Compare commits
7
Commits
| Author | SHA1 | Date | |
|---|---|---|---|
 Abimael MartellandClaude Fable 5
|
7467f842a5 | ||
|
Abimael MartellandClaude Fable 5
|
740488d875 | ||
|
Abimael MartellandClaude Fable 5
|
0b894b7d3f | ||
|
Abimael MartellandClaude Fable 5
|
ebc2f05bcd | ||
|
Abimael MartellandClaude Fable 5
|
dd8aba9e20 | ||
|
Abimael MartellandClaude Fable 5
|
2dcba76d6f | ||
|
Abimael MartellandClaude Fable 5
|
3a389f079e |
@@ -0,0 +1,36 @@
|
||||
name: Deploy landing page
|
||||
|
||||
on:
|
||||
push:
|
||||
branches: [main]
|
||||
paths: ['site/**', '.github/workflows/pages.yml']
|
||||
workflow_dispatch:
|
||||
|
||||
permissions:
|
||||
contents: read
|
||||
pages: write
|
||||
id-token: write
|
||||
|
||||
# Allow one concurrent deployment; don't cancel an in-progress production deploy.
|
||||
concurrency:
|
||||
group: pages
|
||||
cancel-in-progress: false
|
||||
|
||||
jobs:
|
||||
deploy:
|
||||
name: Build & deploy to GitHub Pages
|
||||
runs-on: ubuntu-latest
|
||||
environment:
|
||||
name: github-pages
|
||||
url: ${{ steps.deploy.outputs.page_url }}
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
|
||||
- name: Upload site artifact
|
||||
uses: actions/upload-pages-artifact@v3
|
||||
with:
|
||||
path: site
|
||||
|
||||
- name: Deploy to GitHub Pages
|
||||
id: deploy
|
||||
uses: actions/deploy-pages@v4
|
||||
@@ -0,0 +1,3 @@
|
||||
<svg width="200" height="284" viewBox="0 0 200 284" fill="none" xmlns="http://www.w3.org/2000/svg">
|
||||
<path d="M166.862 90.7716C155.812 94.0514 147.483 101.471 141.383 109.53C140.073 111.26 137.343 109.96 137.863 107.841C149.543 59.8136 134.113 19.896 86.0157 0.247269C83.5758 -0.752669 81.0359 1.43719 81.6759 3.99704C103.555 91.8416 11.5294 84.432 23.1588 184.016C23.3588 185.726 21.4389 186.896 20.039 185.896C15.6792 182.766 10.8095 176.236 7.46963 171.647C6.48968 170.297 4.36978 170.677 3.9198 172.287C1.25994 181.906 0 190.965 0 199.965C0 234.963 17.9891 265.771 45.2177 283.63C46.7777 284.65 48.7776 283.19 48.2476 281.4C46.8477 276.7 46.0577 271.74 45.9977 266.611C45.9977 263.461 46.1977 260.241 46.6877 257.241C47.8276 249.702 50.4475 242.522 54.8473 235.983C69.9365 213.334 100.185 191.455 95.3552 161.747C95.0453 159.867 97.2651 158.627 98.6651 159.917C119.974 179.386 124.194 205.575 120.694 229.063C120.394 231.103 122.954 232.193 124.244 230.593C127.504 226.513 131.483 222.933 135.813 220.244C136.893 219.574 138.333 220.084 138.743 221.284C141.153 228.293 144.733 234.873 148.113 241.452C152.152 249.362 154.302 258.391 153.962 267.951C153.792 272.6 153.022 277.1 151.732 281.38C151.182 283.19 153.162 284.7 154.752 283.66C182.001 265.801 200 234.993 200 199.975C200 187.806 197.87 175.876 193.84 164.697C185.391 141.248 163.952 123.64 169.372 93.0815C169.632 91.6216 168.282 90.3517 166.862 90.7716Z" fill="#FA5D19" style="fill:#FA5D19;fill:color(display-p3 0.9816 0.3634 0.0984);fill-opacity:1;"/>
|
||||
</svg>
|
||||
|
After Width: | Height: | Size: 1.5 KiB |
@@ -0,0 +1,12 @@
|
||||
<svg width="172" height="40" viewBox="0 0 172 40" fill="none" xmlns="http://www.w3.org/2000/svg">
|
||||
<path d="M23.3606 12.8281C21.8137 13.2873 20.6476 14.3261 19.7936 15.4544C19.6102 15.6966 19.228 15.5146 19.3008 15.2178C20.936 8.49401 18.7759 2.90556 12.0422 0.154735C11.7006 0.0147436 11.345 0.321324 11.4346 0.679702C14.4977 12.9779 1.61412 11.9406 3.24224 25.8823C3.27024 26.1217 3.00145 26.2855 2.80546 26.1455C2.19509 25.7073 1.51332 24.7932 1.04575 24.1506C0.908555 23.9616 0.611769 24.0148 0.548773 24.2402C0.176391 25.5869 0 26.8553 0 28.1152C0 33.0149 2.51847 37.328 6.33048 39.8283C6.54887 39.9711 6.82886 39.7667 6.75466 39.5161C6.55867 38.8581 6.44808 38.1638 6.43968 37.4456C6.43968 37.0046 6.46768 36.5539 6.53627 36.1339C6.69587 35.0784 7.06265 34.0732 7.67862 33.1577C9.79111 29.9869 14.0259 26.9239 13.3497 22.7647C13.3063 22.5015 13.6171 22.328 13.8131 22.5085C16.7964 25.2342 17.3871 28.9005 16.8972 32.1889C16.8552 32.4745 17.2135 32.6271 17.3941 32.4031C17.8505 31.832 18.4077 31.3308 19.0138 30.9542C19.165 30.8604 19.3666 30.9318 19.424 31.0998C19.7614 32.0811 20.2626 33.0023 20.7358 33.9234C21.3013 35.0308 21.6023 36.2949 21.5547 37.6332C21.5309 38.2842 21.4231 38.9141 21.2425 39.5133C21.1655 39.7667 21.4427 39.9781 21.6653 39.8325C25.4801 37.3322 28 33.0191 28 28.1166C28 26.4129 27.7018 24.7428 27.1376 23.1777C25.9547 19.8949 22.9533 17.4297 23.712 13.1515C23.7484 12.9471 23.5594 12.7693 23.3606 12.8281Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M41 34.0521V10.9618H55.7586V14.3264H44.7969V21.0226H53.8436V24.2882H44.7969V34.0521H41Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M59.9569 14.7882C58.7352 14.7882 57.7777 13.8976 57.7777 12.6441C57.7777 11.3906 58.7352 10.5 59.9569 10.5C61.1785 10.5 62.136 11.3906 62.136 12.6441C62.136 13.8976 61.1785 14.7882 59.9569 14.7882ZM58.1409 34.0521V17.1632H61.7068V34.0521H58.1409Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M73.5885 17.1632H74.3809V20.4948H72.796C69.6264 20.4948 68.6029 22.9687 68.6029 25.5747V34.0521H65.0371V17.1632H68.2067L68.6029 19.7031C69.4613 18.2847 70.815 17.1632 73.5885 17.1632Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M83.632 34.25C78.3163 34.25 74.9816 30.8194 74.9816 25.6406C74.9816 20.4288 78.3163 16.9653 83.3019 16.9653C88.1884 16.9653 91.457 20.066 91.5561 25.0139C91.5561 25.4427 91.5231 25.9045 91.457 26.3663H78.7125V26.5972C78.8116 29.467 80.6275 31.3472 83.4339 31.3472C85.613 31.3472 87.1979 30.2587 87.6931 28.3785H91.2589C90.6646 31.7101 87.8252 34.25 83.632 34.25ZM78.8446 23.7604H87.8582C87.561 21.2535 85.8112 19.8351 83.3349 19.8351C81.0567 19.8351 79.1087 21.3524 78.8446 23.7604Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M102.033 34.25C96.9151 34.25 93.6465 30.9184 93.6465 25.6406C93.6465 20.4288 97.0142 16.9653 102.132 16.9653C106.49 16.9653 109.197 19.3733 109.891 23.1997H106.16C105.698 21.2205 104.278 20 102.066 20C99.1933 20 97.3113 22.309 97.3113 25.6406C97.3113 28.9392 99.1933 31.2153 102.066 31.2153C104.245 31.2153 105.698 29.9618 106.127 28.0156H109.891C109.23 31.842 106.358 34.25 102.033 34.25Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M121.006 17.1632H121.799V20.4948H120.214C117.044 20.4948 116.021 22.9687 116.021 25.5747V34.0521H112.455V17.1632H115.625L116.021 19.7031C116.879 18.2847 118.233 17.1632 121.006 17.1632Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M130.614 16.9653C135.104 16.9653 137.679 19.1094 137.679 23.1007V34.0521H134.576L134.279 31.6441C133.123 33.1615 131.505 34.25 128.831 34.25C125.133 34.25 122.657 32.4358 122.657 29.3021C122.657 25.8385 125.166 23.8924 129.92 23.8924H134.147V22.8698C134.147 20.9896 132.793 19.8351 130.449 19.8351C128.336 19.8351 126.916 20.8247 126.652 22.309H123.152C123.515 19.0104 126.355 16.9653 130.614 16.9653ZM129.425 31.4792C132.397 31.4792 134.114 29.7309 134.147 27.125V26.5312H129.722C127.51 26.5312 126.289 27.3559 126.289 29.0712C126.289 30.4896 127.477 31.4792 129.425 31.4792Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M144.653 34.0521L139.139 17.1632H142.903L146.766 30.0937L150.629 17.1632H153.897L157.595 30.0937L161.59 17.1632H165.222L159.609 34.0521H155.779L152.214 22.5729L148.516 34.0521H144.653Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
<path d="M166.934 34.0521V10.9618H170.5V34.0521H166.934Z" fill="#262626" style="fill:#262626;fill:color(display-p3 0.1500 0.1500 0.1500);fill-opacity:1;"/>
|
||||
</svg>
|
||||
|
After Width: | Height: | Size: 4.8 KiB |
+428
@@ -0,0 +1,428 @@
|
||||
<!DOCTYPE html>
|
||||
<html lang="en">
|
||||
<head>
|
||||
<meta charset="utf-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1">
|
||||
<title>pdf-inspector — PDF classification & text extraction, no OCR</title>
|
||||
<meta name="description" content="Fast Rust library that classifies PDFs (text-based vs scanned) and extracts clean Markdown — no OCR, no ML models. Bindings for Rust, Python, and Node.js.">
|
||||
<meta property="og:title" content="pdf-inspector">
|
||||
<meta property="og:description" content="Classify PDFs and extract clean Markdown in milliseconds. No OCR. No ML. Pure Rust.">
|
||||
<meta property="og:type" content="website">
|
||||
<link rel="icon" href="data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 100 100'%3E%3Ctext y='.9em' font-size='90'%3E%F0%9F%93%84%3C/text%3E%3C/svg%3E">
|
||||
<link rel="preconnect" href="https://fonts.googleapis.com">
|
||||
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
|
||||
<link href="https://fonts.googleapis.com/css2?family=Bricolage+Grotesque:opsz,wght@12..96,400;12..96,600;12..96,800&family=Hanken+Grotesk:wght@400;500;600&family=JetBrains+Mono:wght@400;500;700&display=swap" rel="stylesheet">
|
||||
<style>
|
||||
:root {
|
||||
--paper: #f4efe4;
|
||||
--paper-2: #eee7d8;
|
||||
--ink: #1b1712;
|
||||
--ink-soft: #4a433a;
|
||||
--muted: #8b8375;
|
||||
--line: #d9cfbb;
|
||||
--accent: #dd3f22;
|
||||
--accent-deep: #b32d15;
|
||||
--card: #faf6ec;
|
||||
--display: "Bricolage Grotesque", serif;
|
||||
--body: "Hanken Grotesk", sans-serif;
|
||||
--mono: "JetBrains Mono", monospace;
|
||||
}
|
||||
* { box-sizing: border-box; margin: 0; padding: 0; }
|
||||
html { scroll-behavior: smooth; }
|
||||
body {
|
||||
background: var(--paper);
|
||||
color: var(--ink);
|
||||
font-family: var(--body);
|
||||
font-size: 17px;
|
||||
line-height: 1.6;
|
||||
-webkit-font-smoothing: antialiased;
|
||||
overflow-x: hidden;
|
||||
background-image:
|
||||
radial-gradient(circle at 1px 1px, rgba(27,23,18,0.05) 1px, transparent 0);
|
||||
background-size: 22px 22px;
|
||||
}
|
||||
::selection { background: var(--accent); color: var(--paper); }
|
||||
a { color: inherit; text-decoration: none; }
|
||||
|
||||
.wrap { max-width: 1120px; margin: 0 auto; padding: 0 28px; }
|
||||
|
||||
/* ── nav ── */
|
||||
nav {
|
||||
position: sticky; top: 0; z-index: 50;
|
||||
background: rgba(244,239,228,0.82);
|
||||
backdrop-filter: blur(10px);
|
||||
border-bottom: 1px solid var(--line);
|
||||
}
|
||||
.nav-in { display: flex; align-items: center; gap: 22px; height: 60px; }
|
||||
