downgrade&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><br \/>\n        miss<span class=\"token punctuation\">.<\/span>update<span class=\"token punctuation\">(<\/span>r<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;missing&#034;<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string\">&#034;n_blocks&#034;<\/span><span class=\"token punctuation\">:<\/span> n<span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;reject_rate&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>rejected <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> n<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">3<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;downgrades&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">(<\/span>downgrades<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;missing_fields&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">(<\/span>miss<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;verdict&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token string\">&#034;\u53ef\u5165\u5e93&#034;<\/span> <span class=\"token keyword\">if<\/span> rejected <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token number\">0<\/span> <span class=\"token keyword\">and<\/span><br \/>\n                       miss<span class=\"token punctuation\">[<\/span><span class=\"token string\">&#034;section&#034;<\/span><span class=\"token punctuation\">]<\/span> <span class=\"token operator\">\/<\/span> <span class=\"token builtin\">max<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">1<\/span><span class=\"token punctuation\">,<\/span> n<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&lt;<\/span> <span class=\"token number\">0.10<\/span> <span class=\"token keyword\">else<\/span> <span class=\"token string\">&#034;\u9700\u4fee\u89e3\u6790\u914d\u7f6e&#034;<\/span><span class=\"token punctuation\">}<\/span><\/p>\n<p><span class=\"token keyword\">if<\/span> __name__ <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token string\">&#034;__main__&#034;<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token keyword\">print<\/span><span class=\"token punctuation\">(<\/span>json<span class=\"token punctuation\">.<\/span>dumps<span class=\"token punctuation\">(<\/span>report<span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;parsed\/blocks.jsonl&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> ensure_ascii<span class=\"token operator\">&#061;<\/span><span class=\"token boolean\">False<\/span><span class=\"token punctuation\">,<\/span> indent<span class=\"token operator\">&#061;<\/span><span class=\"token number\">2<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><\/p>\n<h4>\u9010\u6bb5\u8bfb\u8fd9\u4efd\u811a\u672c<\/h4>\n<ul>\n<li>REQUIRED \/ OPTIONAL_COUNT&#xff1a;\u4e24\u6863\u5b57\u6bb5\u5206\u5f00\u662f\u523b\u610f\u7684\u3002\u5fc5\u586b\u4e09\u6837\u7f3a\u4e00\u6837&#xff0c;\u8fd9\u6761\u5757\u5728\u4e0b\u6e38\u6ca1\u6709\u4efb\u4f55\u7528\u9014&#xff1b;\u53ef\u9009\u4e09\u6837\u7f3a\u4e86\u53ea\u662f\u8ba9\u67d0\u4e00\u73af\u53d8\u5f31&#xff0c;\u4e0d\u8be5\u56e0\u6b64\u4e22\u6389\u6574\u5757\u5185\u5bb9\u3002\u628a\u4e24\u8005\u6df7\u6210\u4e00\u4e2a\u6e05\u5355&#xff0c;\u6700\u5e38\u89c1\u7684\u540e\u679c\u662f section \u7a7a\u503c\u7387\u4e00\u9ad8\u5c31\u6574\u5e93\u62d2\u6536\u3002<\/li>\n<li>check_block()&#xff1a;btype \u4e0d\u8ba4\u8bc6\u5c31\u964d\u7ea7\u6210 text&#xff0c;\u800c\u4e0d\u662f\u62d2\u6536\u2014\u2014\u89e3\u6790\u5668\u7684\u7c7b\u578b\u8bcd\u8868\u4f1a\u968f\u7248\u672c\u6269\u5c55&#xff08;figure \/ code \/ list \u90fd\u53ef\u80fd\u5192\u51fa\u6765&#xff09;&#xff0c;\u62d2\u6536\u4f1a\u8ba9\u89e3\u6790\u5668\u5347\u7ea7\u53d8\u6210\u4e00\u6b21\u7ebf\u4e0a\u4e8b\u6545\u3002table \u5374\u6ca1\u6709 rows \u662f\u552f\u4e00\u8bb0\u300c\u7ed3\u6784\u8fdd\u7ea6\u300d\u7684\u60c5\u5f62&#xff0c;\u56e0\u4e3a\u5b83\u901a\u5e38\u610f\u5473\u7740\u6ca1\u505a\u8868\u683c\u8fd8\u539f\u3002<\/li>\n<li>report()&#xff1a;\u5224\u5b9a\u7ebf\u5199\u5728 verdict \u91cc\u2014\u2014\u96f6\u62d2\u6536\u4e14 section \u7a7a\u503c\u7387 &lt;10%&#xff08;\u4e0e\u7b2c\u4e8c\u8282\u90a3\u5f20\u8868\u7684\u9608\u503c\u5bf9\u9f50&#xff09;\u624d\u7b97\u53ef\u5165\u5e93\u3002\u8fd9\u6761\u7ebf\u662f\u8be5\u6309\u4f60\u81ea\u5df1\u7684\u6587\u6863\u96c6\u91cd\u6807\u7684&#xff0c;10% \u8fd9\u4e2a\u6570\u5bf9\u5e94\u300c\u5f15\u7528\u8fd8\u80fd\u5b9a\u4f4d\u5230\u7ae0\u8282\u300d\u3002<\/li>\n<li>\u5df2\u77e5\u504f\u5dee\u4e24\u6761&#xff1a;\u2460 