MinerU · 全程本地解析 / local pipeline
烂文档 → 干净结构 Messy documents, parsed into clean Markdown & JSON
左边是现实里没法直接用的文档原页，右边是 MinerU 在本地一条命令解析出的结果。三份样例都来自真实运行，未手工修饰。
01公式还原 · LaTeX from a 1953 scan
raw scanNACA-1135 · 1953
→
parsed · LaTeXrender
$$ \nu = -\int_{p_*}^{p} \frac{\sin 2\mu}{2\gamma p}\,dp\ \ [\mathrm{isen,\ therm\ perf}] $$
02表格还原 · structured data table
source pdfNASA CR-2008
→
parsed · tableHTML/MD
| Airfoil | t/c | AR | CD2max measured | AERODAS | StC | Flat-plate |
|---|---|---|---|---|---|---|
| NACA 0012 | 0.12 | Inf. | 2.090 | 2.053 | 1.902 | 1.980 |
| NACA 0012 | 0.12 | Inf. | 2.090 | 2.053 | 1.902 | 1.980 |
| NACA 0015 | 0.15 | Inf. | 1.700 | 2.007 | 1.878 | 1.980 |
| NACA 4409 | 0.09 | Inf. | 2.100 | 2.100 | 1.985 | 1.980 |
| NACA 4412 | 0.12 | Inf. | 2.060 | 2.053 | 1.959 | 1.980 |
| NACA 4415 | 0.15 | Inf. | 2.068 | 2.007 | 1.933 | 1.980 |
| NACA 4418 | 0.18 | Inf. | 2.060 | 1.964 | 1.906 | 1.980 |
| NACA 0012 | 0.12 | Inf. | 2.050 | 2.053 | 1.902 | 1.980 |
| NACA 23012 | 0.12 | Inf. | 2.082 | 2.053 | 1.948 | 1.980 |
| NACA 23017 | 0.17 | Inf. | 2.078 | 1.978 | 1.902 | 1.980 |
| FX-84-W-127 | 0.13 | Inf. | 2.000 | 2.042 | 1.964 | 1.980 |
| FX-84-W-218 | 0.22 | Inf. | 2.040 | 1.911 | 1.939 | 1.980 |
真实单元格与数值，与左图逐格一致 · 节选前 12 行（完整 28 行见下方 .md / .json 下载）
03图表抽取 · extracted figure
raw scanNACA-1135 · chart
→
parsed · imagecropped