.brand { font-family: var(--mono); font-weight: 700; font-size: 15px; letter-spacing: -0.02em; display: flex; align-items: center; gap: 9px; }
|
||||
.brand .dot { width: 9px; height: 9px; background: var(--accent); border-radius: 50%; box-shadow: 0 0 0 3px rgba(221,63,34,0.18); }
|
||||
.nav-links { margin-left: auto; display: flex; gap: 24px; align-items: center; font-size: 14.5px; font-weight: 500; }
|
||||
.nav-links a { color: var(--ink-soft); transition: color .15s; }
|
||||
.nav-links a:hover { color: var(--accent); }
|
||||
.nav-gh { border: 1px solid var(--ink); border-radius: 999px; padding: 6px 15px; color: var(--ink) !important; transition: all .15s; }
|
||||
.nav-gh:hover { background: var(--ink); color: var(--paper) !important; }
|
||||
@media (max-width: 680px) { .nav-links .hide-sm { display: none; } }
|
||||
|
||||
/* ── hero ── */
|
||||
header { padding: 74px 0 40px; position: relative; }
|
||||
.eyebrow { font-family: var(--mono); font-size: 12.5px; letter-spacing: 0.16em; text-transform: uppercase; color: var(--accent-deep); margin-bottom: 22px; }
|
||||
h1 {
|
||||
font-family: var(--display);
|
||||
font-weight: 800;
|
||||
font-size: clamp(2.9rem, 8vw, 6.1rem);
|
||||
line-height: 0.96;
|
||||
letter-spacing: -0.035em;
|
||||
max-width: 15ch;
|
||||
}
|
||||
h1 .em { color: var(--accent); font-style: normal; position: relative; }
|
||||
h1 .strike { position: relative; white-space: nowrap; }
|
||||
h1 .strike::after { content: ""; position: absolute; left: -2%; right: -2%; top: 54%; height: 0.09em; background: var(--accent); transform: rotate(-3deg); }
|
||||
.lede { margin-top: 28px; font-size: clamp(1.05rem, 2.2vw, 1.32rem); color: var(--ink-soft); max-width: 46ch; line-height: 1.5; }
|
||||
.lede b { color: var(--ink); font-weight: 600; }
|
||||
|
||||
.hero-grid { display: grid; grid-template-columns: 1.15fr 0.85fr; gap: 48px; align-items: end; }
|
||||
@media (max-width: 880px) { .hero-grid { grid-template-columns: 1fr; gap: 40px; } }
|
||||
|
||||
/* readout card */
|
||||
.readout {
|
||||
background: var(--ink); color: var(--paper);
|
||||
border-radius: 14px; padding: 22px 22px 20px;
|
||||
font-family: var(--mono); font-size: 13px;
|
||||
box-shadow: 14px 14px 0 rgba(27,23,18,0.09);
|
||||
position: relative;
|
||||
}
|
||||
.readout .rlabel { color: #b8ad98; font-size: 11px; letter-spacing: 0.14em; text-transform: uppercase; margin-bottom: 16px; display: flex; justify-content: space-between; }
|
||||
.readout .rrow { display: flex; justify-content: space-between; align-items: center; padding: 9px 0; border-top: 1px solid rgba(255,255,255,0.09); }
|
||||
.readout .rrow:first-of-type { border-top: none; }
|
||||
.readout .k { color: #cfc6b3; }
|
||||
.readout .v { font-weight: 700; }
|
||||
.readout .v.hot { color: var(--accent); }
|
||||
.bar { height: 6px; background: rgba(255,255,255,0.1); border-radius: 3px; overflow: hidden; margin-top: 3px; width: 96px; }
|
||||
.bar > i { display: block; height: 100%; background: var(--accent); border-radius: 3px; }
|
||||
|
||||
/* ── install row ── */
|
||||
.installs { display: grid; grid-template-columns: repeat(3,1fr); gap: 14px; margin-top: 54px; }
|
||||
@media (max-width: 720px) { .installs { grid-template-columns: 1fr; } }
|
||||
.inst {
|
||||
background: var(--card); border: 1px solid var(--line); border-radius: 11px;
|
||||
padding: 15px 17px; transition: transform .16s, border-color .16s, box-shadow .16s;
|
||||
cursor: pointer;
|
||||
}
|
||||
.inst:hover { transform: translateY(-3px); border-color: var(--accent); box-shadow: 0 8px 22px rgba(27,23,18,0.07); }
|
||||
.inst .reg { font-family: var(--mono); font-size: 11px; letter-spacing: 0.12em; text-transform: uppercase; color: var(--muted); margin-bottom: 8px; display: flex; justify-content: space-between; }
|
||||
.inst code { font-family: var(--mono); font-size: 14px; color: var(--ink); font-weight: 500; }
|
||||
.inst .arrow { color: var(--accent); opacity: 0; transition: opacity .16s; }
|
||||
.inst:hover .arrow { opacity: 1; }
|
||||
|
||||
/* ── section scaffold ── */
|
||||
section { padding: 66px 0; border-top: 1px solid var(--line); }
|
||||
.sec-head { display: flex; align-items: baseline; gap: 16px; margin-bottom: 40px; flex-wrap: wrap; }
|
||||
.sec-num { font-family: var(--mono); font-size: 13px; color: var(--accent); font-weight: 700; }
|
||||
.sec-title { font-family: var(--display); font-weight: 600; font-size: clamp(1.7rem, 4vw, 2.6rem); letter-spacing: -0.025em; }
|
||||
.sec-sub { color: var(--ink-soft); max-width: 52ch; font-size: 1.02rem; }
|
||||
|
||||
/* ── features ── */
|
||||
.feat-grid { display: grid; grid-template-columns: repeat(4,1fr); gap: 1px; background: var(--line); border: 1px solid var(--line); border-radius: 14px; overflow: hidden; }
|
||||
@media (max-width: 900px) { .feat-grid { grid-template-columns: repeat(2,1fr); } }
|
||||
@media (max-width: 560px) { .feat-grid { grid-template-columns: 1fr; } }
|
||||
.feat { background: var(--card); padding: 24px 22px; transition: background .16s; }
|
||||
.feat:hover { background: #fff; }
|
||||
.feat .fn { font-family: var(--mono); font-size: 12px; color: var(--accent); font-weight: 700; }
|
||||
.feat h3 { font-family: var(--display); font-weight: 600; font-size: 1.16rem; margin: 12px 0 8px; letter-spacing: -0.01em; }
|
||||
.feat p { font-size: 14.5px; color: var(--ink-soft); line-height: 1.5; }
|
||||
|
||||
/* ── benchmark ── */
|
||||
.bench {
|
||||
border: 1px solid var(--line); border-radius: 14px; overflow: hidden;
|
||||
background: var(--card);
|
||||
}
|
||||
table { width: 100%; border-collapse: collapse; font-size: 15px; }
|
||||
thead th { font-family: var(--mono); font-size: 11px; letter-spacing: 0.08em; text-transform: uppercase; color: var(--muted); text-align: right; padding: 15px 18px; background: var(--paper-2); border-bottom: 1px solid var(--line); font-weight: 500; }
|
||||
thead th:first-child { text-align: left; }
|
||||
tbody td { padding: 14px 18px; text-align: right; font-family: var(--mono); border-bottom: 1px solid var(--line); }
|
||||
tbody td:first-child { text-align: left; font-family: var(--body); font-weight: 500; }
|
||||
tbody tr:last-child td { border-bottom: none; }
|
||||
tbody tr.us { background: rgba(221,63,34,0.06); }
|
||||
tbody tr.us td:first-child { color: var(--accent-deep); font-weight: 700; }
|
||||
tbody tr.us td:first-child::before { content: "▸ "; color: var(--accent); }
|
||||
.bench-foot { padding: 15px 18px; font-size: 13.5px; color: var(--ink-soft); background: var(--paper-2); border-top: 1px solid var(--line); }
|
||||
.bench-wrap { overflow-x: auto; }
|
||||
.callouts { display: grid; grid-template-columns: 1fr 1fr; gap: 16px; margin-top: 22px; }
|
||||
@media (max-width: 640px) { .callouts { grid-template-columns: 1fr; } }
|
||||
.callout { border-left: 3px solid var(--accent); padding: 4px 0 4px 16px; }
|
||||
.callout .ct { font-family: var(--mono); font-size: 11px; letter-spacing: 0.1em; text-transform: uppercase; color: var(--accent-deep); margin-bottom: 5px; }
|
||||
.callout p { font-size: 14.5px; color: var(--ink-soft); }
|
||||
|
||||
/* ── quickstart tabs ── */
|
||||
.tabs input { position: absolute; opacity: 0; pointer-events: none; }
|
||||
.tablist { display: flex; gap: 6px; margin-bottom: 0; }
|
||||
.tablist label {
|
||||
font-family: var(--mono); font-size: 13px; font-weight: 500;
|
||||
padding: 10px 18px; cursor: pointer; color: var(--muted);
|
||||
border: 1px solid var(--line); border-bottom: none;
|
||||
border-radius: 9px 9px 0 0; background: var(--paper-2); transition: all .15s;
|
||||
}
|
||||
.tablist label:hover { color: var(--ink); }
|
||||
.panel { display: none; }
|
||||
.code {
|
||||
background: var(--ink); border-radius: 0 12px 12px 12px;
|
||||
padding: 22px 24px; overflow-x: auto;
|
||||
font-family: var(--mono); font-size: 13.5px; line-height: 1.7;
|
||||
color: #e9e2d3;
|
||||
box-shadow: 12px 12px 0 rgba(27,23,18,0.07);
|
||||
}
|
||||
.code .cm { color: #8a8069; }
|
||||
.code .kw { color: #ff9f7a; }
|
||||
.code .st { color: #cbb78a; }
|
||||
.code .fn { color: #f4efe4; font-weight: 700; }
|
||||
#t-rust:checked ~ .tablist label[for=t-rust],
|
||||
#t-py:checked ~ .tablist label[for=t-py],
|
||||
#t-node:checked ~ .tablist label[for=t-node],
|
||||
#t-cli:checked ~ .tablist label[for=t-cli] {
|
||||
background: var(--ink); color: var(--paper); border-color: var(--ink);
|
||||
}
|
||||
#t-rust:checked ~ .panels #p-rust,
|
||||
#t-py:checked ~ .panels #p-py,
|
||||
#t-node:checked ~ .panels #p-node,
|
||||
#t-cli:checked ~ .panels #p-cli { display: block; }
|
||||
.code a.ref { color: #ff9f7a; border-bottom: 1px dotted #ff9f7a; }
|
||||
|
||||
/* ── closing split (OSS vs hosted) ── */
|
||||
.split { display: grid; grid-template-columns: 1fr 1fr; gap: 18px; }
|
||||
@media (max-width: 780px) { .split { grid-template-columns: 1fr; } }
|
||||
.path { border: 1px solid var(--line); border-radius: 16px; padding: 32px 30px; background: var(--card); display: flex; flex-direction: column; }
|
||||
.path .ptag { font-family: var(--mono); font-size: 11px; letter-spacing: 0.12em; text-transform: uppercase; color: var(--muted); margin-bottom: 15px; }
|
||||
.path h3 { font-family: var(--display); font-weight: 600; font-size: 1.5rem; letter-spacing: -0.02em; line-height: 1.05; margin-bottom: 12px; }
|
||||
.path p { color: var(--ink-soft); font-size: 15px; line-height: 1.5; flex: 1; margin-bottom: 24px; }
|
||||
.pbtns { display: flex; gap: 12px; flex-wrap: wrap; }
|
||||
.path-pro { background: var(--ink); border-color: var(--ink); box-shadow: 14px 14px 0 rgba(27,23,18,0.09); }
|
||||
.path-pro .ptag { color: #b8ad98; }
|
||||
.path-pro .ptag b { color: var(--accent); font-weight: 700; }
|
||||
.path-pro h3 { color: var(--paper); }
|
||||
.path-pro p { color: #cfc6b3; }
|
||||
.btn { font-family: var(--mono); font-size: 14px; font-weight: 500; padding: 13px 24px; border-radius: 999px; transition: all .15s; border: 1px solid var(--ink); }
|
||||
.btn-p { background: var(--accent); border-color: var(--accent); color: var(--paper); }