page \u8bb0 -1 \u4e0e\u7f3a\u5b57\u6bb5\u90fd\u7b97\u8fdd\u7ea6&#xff0c;\u4f46\u8bed\u4e49\u4e0d\u540c&#xff08;\u524d\u8005\u662f\u300c\u77e5\u9053\u6ca1\u6709\u300d&#xff0c;\u540e\u8005\u662f\u300c\u6ca1\u53d6\u300d&#xff09;&#xff0c;\u8981\u533a\u5206\u5c31\u5f97\u5206\u5f00\u8ba1&#xff1b;\u2461 bbox \u6ca1\u8fdb\u6821\u9a8c&#xff0c;\u9700\u8981\u5750\u6807\u7ea6\u675f\u65f6&#xff08;\u540c\u9875\u591a\u5f20\u540c\u6784\u8868&#xff09;\u5f97\u81ea\u5df1\u8865\u3002<\/li>\n<\/ul>\n<p>\u8dd1\u5728 20 \u7bc7 runbooks \u7684\u6863 A \u4e0e\u6863 B \u8f93\u51fa\u4e0a&#xff0c;\u9884\u671f\u8f93\u51fa\u5982\u4e0b&#xff1a;<\/p>\n<p>\u6863 A: {&#034;n_blocks&#034;: 1842, &#034;reject_rate&#034;: 0.000, &#034;downgrades&#034;: {&#034;table_without_rows&#034;: 63},<br \/>\n       &#034;missing_fields&#034;: {&#034;page&#034;: 0, &#034;section&#034;: 1842, &#034;source_uri&#034;: 1842},<br \/>\n       &#034;verdict&#034;: &#034;\u9700\u4fee\u89e3\u6790\u914d\u7f6e&#034;}<br \/>\n\u6863 B: {&#034;n_blocks&#034;: 2176, &#034;reject_rate&#034;: 0.000, &#034;downgrades&#034;: {&#034;btype-&gt;text&#034;: 41},<br \/>\n       &#034;missing_fields&#034;: {&#034;page&#034;: 12, &#034;section&#034;: 96, &#034;source_uri&#034;: 0},<br \/>\n       &#034;verdict&#034;: &#034;\u53ef\u5165\u5e93&#034;}<\/p>\n<p>\u4e24\u884c\u8f93\u51fa\u7684\u5dee\u522b\u4e0d\u5728 reject_rate&#xff08;\u90fd\u662f 0&#xff09;&#xff0c;\u5728 missing_fields&#xff1a;\u6863 A \u7684 section \u4e0e source_uri \u662f\u6ee1\u683c\u8fdd\u7ea6 1,842 \u6761&#xff0c;\u6863 B \u7684 section \u53ea\u5269 96 \u6761&#xff08;\u7a7a\u503c\u7387 4.4% [\u7b97]&#xff0c;\u8d8a\u8fc7 10% \u90a3\u6761\u7ebf&#xff09;\u3002\u62d2\u6536\u7387\u770b\u4e0d\u51fa\u89e3\u6790\u5668\u7684\u597d\u574f&#xff0c;\u5b57\u6bb5\u8fdd\u7ea6\u5206\u5e03\u624d\u770b\u5f97\u51fa\u2014\u2014\u8fd9\u4e5f\u662f\u4e3a\u4ec0\u4e48\u5951\u7ea6\u6821\u9a8c\u8981\u5355\u72ec\u8dd1&#xff0c;\u4e0d\u80fd\u53ea\u770b\u56db\u628a\u5c3a\u5b50\u3002<\/p>\n<hr \/>\n<h3>\u4e03\u30010.912 \u548c 0.947 \u5230\u5e95\u5dee\u591a\u5c11&#xff1a;\u6362\u6863\u524d\u7684\u5224\u7a33\u4e0e\u91cd\u89e3\u6790\u8303\u56f4<\/h3>\n<p>\u4e0a\u4e00\u8282\u90a3\u5f20\u4e09\u6863\u8868\u6709\u4e2a\u6ca1\u56de\u7b54\u7684\u95ee\u9898&#xff1a;B \u7684\u5b57\u6bb5\u53ec\u56de\u7387 0.912 \u4e0e C \u7684 0.947&#xff0c;\u5dee 3.5pp [\u7b97]&#xff0c;\u591f\u4e0d\u591f\u683c\u591a\u4ed8 48 \u500d\u7684\u94b1&#xff1f;\u53ea\u770b\u5747\u503c\u4e00\u5b9a\u4f1a\u88ab\u9a97\u2014\u201420 \u7bc7\u91cc\u53ea\u8981\u6709\u4e00\u4e24\u7bc7\u542b\u5408\u5e76\u5355\u5143\u683c\u7684\u6587\u6863\u88ab C \u6551\u56de\u6765&#xff0c;\u5747\u503c\u5c31\u88ab\u62ac\u4e0a\u53bb\u4e86\u3002<\/p>\n<p>\u5224\u7a33\u7684\u529e\u6cd5\u662f\u91cd\u91c7\u6837&#xff1a;\u628a 20 \u7bc7\u7684\u5206\u6570\u5f53\u6210\u6837\u672c&#xff0c;\u6709\u653e\u56de\u5730\u62bd 2,000 \u6b21&#xff0c;\u770b\u5dee\u503c\u7684\u5206\u5e03\u8de8\u4e0d\u8de8 0\u3002\u96f6\u4f9d\u8d56\u5b9e\u73b0&#xff1a;<\/p>\n<p>\u8f93\u5165\u662f\u4ec0\u4e48&#xff1a;\u4e24\u6863\u5404\u81ea\u7684\u5206\u7bc7\u5206\u6570\u5217\u8868&#xff08;\u957f\u5ea6\u76f8\u540c\u3001\u6309\u7bc7\u5bf9\u9f50&#xff09;\u3002 \u600e\u4e48\u6267\u884c&#xff1a;\u6210\u5bf9\u76f8\u51cf\u540e\u505a bootstrap&#xff0c;\u8f93\u51fa\u5dee\u503c\u7684\u5747\u503c\u4e0e 95% \u5206\u4f4d\u533a\u95f4\u3002 \u8f93\u51fa\u4ec0\u4e48&#xff1a;\u70b9\u4f30\u8ba1 &#043; \u533a\u95f4\u4e0b\u4e0a\u754c &#043; \u4e00\u53e5\u5224\u5b9a\u3002 \u7ed3\u679c\u4e0d\u5bf9\u67e5\u54ea&#xff1a;\u533a\u95f4\u7279\u522b\u5bbd&#xff08;\u534a\u5bbd &gt;5pp&#xff09;\u901a\u5e38\u4e0d\u662f\u65b9\u6cd5\u95ee\u9898&#xff0c;\u662f\u5206\u7bc7\u5206\u6570\u65b9\u5dee\u592a\u5927&#xff0c;\u5148\u770b\u662f\u4e0d\u662f\u6709\u4e00\u4e24\u7bc7\u79bb\u7fa4&#xff1b;\u6210\u5bf9\u8bbe\u8ba1\u4e0b\u5fc5\u987b\u4e24\u6863\u7528\u540c\u4e00\u6279\u6587\u6863&#xff0c;\u6df7\u4e86\u5c31\u5931\u53bb\u914d\u5bf9\u7684\u610f\u4e49\u3002<\/p>\n<p><span class=\"token comment\"># bootstrap_diff.py \u2014\u2014 \u6210\u5bf9 bootstrap \u5224\u7a33&#xff08;\u96f6\u4f9d\u8d56\u5185\u6838&#xff0c;\u6807\u51c6\u5e93&#xff09;<\/span><br \/>\n<span class=\"token keyword\">import<\/span> random<\/p>\n<p><span class=\"token keyword\">def<\/span> <span class=\"token function\">bootstrap_diff<\/span><span class=\"token punctuation\">(<\/span>a<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">,<\/span> b<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">list<\/span><span class=\"token punctuation\">,<\/span> n_boot<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">int<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">2000<\/span><span class=\"token punctuation\">,<\/span> seed<span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">int<\/span> <span class=\"token operator\">&#061;<\/span> <span class=\"token number\">7<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span><span class=\"token operator\">&gt;<\/span> <span class=\"token builtin\">dict<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token triple-quoted-string string\">&#034;&#034;&#034;a&#061;\u65e7\u6863\u5206\u7bc7\u5206\u6570&#xff0c;b&#061;\u65b0\u6863\u5206\u7bc7\u5206\u6570&#xff0c;\u8fd4\u56de\u5dee\u503c(b-a)\u7684\u70b9\u4f30\u8ba1\u4e0e95%\u533a\u95f4\u3002<br \/>\n    \u6210\u5bf9\u8bbe\u8ba1&#xff1a;\u540c\u4e00\u7bc7\u6587\u6863\u5728\u4e24\u6863\u4e0b\u7684\u5206\u6570\u914d\u5bf9\u76f8\u51cf&#xff0c;\u6d88\u9664\u6587\u6863\u96be\u5ea6\u5dee\u5f02\u3002&#034;&#034;&#034;<\/span><br \/>\n    <span class=\"token keyword\">assert<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>a<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>b<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token string\">&#034;\u5fc5\u987b\u6210\u5bf9&#xff1a;\u4e24\u6863\u8dd1\u540c\u4e00\u6279\u6587\u6863&#034;<\/span><br \/>\n    rnd <span class=\"token operator\">&#061;<\/span> random<span class=\"token punctuation\">.<\/span>Random<span class=\"token punctuation\">(<\/span>seed<span class=\"token punctuation\">)<\/span><br \/>\n    n <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">len<\/span><span class=\"token punctuation\">(<\/span>a<span class=\"token punctuation\">)<\/span><br \/>\n    d <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span>y <span class=\"token operator\">&#8211;<\/span> x <span class=\"token keyword\">for<\/span> x<span class=\"token punctuation\">,<\/span> y <span class=\"token keyword\">in<\/span> <span class=\"token builtin\">zip<\/span><span class=\"token punctuation\">(<\/span>a<span class=\"token punctuation\">,<\/span> b<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span>          <span class=\"token comment\"># \u9010\u7bc7\u5dee\u503c<\/span><br \/>\n    point <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">sum<\/span><span class=\"token punctuation\">(<\/span>d<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">\/<\/span> n<br \/>\n    boots <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    <span class=\"token keyword\">for<\/span> _ <span class=\"token keyword\">in<\/span> <span class=\"token builtin\">range<\/span><span class=\"token punctuation\">(<\/span>n_boot<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">:<\/span><br \/>\n        s <span class=\"token operator\">&#061;<\/span> <span class=\"token builtin\">sum<\/span><span class=\"token punctuation\">(<\/span>d<span class=\"token punctuation\">[<\/span>rnd<span class=\"token punctuation\">.<\/span>randrange<span class=\"token punctuation\">(<\/span>n<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span> <span class=\"token keyword\">for<\/span> _ <span class=\"token keyword\">in<\/span> <span class=\"token builtin\">range<\/span><span class=\"token punctuation\">(<\/span>n<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span> <span class=\"token operator\">\/<\/span> n<br \/>\n        boots<span class=\"token punctuation\">.<\/span>append<span class=\"token punctuation\">(<\/span>s<span class=\"token punctuation\">)<\/span><br \/>\n    boots<span class=\"token punctuation\">.