|
||||
.btn-p:hover { background: var(--accent-deep); border-color: var(--accent-deep); }
|
||||
.btn-s:hover { background: var(--ink); color: var(--paper); }
|
||||
.btn-pro { background: var(--accent); border-color: var(--accent); color: var(--paper); }
|
||||
.btn-pro:hover { background: var(--accent-deep); border-color: var(--accent-deep); }
|
||||
.btn-ghost { color: var(--paper); border-color: rgba(255,255,255,0.3); }
|
||||
.btn-ghost:hover { border-color: var(--paper); background: rgba(255,255,255,0.08); }
|
||||
.fc-mark { height: 34px; width: auto; display: block; margin-bottom: 20px; }
|
||||
.fc-wordmark { height: 15px; width: auto; vertical-align: -2px; transition: opacity .15s; }
|
||||
.fc-wordmark:hover { opacity: 0.65; }
|
||||
footer { border-top: 1px solid var(--line); padding: 34px 0; font-size: 14px; color: var(--muted); }
|
||||
.foot-in { display: flex; justify-content: space-between; align-items: center; gap: 18px; flex-wrap: wrap; }
|
||||
.foot-in a { color: var(--ink-soft); }
|
||||
.foot-in a:hover { color: var(--accent); }
|
||||
.foot-links { display: flex; gap: 20px; font-family: var(--mono); font-size: 13px; }
|
||||
|
||||
/* ── load animation ── */
|
||||
.reveal { opacity: 0; transform: translateY(14px); animation: rise .7s cubic-bezier(.2,.7,.3,1) forwards; }
|
||||
@keyframes rise { to { opacity: 1; transform: none; } }
|
||||
.d1 { animation-delay: .05s; } .d2 { animation-delay: .15s; } .d3 { animation-delay: .25s; }
|
||||
.d4 { animation-delay: .35s; } .d5 { animation-delay: .45s; } .d6 { animation-delay: .55s; }
|
||||
@media (prefers-reduced-motion: reduce) { .reveal { animation: none; opacity: 1; transform: none; } }
|
||||
</style>
|
||||
</head>
|
||||
<body>
|
||||
|
||||
<nav>
|
||||
<div class="wrap nav-in">
|
||||
<a href="#top" class="brand"><span class="dot"></span>pdf-inspector</a>
|
||||
<div class="nav-links">
|
||||
<a href="#features" class="hide-sm">Features</a>
|
||||
<a href="#benchmark" class="hide-sm">Benchmark</a>
|
||||
<a href="#start">Quick start</a>
|
||||
<a class="nav-gh" href="https://github.com/firecrawl/pdf-inspector">GitHub ↗</a>
|
||||
</div>
|
||||
</div>
|
||||
</nav>
|
||||
|
||||
<header id="top">
|
||||
<div class="wrap hero-grid">
|
||||
<div>
|
||||
<div class="eyebrow reveal d1">Rust · Python · Node · CLI</div>
|
||||
<h1 class="reveal d2">Classify PDFs. Extract Markdown. <span class="strike">No OCR.</span></h1>
|
||||
<p class="lede reveal d3">A fast Rust library that tells text-based PDFs from scanned ones, then extracts position-aware text and clean Markdown — <b>locally, in milliseconds</b>. Skip the OCR bill for the ~54% of PDFs that never needed it.</p>
|
||||
</div>
|
||||
<div class="readout reveal d4" aria-hidden="true">
|
||||
<div class="rlabel"><span>classify_pdf()</span><span>~12ms</span></div>
|
||||
<div class="rrow"><span class="k">type</span><span class="v hot">TextBased</span></div>
|
||||
<div class="rrow"><span class="k">confidence</span><span class="v">0.98</span></div>
|
||||
<div class="rrow"><span class="k">needs_ocr</span><span class="v">false</span></div>
|
||||
<div class="rrow"><span class="k">route</span><span class="v">local → md</span></div>
|
||||
<div class="rrow" style="border-top:1px solid rgba(255,255,255,.09);padding-top:13px">
|
||||
<span class="k">signal</span>
|
||||
<span class="bar"><i style="width:98%"></i></span>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div class="wrap">
|
||||
<div class="installs">
|
||||
<a class="inst reveal d4" href="https://crates.io/crates/pdf-inspector">
|
||||
<div class="reg"><span>crates.io</span><span class="arrow">↗</span></div>
|
||||
<code>cargo add pdf-inspector</code>
|
||||
</a>
|
||||
<a class="inst reveal d5" href="https://pypi.org/project/pdf-inspector/">
|
||||
<div class="reg"><span>PyPI</span><span class="arrow">↗</span></div>
|
||||
<code>pip install pdf-inspector</code>
|
||||
</a>
|
||||
<a class="inst reveal d6" href="https://www.npmjs.com/package/@firecrawl/pdf-inspector">
|
||||
<div class="reg"><span>npm</span><span class="arrow">↗</span></div>
|
||||
<code>npm i @firecrawl/pdf-inspector</code>
|
||||
</a>
|
||||
</div>
|
||||
</div>
|
||||
</header>
|
||||
|
||||
<section id="features">
|
||||
<div class="wrap">
|
||||
<div class="sec-head">
|
||||
<span class="sec-num">01</span>
|
||||
<h2 class="sec-title">Built for routing, not just reading</h2>
|
||||
</div>
|
||||
<div class="feat-grid">
|
||||
<div class="feat"><div class="fn">01</div><h3>Smart classification</h3><p>TextBased, Scanned, ImageBased, or Mixed in ~10–50ms by sampling content streams. Returns a confidence score and per-page OCR routing.</p></div>
|
||||
<div class="feat"><div class="fn">02</div><h3>Position-aware text</h3><p>Extraction with font info, X/Y coordinates, and automatic multi-column reading order.</p></div>
|
||||
<div class="feat"><div class="fn">03</div><h3>Markdown conversion</h3><p>Headings, bullet/numbered lists, code blocks, tables, bold/italic, URL linking, and page breaks.</p></div>
|
||||
<div class="feat"><div class="fn">04</div><h3>Table detection</h3><p>Rectangle-based detection from drawing ops plus heuristic alignment detection. Financial tables, footnotes, and cross-page continuations.</p></div>
|
||||
<div class="feat"><div class="fn">05</div><h3>CID font support</h3><p>ToUnicode CMap decoding for Type0/Identity-H fonts, with UTF-16BE, UTF-8, and Latin-1 encodings.</p></div>
|
||||
<div class="feat"><div class="fn">06</div><h3>Multi-column layout</h3><p>Newspaper-style column detection, sequential reading order, and right-to-left text support.</p></div>
|
||||
<div class="feat"><div class="fn">07</div><h3>Encoding checks</h3><p>Flags broken font encodings automatically so callers can fall back to OCR only when it's actually needed.</p></div>
|
||||
<div class="feat"><div class="fn">08</div><h3>Lightweight</h3><p>Pure Rust. No ML models, no external services. A single parse shared between detection and extraction.</p></div>
|
||||
</div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
<section id="benchmark">
|
||||
<div class="wrap">
|
||||
<div class="sec-head">
|
||||
<span class="sec-num">02</span>
|
||||
<h2 class="sec-title">Fastest of the direct-text engines</h2>
|
||||
<p class="sec-sub">Evaluated on the <a href="https://github.com/opendataloader-project/opendataloader-bench" style="color:var(--accent-deep);border-bottom:1px solid var(--line)">opendataloader-bench</a> corpus (200 PDFs). Direct text-extraction engines only — no OCR, no ML. Higher is better.</p>
|
||||
</div>
|
||||
<div class="bench">
|
||||
<div class="bench-wrap">
|
||||
<table>
|
||||
<thead>
|
||||
<tr><th>Engine</th><th>Overall</th><th>Reading order</th><th>Tables</th><th>Headings</th><th>200 docs</th></tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
<tr class="us"><td>pdf-inspector</td><td>0.83</td><td>0.88</td><td>0.66</td><td>0.74</td><td>4s</td></tr>
|
||||
<tr><td>opendataloader</td><td>0.84</td><td>0.91</td><td>0.49</td><td>0.74</td><td>11s</td></tr>
|
||||
<tr><td>pymupdf4llm</td><td>0.73</td><td>0.89</td><td>0.40</td><td>0.41</td><td>18s</td></tr>
|
||||
<tr><td>markitdown</td><td>0.58</td><td>0.88</td><td>0.00</td><td>0.00</td><td>8s</td></tr>
|
||||
</tbody>
|
||||
</table>
|
||||
</div>
|
||||
<div class="bench-foot">OCR/ML engines (docling, marker, mineru) score 0.83–0.88 overall — but take 2–180 minutes on the same corpus.</div>
|
||||
</div>
|
||||
<div class="callouts">
|
||||
<div class="callout"><div class="ct">Where we win</div><p>Fastest engine measured, the best table detection of any engine here, and heading quality now on par with opendataloader — at ~2.5× its speed.</p></div>
|
||||
<div class="callout"><div class="ct">Where we're working</div><p>Reading order still trails opendataloader slightly, and tables that need the visual structure only an OCR engine can see.</p></div>
|
||||
</div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
<section id="start">
|
||||
<div class="wrap">
|
||||
<div class="sec-head">
|
||||
<span class="sec-num">03</span>
|
||||
<h2 class="sec-title">Three lines to Markdown</h2>
|
||||
</div>
|
||||
<div class="tabs">
|
||||
<input type="radio" name="tab" id="t-rust" checked>
|
||||
<input type="radio" name="tab" id="t-py">
|
||||
<input type="radio" name="tab" id="t-node">
|
||||
<input type="radio" name="tab" id="t-cli">
|
||||
<div class="tablist">
|
||||
<label for="t-rust">Rust</label>
|
||||
<label for="t-py">Python</label>
|
||||
<label for="t-node">Node.js</label>
|
||||
<label for="t-cli">CLI</label>
|
||||
</div>
|
||||
<div class="panels">
|
||||
<div class="panel" id="p-rust"><pre class="code"><span class="kw">use</span> pdf_inspector::process_pdf;
|
||||
|
||||
<span class="kw">let</span> result = <span class="fn">process_pdf</span>(<span class="st">"document.pdf"</span>)?;
|
||||
<span class="fn">println!</span>(<span class="st">"Type: {:?}"</span>, result.pdf_type);
|
||||
<span class="kw">if let</span> <span class="kw">Some</span>(markdown) = &result.markdown {
|
||||
<span class="fn">println!</span>(<span class="st">"{}"</span>, markdown);
|
||||
}
|
||||
<span class="cm">// full reference → </span><a class="ref" href="https://github.com/firecrawl/pdf-inspector/blob/main/docs/rust-api.md">docs/rust-api.md</a></pre></div>
|
||||
<div class="panel" id="p-py"><pre class="code"><span class="kw">import</span> pdf_inspector
|
||||
|
||||
result = pdf_inspector.<span class="fn">process_pdf</span>(<span class="st">"document.pdf"</span>)
|
||||
<span class="fn">print</span>(result.pdf_type) <span class="cm"># "text_based" | "scanned" | "image_based" | "mixed"</span>
|
||||
<span class="fn">print</span>(result.markdown) <span class="cm"># Markdown string or None</span>
|
||||
<span class="cm"># full reference → </span><a class="ref" href="https://github.com/firecrawl/pdf-inspector/blob/main/docs/python.md">docs/python.md</a></pre></div>
|
||||
<div class="panel" id="p-node"><pre class="code"><span class="kw">import</span> { readFileSync } <span class="kw">from</span> <span class="st">'fs'</span>;
|
||||
<span class="kw">import</span> { processPdf } <span class="kw">from</span> <span class="st">'@firecrawl/pdf-inspector'</span>;
|
||||
|
||||
<span class="kw">const</span> result = <span class="fn">processPdf</span>(<span class="fn">readFileSync</span>(<span class="st">'document.pdf'</span>));
|
||||
console.<span class="fn">log</span>(result.pdfType); <span class="cm">// "TextBased" | "Scanned" | ...</span>
|
||||
console.<span class="fn">log</span>(result.markdown); <span class="cm">// Markdown string or null</span>
|
||||
<span class="cm">// full reference → </span><a class="ref" href="https://github.com/firecrawl/pdf-inspector/blob/main/napi/README.md">napi/README.md</a></pre></div>
|
||||
<div class="panel" id="p-cli"><pre class="code"><span class="cm"># install the CLI tools</span>
|
||||
cargo <span class="fn">install</span> pdf-inspector
|
||||
|
||||
<span class="cm"># convert a PDF to Markdown</span>
|
||||
<span class="fn">pdf2md</span> document.pdf
|
||||
|
||||
<span class="cm"># classify only — is it scanned?</span>
|
||||
<span class="fn">detect-pdf</span> document.pdf --analyze --json
|
||||
|
||||
<span class="cm"># structured output for pipelines</span>
|
||||
<span class="fn">pdf2md</span> document.pdf --json</pre></div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
<section class="close">
|
||||
<div class="wrap">
|
||||
<div class="sec-head">
|
||||
<span class="sec-num">04</span>
|
||||
<h2 class="sec-title">Two ways to parse</h2>
|
||||
<p class="sec-sub">Run the classifier locally for text-based PDFs; hand the scanned, OCR, and at-scale work to Firecrawl.</p>
|
||||
</div>
|
||||
<div class="split">
|
||||
<div class="path">
|
||||
<div class="ptag">Open source · runs local</div>
|
||||
<h3>Use pdf-inspector yourself</h3>
|
||||
<p>Pure-Rust library and CLI. Classify and extract text-based PDFs on your own machine in milliseconds — no external calls, no OCR bill, MIT licensed.</p>
|
||||
<div class="pbtns">
|
||||
<a class="btn btn-p" href="https://github.com/firecrawl/pdf-inspector">Get started</a>
|
||||
<a class="btn btn-s" href="https://crates.io/crates/pdf-inspector">crates.io</a>
|
||||
</div>
|
||||
</div>
|
||||
<div class="path path-pro">
|
||||
<img class="fc-mark" src="assets/firecrawl-mark.svg" alt="Firecrawl" width="24" height="34">
|
||||
<div class="ptag"><b>Firecrawl Parse</b> · hosted API</div>
|
||||
<h3>Or let Firecrawl handle the hard ones</h3>
|
||||
<p>Scanned documents, OCR, DOCX / XLSX / HTML, and parsing at scale — clean, LLM-ready Markdown from one API call. The downstream route for everything local parsing can't reach.</p>
|
||||
<div class="pbtns">
|
||||
<a class="btn btn-pro" href="https://docs.firecrawl.dev/api-reference/endpoint/parse">Firecrawl Parse ↗</a>
|
||||
<a class="btn btn-ghost" href="https://firecrawl.dev">firecrawl.dev</a>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
<footer>
|
||||
<div class="wrap foot-in">
|
||||
<div style="display:flex;align-items:center;gap:7px">Built by <a href="https://firecrawl.dev"><img class="fc-wordmark" src="assets/firecrawl-wordmark.svg" alt="Firecrawl"></a> · MIT licensed</div>
|
||||
<div class="foot-links">
|
||||
<a href="https://github.com/firecrawl/pdf-inspector">GitHub</a>
|
||||
<a href="https://crates.io/crates/pdf-inspector">crates.io</a>
|
||||
<a href="https://pypi.org/project/pdf-inspector/">PyPI</a>
|
||||
<a href="https://www.npmjs.com/package/@firecrawl/pdf-inspector">npm</a>
|
||||
</div>
|
||||
</div>
|
||||
</footer>
|
||||
|
||||
</body>
|
||||
</html>
|
||||
+10
-492
@@ -39,6 +39,7 @@ pub mod markdown;
|
||||
pub mod process_mode;
|
||||
pub mod structure_tree;
|
||||
pub mod tables;
|
||||
mod text_quality;
|
||||
pub mod text_utils;
|
||||
pub mod tounicode;
|
||||
pub mod types;
|
||||
@@ -60,6 +61,10 @@ pub use types::{LayoutComplexity, PdfLine, PdfRect, TextItem};
|
||||
use lopdf::Document;
|
||||
use std::collections::{BTreeMap, HashMap, HashSet};
|
||||
use std::path::Path;
|
||||
use text_quality::{
|
||||
analyze_text_quality, detect_encoding_issues, is_cid_garbage, is_garbage_text,
|
||||
region_items_have_decoding_issue,
|
||||
};
|
||||
use tounicode::FontCMaps;
|
||||
|
||||
/// OCR reason emitted when the extracted text layer appears garbled due to
|
||||
@@ -3703,283 +3708,15 @@ fn process_document(
|
||||
// Internal helpers
|
||||
// =========================================================================
|
||||
|
||||
/// Detect broken font encodings in extracted markdown text.
|
||||
///
|
||||
/// Two heuristics:
|
||||
/// 1. **U+FFFD**: Any replacement character indicates decode failures.
|
||||
/// 2. **Dollar-as-space**: Pattern like `Word$Word$Word` where `$` is used as a
|
||||
/// word separator due to broken ToUnicode CMaps. Triggers when either:
|
||||
/// - More than 50% of `$` are between letters (clear substitution pattern), OR
|
||||
/// - More than 20 letter-dollar-letter occurrences (even if some `$` are also
|
||||
/// used as trailing/leading separators, 20+ is far beyond normal financial text).
|
||||
fn detect_encoding_issues(markdown: &str) -> bool {
|
||||
// Heuristic 1: U+FFFD replacement characters
|
||||
if markdown.contains('\u{FFFD}') {
|
||||
return true;
|
||||
}
|
||||
|
||||
// Heuristic 2: dollar-as-space pattern
|
||||
if has_dollar_as_space_pattern(markdown) {
|
||||
return true;
|
||||
}
|
||||
|
||||
// Heuristic 3: substitution-cipher letter statistics (broken ToUnicode)
|
||||
let mut stats = CipherGarbleStats::default();
|
||||
stats.add_text(markdown);
|
||||
stats.looks_garbled()
|
||||
}
|
||||
|
||||
fn has_dollar_as_space_pattern(markdown: &str) -> bool {
|
||||
let total_dollars = markdown.matches('$').count();
|
||||
if total_dollars > 10 {
|
||||
let bytes = markdown.as_bytes();
|
||||
let mut letter_dollar_letter = 0usize;
|
||||
for i in 1..bytes.len().saturating_sub(1) {
|
||||
if bytes[i] == b'$'
|
||||
&& bytes[i - 1].is_ascii_alphabetic()
|
||||
&& bytes[i + 1].is_ascii_alphabetic()
|
||||
{
|
||||
letter_dollar_letter += 1;
|
||||
}
|
||||
}
|
||||
if letter_dollar_letter > 20 || letter_dollar_letter * 2 > total_dollars {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
||||
false
|
||||
}
|
||||
|
||||
/// English letter frequencies (percent, a–z). Used as a natural-language
|
||||
/// reference: every Latin-script language in the eval corpus (Swedish,
|
||||
/// Finnish, Turkish, German, romaji) scores ≥ 0.80 cosine similarity against
|
||||
/// it, while substitution-cipher text scores ~0.53.
|
||||
const ENGLISH_LETTER_FREQ: [f64; 26] = [
|
||||
8.2, 1.5, 2.8, 4.3, 12.7, 2.2, 2.0, 6.1, 7.0, 0.15, 0.8, 4.0, 2.4, 6.7, 7.5, 1.9, 0.1, 6.0,
|
||||
6.3, 9.1, 2.8, 1.0, 2.4, 0.15, 2.0, 0.07,
|
||||
];
|
||||
|
||||
/// Letter statistics for detecting substitution-cipher garbling: broken
|
||||
/// ToUnicode CMaps that shift every character by a per-range constant (e.g.
|
||||
/// `Certificate` extracted as `8VceZWZTReV`). Such text is 100% printable
|
||||
/// ASCII with word-like token lengths, so it defeats `is_garbage_text` and
|
||||
/// produces no replacement characters — it needs its own discriminator.
|
||||
#[derive(Debug, Default)]
|
||||
struct CipherGarbleStats {
|
||||
/// Case-folded ASCII letter histogram.
|
||||
letter_counts: [u32; 26],
|
||||
ascii_letters: usize,
|
||||
ascii_vowels: usize,
|
||||
/// Accented Latin letters (Latin-1 Supplement through Latin Extended-B,
|
||||
/// plus Latin Extended Additional). Count toward Latin dominance only.
|
||||
latin_ext_letters: usize,
|
||||
non_latin_letters: usize,
|
||||
/// Adjacent ASCII-letter pairs, and how many of them switch from
|
||||
/// lowercase straight to uppercase mid-word.
|
||||
letter_bigrams: usize,
|
||||
case_shift_bigrams: usize,
|
||||
}
|
||||
|
||||
impl CipherGarbleStats {
|
||||
fn add_text(&mut self, text: &str) {
|
||||
let mut prev: Option<char> = None;
|
||||
for ch in text.chars() {
|
||||
if ch.is_ascii_alphabetic() {
|
||||
let idx = (ch.to_ascii_lowercase() as u8 - b'a') as usize;
|
||||
self.letter_counts[idx] += 1;
|
||||
self.ascii_letters += 1;
|
||||
if matches!(ch.to_ascii_lowercase(), 'a' | 'e' | 'i' | 'o' | 'u') {
|
||||
self.ascii_vowels += 1;
|
||||
}
|
||||
if let Some(p) = prev {
|
||||
self.letter_bigrams += 1;
|
||||
if p.is_ascii_lowercase() && ch.is_ascii_uppercase() {
|
||||
self.case_shift_bigrams += 1;
|
||||
}
|
||||
}
|
||||
prev = Some(ch);
|
||||
} else {
|
||||
if ch.is_alphabetic() {
|
||||
if matches!(ch as u32, 0xC0..=0x24F | 0x1E00..=0x1EFF) {
|
||||
self.latin_ext_letters += 1;
|
||||
} else {
|
||||
self.non_latin_letters += 1;
|
||||
}
|
||||
}
|
||||
prev = None;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Cosine similarity between the observed letter histogram and English
|
||||
/// letter frequencies. A shifted alphabet permutes the histogram, which
|
||||
/// destroys the similarity regardless of the shift amount.
|
||||
fn english_cosine(&self) -> f64 {
|
||||
if self.ascii_letters == 0 {
|
||||
return 1.0;
|
||||
}
|
||||
let n = self.ascii_letters as f64;
|
||||
let mut dot = 0.0;
|
||||
let mut norm_obs = 0.0;
|
||||
for (count, freq) in self.letter_counts.iter().zip(ENGLISH_LETTER_FREQ) {
|
||||
let p = *count as f64 / n;
|
||||
dot += p * freq;
|
||||
norm_obs += p * p;
|
||||
}
|
||||
let norm_en = ENGLISH_LETTER_FREQ
|
||||
.iter()
|
||||
.map(|f| f * f)
|
||||
.sum::<f64>()
|
||||
.sqrt();
|
||||
dot / (norm_obs.sqrt() * norm_en)
|
||||
}
|
||||
|
||||
/// Cosine similarity between the observed histogram and English
|
||||
/// frequencies after sorting BOTH descending — i.e. comparing the *shape*
|
||||
/// of the frequency profile, ignoring which letter sits where. A
|
||||
/// substitution cipher is a bijection, so it preserves this shape exactly
|
||||
/// (att10k 0.97, arbitrary shifts 0.99) regardless of case or offset.
|
||||
/// Non-linguistic ASCII has a different profile: a small alphabet is far
|
||||
/// steeper (random DNA 0.74, hex dumps 0.81), so the shape diverges.