<\/span>sort<span class=\"token punctuation\">(<\/span><span class=\"token punctuation\">)<\/span><br \/>\n    lo <span class=\"token operator\">&#061;<\/span> boots<span class=\"token punctuation\">[<\/span><span class=\"token builtin\">int<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">0.025<\/span> <span class=\"token operator\">*<\/span> n_boot<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    hi <span class=\"token operator\">&#061;<\/span> boots<span class=\"token punctuation\">[<\/span><span class=\"token builtin\">int<\/span><span class=\"token punctuation\">(<\/span><span class=\"token number\">0.975<\/span> <span class=\"token operator\">*<\/span> n_boot<span class=\"token punctuation\">)<\/span> <span class=\"token operator\">&#8211;<\/span> <span class=\"token number\">1<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    <span class=\"token keyword\">return<\/span> <span class=\"token punctuation\">{<\/span><span class=\"token string\">&#034;point_pp&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>point <span class=\"token operator\">*<\/span> <span class=\"token number\">100<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">2<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;ci95_pp&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>lo <span class=\"token operator\">*<\/span> <span class=\"token number\">100<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">2<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token builtin\">round<\/span><span class=\"token punctuation\">(<\/span>hi <span class=\"token operator\">*<\/span> <span class=\"token number\">100<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">2<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">]<\/span><span class=\"token punctuation\">,<\/span><br \/>\n            <span class=\"token string\">&#034;verdict&#034;<\/span><span class=\"token punctuation\">:<\/span> <span class=\"token punctuation\">(<\/span><span class=\"token string\">&#034;\u6362&#034;<\/span> <span class=\"token keyword\">if<\/span> lo <span class=\"token operator\">&gt;<\/span> <span class=\"token number\">0.02<\/span> <span class=\"token keyword\">else<\/span><br \/>\n                        <span class=\"token string\">&#034;\u4e0d\u6362&#034;<\/span> <span class=\"token keyword\">if<\/span> hi <span class=\"token operator\">&lt;<\/span> <span class=\"token number\">0<\/span> <span class=\"token keyword\">else<\/span><br \/>\n                        <span class=\"token string\">&#034;\u4e0d\u53ef\u533a\u5206&#034;<\/span><span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">}<\/span><\/p>\n<p><span class=\"token keyword\">if<\/span> __name__ <span class=\"token operator\">&#061;&#061;<\/span> <span class=\"token string\">&#034;__main__&#034;<\/span><span class=\"token punctuation\">:<\/span><br \/>\n    <span class=\"token comment\"># 20 \u7bc7\u5206\u7bc7\u51c6\u786e\u7387&#xff08;\u26a0\ufe0f \u6784\u9020\u793a\u4f8b&#xff09;&#xff1a;B \u6863 vs C \u6863<\/span><br \/>\n    B <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.65<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.80<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.55<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.75<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.60<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.85<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.65<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.75<\/span><span class=\"token punctuation\">,<\/span><br \/>\n         <span class=\"token number\">0.60<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.80<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.55<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.65<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.75<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.60<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.85<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    C <span class=\"token operator\">&#061;<\/span> <span class=\"token punctuation\">[<\/span><span class=\"token number\">0.75<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.65<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.80<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.60<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.75<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.60<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.85<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.80<\/span><span class=\"token punctuation\">,<\/span><br \/>\n         <span class=\"token number\">0.60<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.85<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.55<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.75<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.65<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.85<\/span><span class=\"token punctuation\">,<\/span> <span class=\"token number\">0.70<\/span><span class=\"token punctuation\">]<\/span><br \/>\n    <span class=\"token keyword\">print<\/span><span class=\"token punctuation\">(<\/span>bootstrap_diff<span class=\"token punctuation\">(<\/span>B<span class=\"token punctuation\">,<\/span> C<span class=\"token punctuation\">)<\/span><span class=\"token punctuation\">)<\/span><\/p>\n<h4>\u9010\u6bb5\u8bfb\u8fd9\u4efd\u811a\u672c<\/h4>\n<ul>\n<li>\u6210\u5bf9\u8bbe\u8ba1&#xff1a;\u5148\u9010\u7bc7\u76f8\u51cf\u518d\u91cd\u91c7\u6837&#xff0c;\u800c\u4e0d\u662f\u5206\u522b\u5bf9\u4e24\u6863\u91cd\u91c7\u6837\u518d\u76f8\u51cf\u3002\u524d\u8005\u6d88\u6389\u4e86\u300c\u8fd9\u7bc7\u6587\u6863\u672c\u6765\u5c31\u96be\u300d\u8fd9\u4e2a\u5171\u540c\u56e0\u7d20&#xff0c;\u533a\u95f4\u4f1a\u7a84\u5f88\u591a\u3002\u4e24\u6863\u5fc5\u987b\u8dd1\u540c\u4e00\u6279\u6587\u6863&#xff0c;\u7bc7\u6570\u4e0d\u7b49\u65f6\u76f4\u63a5\u65ad\u8a00\u5931\u8d25\u3002<\/li>\n<li>seed&#061;7&#xff1a;\u56fa\u5b9a\u79cd\u5b50\u662f\u4e3a\u4e86\u8ba9\u7ed3\u8bba\u53ef\u590d\u73b0\u3002\u6539\u79cd\u5b50\u4f1a\u8ba9\u533a\u95f4\u6296\u52a8\u96f6\u70b9\u51e0\u4e2a pp&#xff0c;\u4f46\u8de8\u4e0d\u8de8 0 \u8fd9\u4e2a\u5224\u5b9a\u901a\u5e38\u4e0d\u53d8\u2014\u2014\u5982\u679c\u53d8\u4e86&#xff0c;\u8bf4\u660e\u6837\u672c\u91cf\u4e0d\u591f&#xff0c;\u522b\u6025\u7740\u4e0b\u7ed3\u8bba\u3002<\/li>\n<li>verdict \u7684\u4e09\u5206\u652f&#xff1a;\u4e0b\u754c &gt;2pp \u624d\u5224\u300c\u6362\u300d&#xff1b;\u4e0a\u754c &lt;0 \u5224\u300c\u4e0d\u6362\u300d&#xff08;\u65b0\u6863\u66f4\u5dee&#xff09;&#xff1b;\u5176\u4f59\u662f\u300c\u4e0d\u53ef\u533a\u5206\u300d\u30022pp \u8fd9\u6761\u7ebf\u6765\u81ea 20 \u7bc7\u7684\u7f6e\u4fe1\u534a\u5bbd \u00b13.5pp [\u7b97] \u4e0e\u300c\u4e0b\u6e38\u6536\u76ca\u8981\u76d6\u5f97\u4f4f\u6210\u672c\u589e\u91cf\u300d\u4e24\u4e2a\u7ea6\u675f&#xff0c;\u4e0d\u662f\u7edf\u8ba1\u663e\u8457\u6027\u9608\u503c\u3002<\/li>\n<li>\u5df2\u77e5\u504f\u5dee\u4e24\u6761&#xff1a;\u2460 bootstrap \u5047\u8bbe\u5206\u7bc7\u72ec\u7acb&#xff0c;\u540c\u6e90\u6587\u6863&#xff08;\u540c\u4e00\u4efd runbook \u7684\u591a\u4e2a\u7248\u672c&#xff09;\u4f1a\u8ba9\u533a\u95f4\u504f\u7a84&#xff1b;\u2461 20 \u7bc7\u7684\u6837\u672c\u91cf\u4e0b&#xff0c;&lt;2pp \u7684\u771f\u5b9e\u5dee\u503c\u57fa\u672c\u68c0\u4e0d\u51fa\u6765&#xff0c;\u9700\u8981\u68c0\u51fa\u6765\u5c31\u5f97\u6309\u7b2c\u56db\u8282\u90a3\u4e2a\u7b97\u6cd5\u628a\u6837\u672c\u52a0\u5230 63 \u7bc7\u3002<\/li>\n<\/ul>\n<p>\u9884\u671f\u8f93\u51fa&#xff1a;<\/p>\n<p>{&#034;point_pp&#034;: 1.00, &#034;ci95_pp&#034;: [-2.40, 4.60], &#034;verdict&#034;: &#034;\u4e0d\u53ef\u533a\u5206&#034;}<\/p>\n<p>\u70b9\u4f30\u8ba1 &#043;1.0pp&#xff0c;\u533a\u95f4\u8de8 0 \u2192 \u5224\u4e0d\u53ef\u533a\u5206 \u2192 \u9009\u4fbf\u5b9c\u7684 B\u3002\u8fd9\u5c31\u662f\u7b2c\u56db\u8282\u90a3\u53e5\u300c\u4e09\u6863\u4e0b\u6e38\u5dee\u503c\u5c0f\u4e8e 2pp \u5c31\u9009\u4fbf\u5b9c\u7684\u90a3\u6863\u300d\u7684\u6765\u5904&#xff1a;\u4e0d\u662f\u62cd\u7684&#xff0c;\u662f\u91cd\u91c7\u6837\u7b97\u51fa\u6765\u7684\u3002<\/p>\n<p>\u6362\u6863\u4e4b\u540e\u8fd8\u6709\u4e00\u4ef6\u4e8b&#xff1a;\u4ec0\u4e48\u65f6\u5019\u91cd\u89e3\u6790\u3002\u5168\u91cf\u91cd\u89e3\u6790 6,000 \u9875\u8981 30 \u5206\u949f\u5230 5 \u5c0f\u65f6&#xff08;\u7b2c\u4e94\u8282\u90a3\u5f20\u8868&#xff09;&#xff0c;\u6bcf\u6b21\u89e3\u6790\u5668\u5347\u7ea7\u90fd\u8dd1\u4e00\u904d\u4e0d\u73b0\u5b9e\u3002\u89e6\u53d1\u6761\u4ef6\u7528\u53cc\u952e (doc_hash, parser_version) \u5224&#xff1a;<\/p>\n<table>\n<tr>\u53d8\u66f4\u7c7b\u578b\u5224\u5b9a\u952e\u91cd\u89e3\u6790\u8303\u56f46,000 \u9875\u4e0b\u7684\u9884\u671f\u5360\u6bd4 \u26a0\ufe0f \u7ecf\u9a8c\u503c<\/tr>\n<tbody>\n<tr>\n<td>\u6587\u6863\u5185\u5bb9\u66f4\u65b0<\/td>\n<td>doc_hash \u53d8\u3001parser_version \u4e0d\u53d8<\/td>\n<td>\u53ea\u91cd\u8dd1\u53d8\u4e86\u7684\u6587\u6863<\/td>\n<td>\u6bcf\u5468 0.5%\u20132%<\/td>\n<\/tr>\n<tr>\n<td>\u89e3\u6790\u5668\u7248\u672c\u5347\u7ea7<\/td>\n<td>parser_version \u53d8<\/td>\n<td>\u5168\u91cf\u91cd\u8dd1<\/td>\n<td>100%<\/td>\n<\/tr>\n<tr>\n<td>\u89e3\u6790\u5668\u914d\u7f6e\u8c03\u6574&#xff08;\u9608\u503c\/\u5206\u680f\u5f00\u5173&#xff09;<\/td>\n<td>parser_version \u52a0\u540e\u7f00<\/td>\n<td>\u5168\u91cf\u91cd\u8dd1&#xff0c;\u53ef\u8d70\u79bb\u7ebf\u961f\u5217<\/td>\n<td>100%<\/td>\n<\/tr>\n<tr>\n<td>\u53ea\u5bf9\u96be\u4f8b\u5347\u6863&#xff08;\u7ed3\u6784\u4fdd\u6301\u7387 &lt;0.5 \u7684\u9875&#xff09;<\/td>\n<td>\u6309\u9875\u6253\u6807<\/td>\n<td>\u53ea\u91cd\u8dd1\u88ab\u6807\u9875<\/td>\n<td>3%\u20138%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u6700\u540e\u4e00\u884c\u7684\u300c\u53ea\u5bf9\u96be\u4f8b\u5347\u6863\u300d\u662f\u8fd9\u4e09\u5f20\u8868\u91cc\u6700\u7701\u94b1\u7684\u4e00\u62db&#xff1a;\u5168\u91cf\u5347 C \u8981\u4ed8 48 \u500d\u6210\u672c&#xff0c;\u53ea\u7ed9\u7ed3\u6784\u4fdd\u6301\u7387 &lt;0.5 \u7684\u90a3 3%\u20138% \u7684\u9875\u5347 C&#xff0c;\u589e\u91cf\u6210\u672c\u53ea\u6709\u5168\u91cf\u7684 1.5\u20134 \u500d [\u7b97]&#xff0c;\u800c\u7ed3\u6784\u4fdd\u6301\u7387\u7684\u62ac\u5347\u51e0\u4e4e\u5168\u90e8\u6765\u81ea\u8fd9\u6279\u9875\u3002\u5224\u5b9a\u952e\u8981\u5199\u8fdb\u7d22\u5f15\u5143\u6570\u636e&#xff0c;\u300c[09-\u68c0\u7d22\u7ba1\u9053\u95e8\u7981 \u8fd8\u5728\u5199]\u300d \u7684\u56de\u6eda\u7b2c\u4e00\u6b65\u9760\u5b83\u3002<\/p>\n<hr \/>\n<h3>\u516b\u3001\u4ec0\u4e48\u65f6\u5019\u4e0d\u8be5\u4e0a VLM \u548c OCR&#xff1f;<\/h3>\n<table>\n<tr>\u60c5\u5f62\u4e3a\u4ec0\u4e48\u4e0d\u8be5\u8be5\u505a\u4ec0\u4e48<\/tr>\n<tbody>\n<tr>\n<td>PDF \u6709\u6587\u672c\u5c42\u4e14\u5b57\u7b26\u6570\u6b63\u5e38<\/td>\n<td>OCR \u4f1a\u628a\u5df2\u6b63\u786e\u7684\u5b57\u7b26\u91cd\u8bc6\u522b\u4e00\u904d&#xff0c;\u5b57\u6bb5\u53ec\u56de\u7387\u4e0d\u5347\u53cd\u964d<\/td>\n<td>\u5148\u8dd1\u7a7a\u9875\u7387\u68c0\u6d4b&#xff0c;\u53ea\u5bf9\u7a7a\u9875 OCR<\/td>\n<\/tr>\n<tr>\n<td>\u6587\u6863\u4ee5\u7eaf\u6b63\u6587\u4e3a\u4e3b&#xff0c;\u8868\u683c\u5bc6\u5ea6 &lt;5%<\/td>\n<td>\u6863 B \u7684\u7ed3\u6784\u589e\u76ca\u7528\u4e0d\u4e0a&#xff0c;\u6863 C \u66f4\u7528\u4e0d\u4e0a<\/td>\n<td>\u6863 A &#043; \u62bd\u68c0&#xff0c;\u7701\u4e0b\u7684\u9884\u7b97\u653e\u5230 \u300c[03-\u5207\u5206-\u8bed\u4e49\u5b8c\u6574\u4e0e\u68c0\u7d22\u7cbe\u5ea6 \u8fd8\u5728\u5199]\u300d<\/td>\n<\/tr>\n<tr>\n<td>\u6587\u6863\u91cf &gt;10 \u4e07\u9875&#xff0c;\u6216\u7248\u5f0f\u4ecd\u5728\u9891\u7e41\u53d8\u66f4<\/td>\n<td>\u5168\u91cf C \u7684\u6210\u672c\u4f1a\u6324\u6389\u522b\u7684\u73af\u8282&#xff0c;\u57fa\u7ebf\u672a\u7a33\u65f6\u5347\u6863\u6536\u4e0d\u56de<\/td>\n<td>\u6863 B \u5168\u91cf &#043; \u6863 C \u53ea\u8dd1\u7ed3\u6784\u4fdd\u6301\u7387 &lt;0.5 \u7684\u9875<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u4e00\u7ec4\u5bf9\u7167&#xff1a;<\/p>\n<p>\u274c \u300c\u6574\u5e93 12,000 \u6761\u90fd\u4e0a VLM&#xff0c;\u4e00\u6b65\u5230\u4f4d\u300d    \u2014\u2014 \u5168\u91cf\u6210\u672c\u6362\u6765\u7684\u589e\u91cf\u96c6\u4e2d\u5728\u4e0d\u5230 10% \u7684\u9875\u9762<br \/>\n\u2705 \u300c\u6863 B \u5168\u91cf&#xff0c;\u7ed3\u6784\u4fdd\u6301\u7387 &lt;0.5 \u7684\u9875\u91cd\u8dd1\u6863 C\u300d \u2014\u2014 \u628a\u94b1\u82b1\u5728\u786e\u5b9e\u96be\u7684\u90a3\u4e9b\u9875\u4e0a<\/p>\n<p>\u8fd8\u6709\u4e00\u79cd\u60c5\u5f62\u5bb9\u6613\u88ab\u5ffd\u7565&#xff1a;\u95ee\u9898\u53ea\u95ee\u300c\u8fd9\u6bb5\u8bdd\u5728\u4e0d\u5728\u300d&#xff0c;\u4e0d\u8981\u6c42\u7cbe\u786e\u53d6\u503c\u3002\u8fd9\u65f6\u7ed3\u6784\u4e0d\u662f\u74f6\u9888&#xff0c;\u53ec\u56de\u624d\u662f&#xff0c;\u5347\u6863\u7684\u94b1\u4e0d\u5982\u653e\u5230\u300c [05-\u6df7\u5408\u68c0\u7d22\u4e0e\u91cd\u6392 &#8211; \u8fd9\u4e2a\u4e5f\u8fd8\u5728\u5199] \u300d\u4e0a\u3002<\/p>\n<p>\u300c\u5148\u91cf\u518d\u9009\u300d\u4e5f\u6709\u4e0d\u5212\u7b97\u7684\u65f6\u5019&#xff1a;\u6587\u6863\u603b\u91cf &lt;50 \u7bc7\u4e14\u7248\u5f0f\u5355\u4e00&#xff0c;\u76f4\u63a5\u9009 B \u8dd1\u5b8c\u770b\u4e00\u773c\u8f93\u51fa\u5373\u53ef&#xff0c;\u5efa gold \u6807\u6ce8\u7684\u6210\u672c\u6bd4\u5347\u6863\u7701\u4e0b\u7684\u94b1\u8fd8\u591a\u3002<\/p>\n<hr \/>\n<h3>\u4e5d\u3001\u5c0f\u7ed3<\/h3>\n<p>\u5168\u7bc7\u603b\u7ed3&#xff1a;\u89e3\u6790\u8d28\u91cf\u7684\u5ea6\u91cf\u5bf9\u8c61\u4e0d\u662f\u300c\u5b57\u5bf9\u4e0d\u5bf9\u300d&#xff0c;\u662f\u300c\u7ed3\u6784\u5728\u4e0d\u5728\u300d\u3002 \u5b57\u9519\u4e86\u6709\u544a\u8b66&#xff0c;\u7ed3\u6784\u6ca1\u4e86\u6ca1\u6709\u544a\u8b66&#xff0c;\u53ea\u5728\u53ec\u56de\u6392\u5e8f\u91cc\u5b89\u9759\u5730\u6263\u6389\u51e0\u4e2a\u767e\u5206\u70b9\u3002<\/p>\n<p>\u56db\u6761\u5224\u636e&#xff1a;<\/p>\n<li>\u4e94\u73af\u5404\u88c5\u4e00\u4e2a\u68c0\u6d4b\u70b9&#xff08;\u7a7a\u9875\u7387 \/ \u6807\u9898\u5c42\u7ea7 \/ \u884c\u5217\u5bf9\u9f50 \/ \u8de8\u9875\u7eed\u63a5 \/ \u9875\u7801\u8def\u5f84&#xff09;&#xff0c;\u7f3a\u54ea\u4e00\u73af\u7684\u68c0\u6d4b&#xff0c;\u54ea\u4e00\u73af\u7684\u635f\u5931\u5c31\u770b\u4e0d\u89c1\u3002<\/li>\n<li>\u56db\u628a\u5c3a\u5b50\u91cc&#xff0c;\u7ed3\u6784\u4fdd\u6301\u7387\u4e0e\u4e0b\u6e38\u51c6\u786e\u7387\u7684\u76f8\u5173\u6027\u6700\u5f3a&#xff0c;\u5b57\u6bb5\u53ec\u56de\u7387\u6b21\u4e4b&#xff0c;\u5355\u4f4d\u6210\u672c\u662f\u7ea6\u675f\u800c\u975e\u76ee\u6807\u3002<\/li>\n<li>\u9009\u578b\u6309\u300c\u5148\u91cf\u518d\u9009\u300d&#xff1a;20 \u7bc7\u81ea\u5df1\u7684\u6587\u6863\u8dd1\u51fa\u56db\u628a\u5c3a\u5b50\u518d\u5b9a\u6863&#xff0c;\u522b\u6309\u6587\u6863\u7c7b\u578b\u731c&#xff1b;\u6362\u6863\u524d\u518d\u8865\u4e00\u6b21\u6210\u5bf9 bootstrap&#xff0c;\u533a\u95f4\u8de8 0 \u5c31\u522b\u6362\u3002<\/li>\n<li>\u5951\u7ea6\u6821\u9a8c\u4e0e\u56db\u628a\u5c3a\u5b50\u5206\u5f00\u8dd1&#xff1a;\u62d2\u6536\u7387\u770b\u4e0d\u51fa\u89e3\u6790\u5668\u597d\u574f&#xff0c;section \/ source_uri \/ rows \u7684\u8fdd\u7ea6\u5206\u5e03\u624d\u770b\u5f97\u51fa\u3002<\/li>\n<li><\/li>\n<p>\u4e00\u4e2a\u5341\u5206\u949f\u80fd\u505a\u5b8c\u7684\u6478\u5e95&#xff1a;\u4ece\u77e5\u8bc6\u5e93\u91cc\u6311\u4e00\u4efd\u6709\u8868\u683c\u7684\u6587\u6863&#xff0c;\u7528\u73b0\u5728\u7684\u89e3\u6790\u5668\u8dd1\u4e00\u904d&#xff0c;\u6570\u4e24\u4ef6\u4e8b\u2014\u2014\u8f93\u51fa\u91cc rows \u975e\u7a7a\u7684\u5757\u6709\u51e0\u4e2a&#xff0c;section \u975e\u7a7a\u7684\u5757\u5360\u6bd4\u662f\u591a\u5c11\u3002\u524d\u8005\u662f 0&#xff0c;\u8bf4\u660e\u538b\u6839\u6ca1\u505a\u8868\u683c\u8fd8\u539f&#xff0c;\u540e\u9762\u6240\u6709\u8c03\u53c2\u90fd\u662f\u5728\u6ca1\u6709\u7ed3\u6784\u7684\u6587\u672c\u4e0a\u505a&#xff1b;\u540e\u8005\u4f4e\u4e8e 90%&#xff0c;\u300c[06-\u5f15\u7528\u4e0e\u53ef\u6eaf\u6e90 \u8fd8\u5728\u5199]\u300d \u7684\u5f15\u7528\u5b9a\u4f4d\u8fdf\u65e9\u8981\u8fd4\u5de5&#xff0c;\u8fd4\u5de5\u5bf9\u8c61\u662f\u6574\u6761\u7d22\u5f15\u3002\u8fd9\u4e24\u4e2a\u6570\u4e5f\u51b3\u5b9a\u4e86\u4f60\u8981\u4e0d\u8981\u7ee7\u7eed\u8bfb \u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d&#xff0c;\u4ee5\u53ca\u8bfb\u5b8c\u4e4b\u540e\u8be5\u5347\u6863\u8fd8\u662f\u8be5\u5148\u628a\u5951\u7ea6\u8865\u9f50\u3002<\/p>\n<p>\u4e0b\u4e00\u7bc7&#xff1a;\u300c02B-\u6587\u6863\u89e3\u6790\u4e0b-\u8868\u683c\u8de8\u9875\u4e0e\u622a\u65ad\u300d&#xff0c;\u89e3\u51b3\u300c\u8868\u683c\u88ab\u62c9\u5e73\u3001\u6bb5\u843d\u88ab\u5207\u65ad\u300d\u8fd9\u4e00\u7c7b\u2014\u2014\u4e94\u6761\u65ad\u88c2\u68c0\u6d4b\u89c4\u5219\u91cc&#xff0c;\u7b2c\u4e00\u6761\u5c31\u80fd\u628a\u672c\u6587 cross_block 0.742 \u8fd9\u4e2a\u6570\u62ac\u4e0a\u53bb\u3002<\/p>\n<hr \/>\n<h4>\u53c2\u8003<\/h4>\n<ul>\n<li>pdfplumber&#xff08;\u6863 A \u6837\u4f8b\u5b9e\u73b0&#xff0c;\u652f\u6491\u300c\u6709\u6587\u672c\u5c42\u4e0d\u5fc5 OCR\u300d&#xff09;<\/li>\n<li>PyMuPDF&#xff08;\u89e3\u7801\u73af\u8282\u7684\u6587\u672c\u5c42\u53d6\u7528&#xff0c;\u652f\u6491\u7a7a\u9875\u7387\u68c0\u6d4b\u70b9&#xff09;<\/li>\n<li>Unstructured \u00b7 Marker \u00b7 MinerU<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>\u6587\u6863\u89e3\u6790\u4e22\u7684\u4e0d\u662f\u5b57&#xff0c;\u662f\u7ed3\u6784&#xff1a;\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u628a\u5c3a\u5b50 \u7cfb\u5217\u5bfc\u822a&#xff1a;00 \u7cfb\u5217\u5bfc\u822a \u00b7 01 \u56db\u7c7b\u77e5\u8bc6 \u00b7 \u672c\u7bc7&#xff08;02A&#xff09; \u00b7 02B \u65ad\u88c2\u4fee\u590d \u00b7 03 \u5207\u5206\u6807\u5b9a \u00b7 04 \u67e5\u8be2\u7406\u89e3 \u00b7 05 \u6df7\u5408\u68c0\u7d22\u4e0e\u91cd\u6392 \u00b7 06 \u5f15\u7528\u9a8c\u8bc1 \u00b7 07 \u77e5\u8bc6\u65b0\u9c9c\u5ea6 \u00b7 08 \u591a\u6e90\u4e00\u81f4\u6027 \u00b7 09 \u7ba1\u9053\u95e8\u7981 \u5c0f\u63d0\u4e00\u53e5&#xff1a; \u7531\u4e8e\u5e73\u53f0\u6bcf\u65e5\u53d1\u5e03\u7b14\u8bb0\u6570\u91cf\u9650\u5236&#xff0c;\u76ee\u524d\u7cfb\u5217\u7b14\u8bb0\u8fd8\u6ca1\u5168\u90e8\u4e0a\u4f20\u3002\u540e\u7eed\u5185\u5bb9\u4f1a\u6301\u7eed\u66f4\u65b0&amp;#x<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[53,81,50,207,51],"topic":[],"class_list":["post-119483","post","type-post","status-publish","format-standard","hentry","category-server","tag-ai","tag-python","tag-50","tag-207","tag-51"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v20.3 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/119483.