|
||||
fn english_shape_cosine(&self) -> f64 {
|
||||
if self.ascii_letters == 0 {
|
||||
return 1.0;
|
||||
}
|
||||
let n = self.ascii_letters as f64;
|
||||
let mut obs: [f64; 26] = std::array::from_fn(|i| self.letter_counts[i] as f64 / n);
|
||||
obs.sort_unstable_by(|a, b| b.total_cmp(a));
|
||||
let mut en = ENGLISH_LETTER_FREQ;
|
||||
en.sort_unstable_by(|a, b| b.total_cmp(a));
|
||||
|
||||
let dot: f64 = obs.iter().zip(en).map(|(o, e)| o * e).sum();
|
||||
let norm_obs = obs.iter().map(|o| o * o).sum::<f64>().sqrt();
|
||||
let norm_en = en.iter().map(|e| e * e).sum::<f64>().sqrt();
|
||||
dot / (norm_obs * norm_en)
|
||||
}
|
||||
|
||||
/// Thresholds validated against the 380-document pdf-evals snapshot
|
||||
/// corpus (0 false positives) and the garbled ParseBench `att10k` page
|
||||
/// (vowel ratio 0.245, case-shift rate 0.225, cosine 0.532). Closest
|
||||
/// legitimate document on each axis: vowel ratio 0.264 (circuit
|
||||
/// schematic), case-shift rate 0.021, cosine 0.801.
|
||||
fn looks_garbled(&self) -> bool {
|
||||
// Need a statistically meaningful, Latin-dominant sample.
|
||||
if self.ascii_letters < 200
|
||||
|| self.non_latin_letters > self.ascii_letters + self.latin_ext_letters
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// Real Latin-script text keeps vowels above ~30% of letters even in
|
||||
// acronym- and part-number-heavy documents; shifted text starves them.
|
||||
let vowel_ratio = self.ascii_vowels as f64 / self.ascii_letters as f64;
|
||||
if vowel_ratio > 0.30 {
|
||||
return false;
|
||||
}
|
||||
|
||||
// Signal 1: lowercase→uppercase transitions inside words. A shifted
|
||||
// lowercase alphabet straddles the ASCII uppercase block ('i'→'Z',
|
||||
// 't'→'e'), so garbled words flip case constantly. Real documents
|
||||
// stay ≤ 0.02 even with camelCase identifiers.
|
||||
let case_shifts = self.letter_bigrams >= 100
|
||||
&& self.case_shift_bigrams as f64 >= self.letter_bigrams as f64 * 0.10;
|
||||
|
||||
// Signal 2: the histogram is a permutation of natural language — an
|
||||
// English-like frequency SHAPE (sorted cosine high) but with letters
|
||||
// in the wrong POSITIONS (unsorted cosine low). This is the signature
|
||||
// of a substitution cipher and is case-independent, so it catches
|
||||
// all-lowercase and all-uppercase shifts as well as case-straddling
|
||||
// ones. Genuinely non-linguistic ASCII that is merely "unlike English"
|
||||
// fails one of the two halves: DNA/hex dumps have too steep a profile
|
||||
// (shape cosine < 0.90), while protein sequences, ticker symbols and
|
||||
// base64 are not sufficiently unlike English in position (unsorted
|
||||
// cosine ≥ 0.60) — so none of them are routed to OCR.
|
||||
let permuted_language = self.english_cosine() < 0.60 && self.english_shape_cosine() >= 0.90;
|
||||
|
||||
case_shifts || permuted_language
|
||||
}
|
||||
}
|
||||
|
||||
#[derive(Debug, Default)]
|
||||
struct TextQualityReport {
|
||||
pages_needing_ocr: Vec<u32>,
|
||||
has_encoding_issues: bool,
|
||||
reasons_by_page: BTreeMap<u32, Vec<String>>,
|
||||
}
|
||||
|
||||
#[derive(Debug, Default)]
|
||||
struct PageTextQualityEvidence {
|
||||
chars: usize,
|
||||
replacement_chars: usize,
|
||||
replacement_spans: usize,
|
||||
longest_replacement_run: usize,
|
||||
cipher_garble: CipherGarbleStats,
|
||||
}
|
||||
|
||||
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
|
||||
enum TextSpanIssueKind {
|
||||
Replacement,
|
||||
Strong,
|
||||
}
|
||||
|
||||
fn analyze_text_quality(items: &[TextItem]) -> TextQualityReport {
|
||||
let mut reasons_by_page = BTreeMap::new();
|
||||
let mut evidence_by_page = BTreeMap::<u32, PageTextQualityEvidence>::new();
|
||||
|
||||
for item in items {
|
||||
if !matches!(item.item_type, crate::types::ItemType::Text) {
|
||||
continue;
|
||||
}
|
||||
|
||||
let evidence = evidence_by_page.entry(item.page).or_default();
|
||||
evidence.chars += item.text.chars().filter(|ch| !ch.is_whitespace()).count();
|
||||
evidence.cipher_garble.add_text(&item.text);
|
||||
|
||||
match text_span_decoding_issue_kind(&item.text) {
|
||||
Some(TextSpanIssueKind::Strong) => {
|
||||
add_ocr_reason(
|
||||
&mut reasons_by_page,
|
||||
item.page,
|
||||
OCR_REASON_SUSPECTED_GARBLED_TEXT,
|
||||
);
|
||||
}
|
||||
Some(TextSpanIssueKind::Replacement) => {
|
||||
let stats = replacement_text_stats(&item.text);
|
||||
evidence.replacement_chars += stats.0;
|
||||
evidence.replacement_spans += 1;
|
||||
evidence.longest_replacement_run = evidence.longest_replacement_run.max(stats.1);
|
||||
}
|
||||
None => {}
|
||||
}
|
||||
}
|
||||
|
||||
for (page, evidence) in evidence_by_page {
|
||||
if reasons_by_page.contains_key(&page) {
|
||||
continue;
|
||||
}
|
||||
if page_replacement_evidence_needs_ocr(&evidence) || evidence.cipher_garble.looks_garbled()
|
||||
{
|
||||
add_ocr_reason(
|
||||
&mut reasons_by_page,
|
||||
page,
|
||||
OCR_REASON_SUSPECTED_GARBLED_TEXT,
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
let pages_needing_ocr: Vec<u32> = reasons_by_page.keys().copied().collect();
|
||||
TextQualityReport {
|
||||
has_encoding_issues: !pages_needing_ocr.is_empty(),
|
||||
pages_needing_ocr,
|
||||
reasons_by_page,
|
||||
}
|
||||
}
|
||||
|
||||
fn suspected_garbled_reason() -> String {
|
||||
OCR_REASON_SUSPECTED_GARBLED_TEXT.to_string()
|
||||
}
|
||||
|
||||
fn add_ocr_reason(reasons_by_page: &mut BTreeMap<u32, Vec<String>>, page: u32, reason: &str) {
|
||||
pub(crate) fn add_ocr_reason(
|
||||
reasons_by_page: &mut BTreeMap<u32, Vec<String>>,
|
||||
page: u32,
|
||||
reason: &str,
|
||||
) {
|
||||
let reasons = reasons_by_page.entry(page).or_default();
|
||||
if !reasons.iter().any(|existing| existing == reason) {
|
||||
reasons.push(reason.to_string());
|
||||
@@ -4011,225 +3748,6 @@ fn page_ocr_reasons_vec(reasons_by_page: BTreeMap<u32, Vec<String>>) -> Vec<Page
|
||||
.collect()
|
||||
}
|
||||
|
||||
fn region_items_have_decoding_issue(items: &[TextItem]) -> bool {
|
||||
items.iter().any(|item| {
|
||||
matches!(item.item_type, crate::types::ItemType::Text)
|
||||
&& text_span_has_decoding_issue(&item.text)
|
||||
})
|
||||
}
|
||||
|
||||
fn text_span_has_decoding_issue(text: &str) -> bool {
|
||||
text_span_decoding_issue_kind(text).is_some()
|
||||
}
|
||||
|
||||
fn text_span_decoding_issue_kind(text: &str) -> Option<TextSpanIssueKind> {
|
||||
let text = text.trim();
|
||||
if text.is_empty() {
|
||||
return None;
|
||||
}
|
||||
|
||||
if has_dollar_as_space_pattern(text)
|
||||
|| has_private_use_text_run(text)
|
||||
|| is_cid_garbage(text)
|
||||
|| has_cid_control_token(text)
|
||||
{
|
||||
return Some(TextSpanIssueKind::Strong);
|
||||
}
|
||||
|
||||
if has_replacement_text_run(text) {
|
||||
return Some(TextSpanIssueKind::Replacement);
|
||||
}
|
||||
|
||||
None
|
||||
}
|
||||
|
||||
fn replacement_text_stats(text: &str) -> (usize, usize) {
|
||||
let mut replacement = 0usize;
|
||||
let mut current_run = 0usize;
|
||||
let mut longest_run = 0usize;
|
||||
|
||||
for ch in text.chars() {
|
||||
if ch == '\u{FFFD}' {
|
||||
replacement += 1;
|
||||
current_run += 1;
|
||||
longest_run = longest_run.max(current_run);
|
||||
} else {
|
||||
current_run = 0;
|
||||
}
|
||||
}
|
||||
|
||||
(replacement, longest_run)
|
||||
}
|
||||
|
||||
fn page_replacement_evidence_needs_ocr(evidence: &PageTextQualityEvidence) -> bool {
|
||||
if evidence.replacement_chars == 0 || evidence.chars == 0 {
|
||||
return false;
|
||||
}
|
||||
|
||||
// If the entire page is only a short broken text layer, even a short
|
||||
// replacement run is enough evidence. On otherwise text-heavy pages,
|
||||
// require density so math formulas do not force full-page OCR.
|
||||
if evidence.chars <= 80 && evidence.longest_replacement_run >= 2 {
|
||||
return true;
|
||||
}
|
||||
|
||||
let replacement_density_bps = evidence.replacement_chars * 10_000 / evidence.chars;
|
||||
let enough_bad_text = evidence.replacement_chars >= 12 && replacement_density_bps >= 500;
|
||||
let repeated_bad_spans = evidence.replacement_spans >= 3 && replacement_density_bps >= 250;
|
||||
let long_bad_run = evidence.longest_replacement_run >= 8 && replacement_density_bps >= 250;
|
||||
|
||||
enough_bad_text || repeated_bad_spans || long_bad_run
|
||||
}
|
||||
|
||||
fn has_replacement_text_run(text: &str) -> bool {
|
||||
let (replacement, longest_run) = replacement_text_stats(text);
|
||||
longest_run >= 2 || replacement >= 3
|
||||
}
|
||||
|
||||
fn has_private_use_text_run(text: &str) -> bool {
|
||||
let mut total = 0usize;
|
||||
let mut private_use = 0usize;
|
||||
let mut current_run = 0usize;
|
||||
let mut longest_run = 0usize;
|
||||
|
||||
for ch in text.chars() {
|
||||
if ch.is_whitespace() {
|
||||
current_run = 0;
|
||||
continue;
|
||||
}
|
||||
total += 1;
|
||||
if is_private_use_char(ch) {
|
||||
private_use += 1;
|
||||
current_run += 1;
|
||||
longest_run = longest_run.max(current_run);
|
||||
} else {
|
||||
current_run = 0;
|
||||
}
|
||||
}
|
||||
|
||||
if private_use == 0 {
|
||||
return false;
|
||||
}
|
||||
|
||||
longest_run >= 3 || (total >= 5 && private_use >= 2 && private_use * 2 >= total)
|
||||
}
|
||||
|
||||
fn has_cid_control_token(text: &str) -> bool {
|
||||
text.split_whitespace().any(token_has_cid_control)
|
||||
}
|
||||
|
||||
fn token_has_cid_control(token: &str) -> bool {
|
||||
let mut total = 0usize;
|
||||
let mut c1_control = 0usize;
|
||||
|
||||
for ch in token.chars() {
|
||||
total += 1;
|
||||
if ('\u{0080}'..='\u{009F}').contains(&ch) {
|
||||
c1_control += 1;
|
||||
}
|
||||
}
|
||||
|
||||
total >= 5 && c1_control >= 2 && c1_control * 20 >= total
|
||||
}
|
||||
|
||||
fn is_private_use_char(ch: char) -> bool {
|
||||
matches!(
|
||||
ch as u32,
|
||||
0xE000..=0xF8FF | 0xF0000..=0xFFFFD | 0x100000..=0x10FFFD
|
||||
)
|
||||
}
|
||||
|
||||
/// Check if extracted text is predominantly garbage (non-alphanumeric).