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u6587\u6863\u89e3\u6790\u4e22\u7684\u4e0d\u662f\u5b57&#xff0c;\u662f\u7ed3\u6784&#xff1a;\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u628a\u5c3a\u5b50 \u7cfb\u5217\u5bfc\u822a&#xff1a;00 \u7cfb\u5217\u5bfc\u822a \u00b7 01 \u56db\u7c7b\u77e5\u8bc6 \u00b7 \u672c\u7bc7&#xff08;02A&#xff09; \u00b7 02B \u65ad\u88c2\u4fee\u590d \u00b7 03 \u5207\u5206\u6807\u5b9a \u00b7 04 \u67e5\u8be2\u7406\u89e3 \u00b7 05 \u6df7\u5408\u68c0\u7d22\u4e0e\u91cd\u6392 \u00b7 06 \u5f15\u7528\u9a8c\u8bc1 \u00b7 07 \u77e5\u8bc6\u65b0\u9c9c\u5ea6 \u00b7 08 \u591a\u6e90\u4e00\u81f4\u6027 \u00b7 09 \u7ba1\u9053\u95e8\u7981 \u5c0f\u63d0\u4e00\u53e5&#xff1a; \u7531\u4e8e\u5e73\u53f0\u6bcf\u65e5\u53d1\u5e03\u7b14\u8bb0\u6570\u91cf\u9650\u5236&#xff0c;\u76ee\u524d\u7cfb\u5217\u7b14\u8bb0\u8fd8\u6ca1\u5168\u90e8\u4e0a\u4f20\u3002\u540e\u7eed\u5185\u5bb9\u4f1a\u6301\u7eed\u66f4\u65b0&amp;#x\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/119483.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-29T13:01:01+00:00\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"19 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/119483.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/119483.html\",\"name\":\"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-10-29T13:01:01+00:00\",\"dateModified\":\"2026-10-29T13:01:01+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/119483.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/119483.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/119483.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/119483.html","og_locale":"zh_CN","og_type":"article","og_title":"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u6587\u6863\u89e3\u6790\u4e22\u7684\u4e0d\u662f\u5b57&#xff0c;\u662f\u7ed3\u6784&#xff1a;\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u628a\u5c3a\u5b50 \u7cfb\u5217\u5bfc\u822a&#xff1a;00 \u7cfb\u5217\u5bfc\u822a \u00b7 01 \u56db\u7c7b\u77e5\u8bc6 \u00b7 \u672c\u7bc7&#xff08;02A&#xff09; \u00b7 02B \u65ad\u88c2\u4fee\u590d \u00b7 03 \u5207\u5206\u6807\u5b9a \u00b7 04 \u67e5\u8be2\u7406\u89e3 \u00b7 05 \u6df7\u5408\u68c0\u7d22\u4e0e\u91cd\u6392 \u00b7 06 \u5f15\u7528\u9a8c\u8bc1 \u00b7 07 \u77e5\u8bc6\u65b0\u9c9c\u5ea6 \u00b7 08 \u591a\u6e90\u4e00\u81f4\u6027 \u00b7 09 \u7ba1\u9053\u95e8\u7981 \u5c0f\u63d0\u4e00\u53e5&#xff1a; \u7531\u4e8e\u5e73\u53f0\u6bcf\u65e5\u53d1\u5e03\u7b14\u8bb0\u6570\u91cf\u9650\u5236&#xff0c;\u76ee\u524d\u7cfb\u5217\u7b14\u8bb0\u8fd8\u6ca1\u5168\u90e8\u4e0a\u4f20\u3002\u540e\u7eed\u5185\u5bb9\u4f1a\u6301\u7eed\u66f4\u65b0&amp;#x","og_url":"https:\/\/www.wsisp.com\/helps\/119483.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-10-29T13:01:01+00:00","author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"19 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/119483.html","url":"https:\/\/www.wsisp.com\/helps\/119483.html","name":"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-10-29T13:01:01+00:00","dateModified":"2026-10-29T13:01:01+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/119483.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/119483.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/119483.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"SA-K-02A \u77e5\u8bc6\u4e0e\u68c0\u7d22\u5de5\u7a0b \u6587\u6863\u89e3\u6790\u4e0a-\u7ba1\u7ebf\u4e94\u73af\u4e0e\u8d28\u91cf\u56db\u5c3a"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/119483","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=119483"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/119483\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=119483"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=119483"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=119483"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=119483"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}