|
||||
///
|
||||
/// Broken font encodings produce text like "----1-.-.-.___ --.-. .._ I_---."
|
||||
/// where most characters are punctuation/symbols. Real text in any language
|
||||
/// has >50% alphanumeric characters.
|
||||
fn is_garbage_text(markdown: &str) -> bool {
|
||||
let mut alphanum = 0usize;
|
||||
let mut non_alphanum = 0usize;
|
||||
|
||||
let chars: Vec<char> = markdown.chars().collect();
|
||||
let mut i = 0usize;
|
||||
while i < chars.len() {
|
||||
let ch = chars[i];
|
||||
let mut run_end = i + 1;
|
||||
while run_end < chars.len() && chars[run_end] == ch {
|
||||
run_end += 1;
|
||||
}
|
||||
|
||||
let is_decorative_leader = matches!(ch, '.' | '_' | '·') && run_end - i >= 3;
|
||||
if !is_decorative_leader {
|
||||
for &run_ch in &chars[i..run_end] {
|
||||
if run_ch.is_whitespace() {
|
||||
continue;
|
||||
}
|
||||
// Skip markdown syntax chars that we add (not from the PDF)
|
||||
if matches!(run_ch, '#' | '*' | '|' | '-' | '\n') {
|
||||
continue;
|
||||
}
|
||||
if run_ch.is_alphanumeric() {
|
||||
alphanum += 1;
|
||||
} else {
|
||||
non_alphanum += 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
i = run_end;
|
||||
}
|
||||
|
||||
let total = alphanum + non_alphanum;
|
||||
total >= 50 && alphanum * 2 < total
|
||||
}
|
||||
|
||||
/// Detect garbage from failed CID-to-Unicode mapping on Identity-H fonts.
|
||||
///
|
||||
/// When CID values don't correspond to Unicode codepoints, the raw bytes often
|
||||
/// produce characters in the C1 control range (U+0080–U+009F) or Private Use
|
||||
/// Area, mixed with random Latin Extended characters. Valid text in any
|
||||
/// language almost never contains C1 controls. We also fall back to the
|
||||
/// general `is_garbage_text` check for non-alphanumeric-heavy patterns.
|
||||
fn is_cid_garbage(text: &str) -> bool {
|
||||
if is_garbage_text(text) {
|
||||
return true;
|
||||
}
|
||||
let mut total = 0usize;
|
||||
let mut c1_control = 0usize;
|
||||
let mut high_latin = 0usize;
|
||||
for ch in text.chars() {
|
||||
if ch.is_whitespace() {
|
||||
continue;
|
||||
}
|
||||
total += 1;
|
||||
// C1 control characters (U+0080–U+009F) — almost never in real text
|
||||
if ch == '·' {
|
||||
continue;
|
||||
}
|
||||
if ('\u{0080}'..='\u{009F}').contains(&ch) {
|
||||
c1_control += 1;
|
||||
}
|
||||
// High Latin-1 (U+00A0–U+00FF) — legitimate in Western European text
|
||||
// but when combined with ASCII in CID passthrough, indicates mojibake
|
||||
// from CID values being misinterpreted as Latin-1 characters.
|
||||
if ('\u{00A0}'..='\u{00FF}').contains(&ch) {
|
||||
high_latin += 1;
|
||||
}
|
||||
}
|
||||
if total < 5 {
|
||||
return false;
|
||||
}
|
||||
// If ≥5% of non-whitespace chars are C1 controls, it's garbage
|
||||
if c1_control >= 2 && c1_control * 20 >= total {
|
||||
return true;
|
||||
}
|
||||
// If ≥40% of non-whitespace chars are high Latin-1 AND the text has few
|
||||
// ASCII letters, it's likely CID-as-Latin-1 mojibake (Japanese/CJK PDFs
|
||||
// where CID values 0x80-0xFF become accented Latin characters). Keep a
|
||||
// minimum length so short math tokens like "2×()×" do not route a clean
|
||||
// page to OCR.
|
||||
let ascii_letters = text.chars().filter(|c| c.is_ascii_alphabetic()).count();
|
||||
total >= 20 && high_latin * 5 >= total * 2 && ascii_letters * 3 < total
|
||||
}
|
||||
|
||||
/// Detect markdown tables with suspicious structure that suggest the heuristic
|
||||
/// missed/mangled rows or columns. Returns true when the caller should treat
|
||||
/// the result as `needs_ocr` and fall back to GPU OCR.
|
||||
|
||||
@@ -153,6 +153,85 @@ pub(crate) fn is_toc_entry_line(text: &str) -> bool {
|
||||
dots >= 3
|
||||
}
|
||||
|
||||
/// A heading that announces a table of contents ("Contents", "Table of
|
||||
/// Contents"). Lines after it on the same page are ToC entries — section
|
||||
/// titles that look exactly like headings but must not be promoted.
|
||||
pub(crate) fn is_toc_marker_heading(text: &str) -> bool {
|
||||
let t = text.trim().trim_end_matches(':').trim().to_lowercase();
|
||||
matches!(t.as_str(), "contents" | "table of contents")
|
||||
}
|
||||
|
||||
/// Lines that resemble headings structurally but are display-math fragments:
|
||||
/// equations ending in an equation number ("S = kB ln W, (2)") or equation
|
||||
/// lead-ins ("Rearranging Equation (8) gives:"). Both carry an "(N)" equation
|
||||
/// reference — but a trailing "(N)" alone is not enough: real headings end
|
||||
/// with parenthesized numbers too ("Nicaea (325)", appendix numbering), so
|
||||
/// the suffix form additionally requires math evidence — an "=" in the line
|
||||
/// or a comma immediately before the number, both present in every display
|
||||
/// equation and absent from name-plus-number headings. A bare trailing colon
|
||||
/// is NOT a fragment signal either: real headings frequently end with colons
|
||||
/// ("Procedure:", "Steps for Using the Microscope:").
|
||||
pub(crate) fn is_heading_fragment(text: &str) -> bool {
|
||||
let t = text.trim_end();
|
||||
|
||||
fn is_equation_number(s: &str) -> bool {
|
||||
s.strip_prefix('(')
|
||||
.and_then(|r| r.strip_suffix(')'))
|
||||
.is_some_and(|inner| {
|
||||
!inner.is_empty() && inner.len() <= 3 && inner.chars().all(|c| c.is_ascii_digit())
|
||||
})
|
||||
}
|
||||
|
||||
// Equation-number suffix with math evidence: "S = kB ln W, (2)"
|
||||
let mut rev = t.rsplit(' ');
|
||||
let last = rev.next().unwrap_or("");
|
||||
if is_equation_number(last) {
|
||||
// Page-of-total running headers: "LIVSMEDELSVERKET PM 2 (10)"
|
||||
if let Some(prev_word) = t.rsplit(' ').nth(1) {
|
||||
if let (Ok(page), Some(total)) = (
|
||||
prev_word.parse::<u32>(),
|
||||
last.trim_start_matches('(')
|
||||
.trim_end_matches(')')
|
||||
.parse::<u32>()
|
||||
.ok(),
|
||||
) {
|
||||
if page <= total {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
}
|
||||
let punct_before = rev
|
||||
.next()
|
||||
.is_some_and(|w| w.ends_with(',') || w.ends_with(':'));
|
||||
let has_math_op = t.chars().any(|c| {
|
||||
matches!(
|
||||
c,
|
||||
'=' | '<'
|
||||
| '>'
|
||||
| '≤'
|
||||
| '≥'
|
||||
| '≪'
|
||||
| '≫'
|
||||
| '≈'
|
||||
| '≠'
|
||||
| '±'
|
||||
| '∑'
|
||||
| '∫'
|
||||
| '√'
|
||||
| '∝'
|
||||
)
|
||||
});
|
||||
if punct_before || has_math_op {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
// Lead-in: ends with a colon AND references an equation number inline
|
||||
if t.ends_with(':') && t.split_whitespace().any(is_equation_number) {
|
||||
return true;
|
||||
}
|
||||
false
|
||||
}
|
||||
|
||||
/// Compute the Y-gap threshold for paragraph break detection.
|
||||
///
|
||||
/// Instead of using a fixed multiple of base_size (which fails for double-spaced
|
||||
@@ -367,4 +446,41 @@ mod tests {
|
||||
// Long numbers are data, not page refs
|
||||
assert!(!is_toc_entry_line("ISBN ... 97814"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn toc_marker_headings() {
|
||||
assert!(is_toc_marker_heading("Contents"));
|
||||
assert!(is_toc_marker_heading("CONTENTS"));
|
||||
assert!(is_toc_marker_heading("Table of Contents"));
|
||||
assert!(is_toc_marker_heading("Table of contents:"));
|
||||
assert!(!is_toc_marker_heading("Contents of the Shipment"));
|
||||
assert!(!is_toc_marker_heading("Introduction"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn heading_fragments() {
|
||||
// Equation lead-ins: colon ending + inline equation reference
|
||||
assert!(is_heading_fragment("Rearranging Equation (8) gives:"));
|
||||
// Display-equation neighbours ending in an equation number
|
||||
assert!(is_heading_fragment("S = kB ln W, (2)"));
|
||||
assert!(is_heading_fragment("E = mc2 (12)"));
|
||||
assert!(is_heading_fragment("x + y = z, (3)"));
|
||||
// Page-of-total running headers
|
||||
assert!(is_heading_fragment("LIVSMEDELSVERKET PM 2 (10)"));
|
||||
// Comparison-operator evidence and colon-before-number
|
||||
assert!(is_heading_fragment(
|
||||
"PLL\u{fe} PHH\u{226a} PLH\u{fe} PHL: (12)"
|
||||
));
|
||||
// Real headings pass — including name-plus-number and colon-ended ones
|
||||
assert!(!is_heading_fragment("Nicaea (325)"));
|
||||
assert!(!is_heading_fragment(
|
||||
"\u{627}\u{644}\u{645}\u{644}\u{62d}\u{642} \u{631}\u{642}\u{645} (1)"
|
||||
));
|
||||
assert!(!is_heading_fragment("4. Entropy"));
|
||||
assert!(!is_heading_fragment("Procedure:"));
|
||||
assert!(!is_heading_fragment("Steps for Using the Microscope:"));
|
||||
assert!(!is_heading_fragment("Changing objectives:"));
|
||||
assert!(!is_heading_fragment("Sales by Region (2024)"));
|
||||
assert!(!is_heading_fragment("Results (preliminary)"));
|
||||
}
|
||||
}
|
||||
|
||||
+64
-3
@@ -7,7 +7,8 @@ use crate::types::TextLine;
|
||||
|
||||
use super::analysis::{
|
||||
bold_heading_level, calculate_font_stats, compute_heading_tiers, compute_paragraph_threshold,
|
||||
detect_header_level, font_size_rarity, has_dot_leaders, is_toc_entry_line,
|
||||
detect_header_level, font_size_rarity, has_dot_leaders, is_heading_fragment, is_toc_entry_line,
|
||||
is_toc_marker_heading,
|
||||
};
|
||||
use super::classify::{
|
||||
format_list_item, is_caption_line, is_list_item, is_monospace_font, starts_with_bullet_marker,
|
||||
@@ -140,8 +141,11 @@ fn find_isolated_lines(lines: &[TextLine], base_size: f32, para_threshold: f32)
|
||||
}
|
||||
}
|
||||
for (&page, &(total, isolated)) in &page_line_counts {
|
||||
if total > 0 && isolated as f32 / total as f32 > 0.25 {
|
||||
// Too many isolated lines on this page — remove them all
|
||||
// The ratio only means something on pages dense enough for a
|
||||
// multi-column misfire; on sparse pages (covers, ToC pages with a
|
||||
// lone title) one isolated line is 25%+ of the page and exactly the
|
||||
// line isolation exists to find.
|
||||
if total >= 10 && isolated as f32 / total as f32 > 0.25 {
|
||||
set.retain(|&i| lines[i].page != page);
|
||||
}
|
||||
}
|
||||
@@ -486,6 +490,7 @@ pub(super) fn to_markdown_from_lines_with_tables_and_images(
|
||||
let mut in_code_block = false;
|
||||
let mut prev_had_dot_leaders = false;
|
||||
let mut paragraph_in_wrapped_bold_run = false;
|
||||
let mut toc_suppress_page: Option<u32> = None;
|
||||
let mut inserted_tables: HashSet<(u32, usize)> = HashSet::new();
|
||||
let mut inserted_images: HashSet<(u32, usize)> = HashSet::new();
|
||||
|
||||
@@ -698,12 +703,22 @@ pub(super) fn to_markdown_from_lines_with_tables_and_images(
|
||||
_ => false,
|
||||
};
|
||||
|
||||
// Lines explicitly tagged with a non-heading content role must never
|
||||
// be promoted by the visual heuristic — a tagged list item, quote, or
|
||||
// code line can look exactly like a heading (short, isolated).
|
||||
let non_heading_role = struct_role
|
||||
.as_ref()
|
||||
.is_some_and(StructRole::is_non_heading_content);
|
||||
let heuristic_heading = if options.detect_headers
|
||||
&& !non_heading_role
|
||||
&& !is_code_line
|
||||
&& !looks_like_list_continuation
|
||||
&& plain_trimmed.len() > 3
|
||||
&& plain_trimmed.split_whitespace().count() <= 15
|
||||
&& !starts_with_bullet_marker(plain_trimmed)
|
||||
&& !is_toc_entry_line(plain_trimmed)
|
||||
&& !is_heading_fragment(plain_trimmed)
|
||||
&& toc_suppress_page != Some(line.page)
|
||||
{
|
||||
let line_font_size = line.items.first().map(|i| i.font_size).unwrap_or(base_size);
|
||||
detect_header_level(line_font_size, base_size, &heading_tiers).or_else(|| {
|
||||
@@ -768,6 +783,9 @@ pub(super) fn to_markdown_from_lines_with_tables_and_images(
|
||||
plain_text.clone()
|
||||
};
|
||||
output.push_str(&format!("{} {}\n\n", prefix, heading_text.trim()));
|
||||
if is_toc_marker_heading(plain_trimmed) {
|
||||
toc_suppress_page = Some(line.page);
|
||||
}
|
||||
in_list = false;
|
||||
continue;
|
||||
}
|
||||
@@ -964,6 +982,7 @@ pub fn to_markdown_from_lines(lines: Vec<TextLine>, options: MarkdownOptions) ->
|
||||
let mut last_list_x: Option<f32> = None;
|
||||
let mut prev_had_dot_leaders = false;
|
||||
let mut paragraph_in_wrapped_bold_run = false;
|
||||
let mut toc_suppress_page: Option<u32> = None;
|
||||
|
||||
for (line_idx, line) in lines.iter().enumerate() {
|
||||
// Page break
|
||||
@@ -1043,6 +1062,9 @@ pub fn to_markdown_from_lines(lines: Vec<TextLine>, options: MarkdownOptions) ->
|
||||
&& plain_trimmed.len() > 3
|
||||
&& plain_trimmed.split_whitespace().count() <= 15
|
||||
&& !is_toc_entry_line(plain_trimmed)
|
||||
&& !is_heading_fragment(plain_trimmed)
|
||||
&& toc_suppress_page != Some(line.page)
|
||||
&& !(options.detect_code && line.items.iter().any(|i| is_monospace_font(&i.font)))
|
||||
{
|
||||
let line_font_size = line.items.first().map(|i| i.font_size).unwrap_or(base_size);
|
||||
if let Some(header_level) =
|
||||
@@ -1086,6 +1108,9 @@ pub fn to_markdown_from_lines(lines: Vec<TextLine>, options: MarkdownOptions) ->
|
||||
plain_text.clone()
|
||||
};
|
||||
output.push_str(&format!("{} {}\n\n", prefix, heading_text.trim()));
|
||||
if is_toc_marker_heading(plain_trimmed) {
|
||||
toc_suppress_page = Some(line.page);
|
||||
}
|
||||
in_list = false;
|
||||
continue;
|
||||
}
|
||||
@@ -1214,6 +1239,42 @@ mod tests {
|
||||
}
|
||||
}
|
||||
|
||||
fn line_at(text: &str, page: u32, y: f32) -> TextLine {
|
||||
let mut item = make_item(text, page, None);
|
||||
item.y = y;
|
||||
make_line(vec![item])
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn isolated_lines_kept_on_sparse_pages() {
|
||||
// A ToC page with a lone title and one entry far below: the density
|
||||
// ratio is 50% but the page is too sparse for the multi-column
|
||||
// misfire the guard targets — the title must stay isolated.
|
||||
let lines = vec![
|
||||
line_at("CONTENTS", 1, 700.0),
|
||||
line_at("Chapter One 5", 1, 500.0),
|
||||
];
|
||||
let isolated = find_isolated_lines(&lines, 12.0, 20.0);
|
||||
assert!(
|
||||
isolated.contains(&0),
|
||||
"sparse-page title must stay isolated"
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn isolated_lines_wiped_on_dense_pages() {
|
||||
// 12 short lines all with paragraph gaps — the multi-column misfire
|
||||
// shape. The guard must clear them all.
|
||||
let lines: Vec<TextLine> = (0..12)
|
||||
.map(|i| line_at("Short column line", 1, 700.0 - i as f32 * 50.0))
|
||||
.collect();
|
||||
let isolated = find_isolated_lines(&lines, 12.0, 20.0);
|
||||
assert!(
|
||||
isolated.is_empty(),
|
||||
"dense page of isolated lines must be wiped"
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_struct_role_heading() {
|
||||
let lines = vec![
|
||||
|
||||
+100
-1
@@ -3,7 +3,7 @@
|
||||
use std::collections::{HashMap, HashSet};
|
||||
|
||||
use crate::structure_tree::StructRole;
|
||||
use crate::types::TextLine;
|
||||
use crate::types::{TextItem, TextLine};
|
||||
|
||||
use super::analysis::detect_header_level;
|
||||
|
||||
@@ -87,6 +87,41 @@ pub(crate) fn merge_heading_lines(
|
||||
false
|
||||
};
|
||||
|
||||
// Bold headings at body font size never reach a tier, so wrapped ones
|
||||
// split into two output headings ("…of wood pellets and cost" /
|
||||
// "structure in Japan"). Merge a fully-bold line into the previous
|
||||
// fully-bold line when it reads as a wrap continuation: starts
|
||||
// lowercase, tiny Y gap, and the previous line has no terminal
|
||||
// punctuation. Kept deliberately narrow — bold list labels and bold
|
||||
// sentences start with markers or capitals and are unaffected.
|
||||
let should_merge = should_merge
|
||||
|| if let Some(prev) = result.last() {
|
||||
let all_bold = |l: &TextLine| {
|
||||
!l.items.is_empty() && l.items.iter().all(|i: &TextItem| i.is_bold)
|
||||
};
|
||||
let prev_text = prev.text();
|
||||
let prev_trim = prev_text.trim_end();
|
||||
let curr_text = line.text();
|
||||
let curr_trim = curr_text.trim();
|
||||
let y_gap = prev.y - line.y;
|
||||
// Both lines must be tier-less: a tiered/tagged bold heading
|
||||
// followed by bold body text must not absorb it.
|
||||
line_level.is_none()
|
||||
&& effective_heading_level(prev, base_size, heading_tiers, struct_roles)
|
||||
.is_none()
|
||||
&& prev.page == line.page
|
||||
&& all_bold(prev)
|
||||
&& all_bold(&line)
|
||||
&& y_gap > 0.0
|
||||
&& y_gap < line_font * 1.6
|
||||
&& curr_trim.chars().next().is_some_and(|c| c.is_lowercase())
|
||||
&& !prev_trim.ends_with(['.', ':', ';', '!', '?'])
|
||||
&& prev_trim.split_whitespace().count() + curr_trim.split_whitespace().count()
|
||||
<= 20
|
||||
} else {
|
||||
false
|
||||
};
|
||||
|
||||
if should_merge {
|
||||
// Append this line's items to the previous line
|
||||
let prev = result.last_mut().unwrap();
|
||||
@@ -685,4 +720,68 @@ mod tests {
|
||||
.unwrap();
|
||||
assert_eq!(first_header.page, 1, "first occurrence should be on page 1");
|
||||
}
|
||||
|
||||
fn make_bold_line(text: &str, page: u32, y: f32) -> TextLine {
|
||||
let mut item = make_item(text, 12.0, None);
|
||||
item.is_bold = true;
|
||||
TextLine {
|
||||
items: vec![item],
|
||||
y,
|
||||
page,
|
||||
adaptive_threshold: 0.10,
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn merge_wrapped_bold_heading_lowercase_continuation() {
|
||||
// Bold-at-body-size heading wrapped across two lines: the second line
|
||||
// starts lowercase and must merge into the first.
|
||||
let lines = vec![
|
||||
make_bold_line(
|
||||
"3. Perspective of supply and demand balance and cost",
|
||||
1,
|
||||
700.0,
|
||||
),
|
||||
make_bold_line("structure in Japan", 1, 686.0),
|
||||
make_line("Body text paragraph follows here.", 12.0, 1, 660.0, None),
|
||||
];
|
||||
let result = merge_heading_lines(lines, 12.0, &[], None);
|
||||
assert_eq!(result.len(), 2, "wrapped bold heading should merge");
|
||||
assert!(result[0].text().contains("cost structure in Japan"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn no_merge_for_bold_sentences_or_new_headings() {
|
||||
// Second bold line starts with a capital — a new heading or label,
|
||||
// not a wrap continuation.
|
||||
let lines = vec![
|
||||
make_bold_line("Replace", 1, 700.0),
|
||||
make_bold_line("Trash", 1, 686.0),
|
||||
];
|
||||
let result = merge_heading_lines(lines, 12.0, &[], None);
|
||||
assert_eq!(result.len(), 2, "distinct bold lines must not merge");
|
||||
|
||||
// Previous line ends a sentence — continuation must not merge.
|
||||
let lines = vec![
|
||||
make_bold_line("This is a bold sentence.", 1, 700.0),
|
||||
make_bold_line("another bold line", 1, 686.0),
|
||||
];
|
||||
let result = merge_heading_lines(lines, 12.0, &[], None);
|
||||
assert_eq!(result.len(), 2, "sentence-final bold line must not merge");
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn tiered_bold_heading_does_not_absorb_bold_body() {
|
||||
// Previous line is a tier-level bold heading (16pt vs 12pt body);
|
||||
// a following lowercase bold body line must NOT merge into it.
|
||||
let mut heading = make_bold_line("Section Title", 1, 700.0);
|
||||
heading.items[0].font_size = 16.0;
|
||||
heading.items[0].height = 16.0;
|
||||
let lines = vec![
|
||||
heading,
|
||||
make_bold_line("emphasized body text continues here", 1, 686.0),
|
||||
];
|
||||
let result = merge_heading_lines(lines, 12.0, &[16.0], None);
|
||||
assert_eq!(result.len(), 2, "tiered heading must not absorb bold body");
|
||||
}
|
||||
}
|
||||
|
||||
@@ -76,6 +76,52 @@ pub enum StructRole {
|
||||
}
|
||||
|
||||
impl StructRole {
|
||||
/// Content roles whose text must never be promoted to a heading by the
|
||||
/// visual heuristic. These carry an explicit non-heading meaning in the
|
||||
/// struct tree (lists, quotes, notes, references, captions, formulas,
|
||||
/// forms, ToC entries), yet their text is often short and visually
|
||||
/// isolated — exactly what the heuristic keys on. Heading roles (H, H1–H6)
|
||||
/// and generic container/flow roles (P, Div, Sect, Span, …) are excluded
|
||||
/// so the heuristic can still fire there.
|
||||
///
|
||||
/// `Figure` is deliberately NOT in this set: cover/banner pages routinely
|
||||
/// tag the document title inside a Figure (alongside a seal or logo), and
|
||||
/// that title is a real heading. `Formula` and `Form` stay — a line
|
||||
/// explicitly tagged as an equation or form field is never a heading.
|
||||
///
|
||||
/// Table roles (Table/TR/TH/TD/THead/TBody/TFoot) are included so that
|
||||
/// when table reconstruction falls back and cells reach the line loop as
|
||||
/// plain text, a short isolated cell — a `TH` column header especially —
|
||||
/// is not promoted to a heading.
|
||||
pub(crate) fn is_non_heading_content(&self) -> bool {
|
||||
matches!(
|
||||
self,
|
||||
Self::L
|
||||
| Self::LI
|
||||
| Self::Lbl
|
||||
| Self::LBody
|
||||
| Self::BlockQuote
|
||||
| Self::Quote
|
||||
| Self::Caption
|
||||
| Self::TOC
|
||||
| Self::TOCI
|
||||
| Self::Index
|
||||
| Self::Note
|
||||
| Self::Reference
|
||||
| Self::BibEntry
|
||||
| Self::Code
|
||||
| Self::Formula
|
||||
| Self::Form
|
||||
| Self::Table
|
||||
| Self::TR
|
||||
| Self::TH
|
||||
| Self::TD
|
||||
| Self::THead
|
||||
| Self::TBody
|
||||
| Self::TFoot
|
||||
)
|
||||
}
|
||||
|
||||
fn from_name(name: &str) -> Self {
|
||||
match name {
|
||||
"Document" => Self::Document,
|
||||
@@ -856,6 +902,54 @@ fn contains_bytes(haystack: &[u8], needle: &[u8]) -> bool {
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn non_heading_content_roles() {
|
||||
for r in [
|
||||
StructRole::L,
|
||||
StructRole::LI,
|
||||
StructRole::BlockQuote,
|
||||
StructRole::Quote,
|
||||
StructRole::Caption,
|
||||
StructRole::TOC,
|
||||
StructRole::TOCI,
|
||||
StructRole::Index,
|
||||
StructRole::Note,
|
||||
StructRole::Reference,
|
||||
StructRole::BibEntry,
|
||||
StructRole::Code,
|
||||
StructRole::Formula,
|
||||
StructRole::Form,
|
||||
StructRole::Table,
|
||||
StructRole::TR,
|
||||
StructRole::TH,
|
||||
StructRole::TD,
|
||||
StructRole::THead,
|
||||
StructRole::TBody,
|
||||
StructRole::TFoot,
|
||||
] {
|
||||
assert!(
|
||||
r.is_non_heading_content(),
|
||||
"{r:?} should block heading promotion"
|
||||
);
|
||||
}
|
||||
// Heading and generic container/flow roles must NOT block promotion
|
||||
for r in [
|
||||
StructRole::H,
|
||||
StructRole::H1,
|
||||
StructRole::H3,
|
||||
StructRole::P,
|
||||
StructRole::Div,
|
||||
StructRole::Sect,
|
||||
StructRole::Span,
|
||||
StructRole::Figure,
|
||||
] {
|
||||
assert!(
|
||||
!r.is_non_heading_content(),
|
||||
"{r:?} should allow heading promotion"
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_struct_role_from_name() {
|
||||
assert_eq!(StructRole::from_name("H1"), StructRole::H1);
|
||||
|
||||
@@ -0,0 +1,520 @@
|
||||
//! Text-quality detection: deciding when an extracted text layer is too broken
|
||||
//! to serve and a page should fall back to OCR.
|
||||
//!
|
||||
//! Extraction can produce plausible-looking bytes that are actually garbage —
|
||||
//! failed CID→Unicode mappings, broken ToUnicode CMaps, mojibake. These
|
||||
//! detectors catch that and let callers set `needs_ocr`. They come in two
|
||||
//! layers, sharing the same primitives:
|
||||
//!
|
||||
//! - **Markdown-level** ([`detect_encoding_issues`], [`is_garbage_text`],
|
||||
//! [`is_cid_garbage`]) run on a page's final markdown string. Used as a
|
||||
//! backstop on the region-extraction and whole-document paths.
|
||||
//! - **Item/span-level** ([`analyze_text_quality`],
|
||||
//! [`region_items_have_decoding_issue`]) run on individual `TextItem`s and
|
||||
//! accumulate per-page evidence, so localized garbled spans on an otherwise
|
||||
//! clean page are caught without a single span having to condemn the page.
|
||||
//!
|
||||
//! Detection classes, roughly by signal:
|
||||
//! - **Replacement runs**: U+FFFD clusters ([`has_replacement_text_run`]).
|
||||
//! - **Private-use / C1-control runs**: CID passthrough landing in PUA or the
|
||||
//! C1 block ([`has_private_use_text_run`], [`has_cid_control_token`]).
|
||||
//! - **Dollar-as-space**: `Word$Word$Word` from broken CMaps
|
||||
//! ([`has_dollar_as_space_pattern`]).
|
||||
//! - **Non-alphanumeric dominance**: symbol soup ([`is_garbage_text`]).
|
||||
//! - **Substitution-cipher letter statistics**: pure-ASCII output whose letter
|
||||
//! distribution is a permutation of natural language ([`CipherGarbleStats`]).
|
||||
|
||||
use crate::types::TextItem;
|
||||
use crate::{add_ocr_reason, OCR_REASON_SUSPECTED_GARBLED_TEXT};
|
||||
use std::collections::BTreeMap;
|
||||
|
||||
/// Detect broken font encodings in extracted markdown text.
|
||||
///
|
||||
/// Two heuristics:
|
||||
/// 1. **U+FFFD**: Any replacement character indicates decode failures.
|
||||
/// 2. **Dollar-as-space**: Pattern like `Word$Word$Word` where `$` is used as a
|
||||
/// word separator due to broken ToUnicode CMaps. Triggers when either:
|
||||
/// - More than 50% of `$` are between letters (clear substitution pattern), OR
|
||||
/// - More than 20 letter-dollar-letter occurrences (even if some `$` are also
|
||||
/// used as trailing/leading separators, 20+ is far beyond normal financial text).
|
||||
pub(crate) fn detect_encoding_issues(markdown: &str) -> bool {
|
||||
// Heuristic 1: U+FFFD replacement characters
|
||||
if markdown.contains('\u{FFFD}') {
|
||||
return true;
|
||||
}
|
||||
|
||||
// Heuristic 2: dollar-as-space pattern
|
||||
if has_dollar_as_space_pattern(markdown) {
|
||||
return true;
|
||||
}
|
||||
|
||||
// Heuristic 3: substitution-cipher letter statistics (broken ToUnicode)
|
||||
let mut stats = CipherGarbleStats::default();
|
||||
stats.add_text(markdown);
|
||||
stats.looks_garbled()
|
||||
}
|
||||
|
||||
fn has_dollar_as_space_pattern(markdown: &str) -> bool {
|
||||
let total_dollars = markdown.matches('$').count();
|
||||
if total_dollars > 10 {
|
||||
let bytes = markdown.as_bytes();
|
||||
let mut letter_dollar_letter = 0usize;
|
||||
for i in 1..bytes.len().saturating_sub(1) {
|
||||
if bytes[i] == b'$'
|
||||
&& bytes[i - 1].is_ascii_alphabetic()
|
||||
&& bytes[i + 1].is_ascii_alphabetic()
|
||||
{
|
||||
letter_dollar_letter += 1;
|
||||
}
|
||||
}
|
||||
if letter_dollar_letter > 20 || letter_dollar_letter * 2 > total_dollars {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
||||
false
|
||||
}
|
||||
|
||||
/// English letter frequencies (percent, a–z). Used as a natural-language
|
||||
/// reference: every Latin-script language in the eval corpus (Swedish,
|
||||
/// Finnish, Turkish, German, romaji) scores ≥ 0.80 cosine similarity against
|
||||
/// it, while substitution-cipher text scores ~0.53.
|
||||
const ENGLISH_LETTER_FREQ: [f64; 26] = [
|
||||
8.2, 1.5, 2.8, 4.3, 12.7, 2.2, 2.0, 6.1, 7.0, 0.15, 0.8, 4.0, 2.4, 6.7, 7.5, 1.9, 0.1, 6.0,
|
||||
6.3, 9.1, 2.8, 1.0, 2.4, 0.15, 2.0, 0.07,
|
||||
];
|
||||
|
||||
/// Letter statistics for detecting substitution-cipher garbling: broken
|
||||
/// ToUnicode CMaps that shift every character by a per-range constant (e.g.
|
||||
/// `Certificate` extracted as `8VceZWZTReV`). Such text is 100% printable
|
||||
/// ASCII with word-like token lengths, so it defeats `is_garbage_text` and
|
||||
/// produces no replacement characters — it needs its own discriminator.
|
||||
#[derive(Debug, Default)]
|
||||
struct CipherGarbleStats {
|
||||
/// Case-folded ASCII letter histogram.
|
||||
letter_counts: [u32; 26],
|
||||
ascii_letters: usize,
|
||||
ascii_vowels: usize,
|
||||
/// Accented Latin letters (Latin-1 Supplement through Latin Extended-B,
|
||||
/// plus Latin Extended Additional). Count toward Latin dominance only.
|
||||
latin_ext_letters: usize,
|
||||
non_latin_letters: usize,
|
||||
/// Adjacent ASCII-letter pairs, and how many of them switch from
|
||||
/// lowercase straight to uppercase mid-word.
|
||||
letter_bigrams: usize,
|
||||
case_shift_bigrams: usize,
|
||||
}
|
||||
|
||||
impl CipherGarbleStats {
|
||||
fn add_text(&mut self, text: &str) {
|
||||
let mut prev: Option<char> = None;
|
||||
for ch in text.chars() {
|
||||
if ch.is_ascii_alphabetic() {
|
||||
let idx = (ch.to_ascii_lowercase() as u8 - b'a') as usize;
|
||||
self.letter_counts[idx] += 1;
|
||||
self.ascii_letters += 1;
|
||||
if matches!(ch.to_ascii_lowercase(), 'a' | 'e' | 'i' | 'o' | 'u') {
|
||||
self.ascii_vowels += 1;
|
||||
}
|
||||
if let Some(p) = prev {
|
||||
self.letter_bigrams += 1;
|
||||
if p.is_ascii_lowercase() && ch.is_ascii_uppercase() {
|
||||
self.case_shift_bigrams += 1;
|
||||
}
|
||||
}
|
||||
prev = Some(ch);
|
||||
} else {
|
||||
if ch.is_alphabetic() {
|
||||
if matches!(ch as u32, 0xC0..=0x24F | 0x1E00..=0x1EFF) {
|
||||
self.latin_ext_letters += 1;
|
||||
} else {
|
||||
self.non_latin_letters += 1;
|
||||
}
|
||||
}
|
||||
prev = None;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Cosine similarity between the observed letter histogram and English
|
||||
/// letter frequencies. A shifted alphabet permutes the histogram, which
|
||||
/// destroys the similarity regardless of the shift amount.
|
||||
fn english_cosine(&self) -> f64 {
|
||||
if self.ascii_letters == 0 {
|
||||
return 1.0;
|
||||
}
|
||||
let n = self.ascii_letters as f64;
|
||||
let mut dot = 0.0;
|
||||
let mut norm_obs = 0.0;
|
||||
for (count, freq) in self.letter_counts.iter().zip(ENGLISH_LETTER_FREQ) {
|
||||
let p = *count as f64 / n;
|
||||
dot += p * freq;
|
||||
norm_obs += p * p;
|
||||
}
|
||||
let norm_en = ENGLISH_LETTER_FREQ
|
||||
.iter()
|
||||
.map(|f| f * f)
|
||||
.sum::<f64>()
|
||||
.sqrt();
|
||||
dot / (norm_obs.sqrt() * norm_en)
|
||||
}
|
||||
|
||||
/// Cosine similarity between the observed histogram and English
|
||||
/// frequencies after sorting BOTH descending — i.e. comparing the *shape*
|
||||
/// of the frequency profile, ignoring which letter sits where. A
|
||||
/// substitution cipher is a bijection, so it preserves this shape exactly
|
||||
/// (att10k 0.97, arbitrary shifts 0.99) regardless of case or offset.
|
||||
/// Non-linguistic ASCII has a different profile: a small alphabet is far
|
||||
/// steeper (random DNA 0.74, hex dumps 0.81), so the shape diverges.
|
||||
fn english_shape_cosine(&self) -> f64 {
|
||||
if self.ascii_letters == 0 {
|
||||
return 1.0;
|
||||
}
|
||||
let n = self.ascii_letters as f64;
|
||||
let mut obs: [f64; 26] = std::array::from_fn(|i| self.letter_counts[i] as f64 / n);
|
||||
obs.sort_unstable_by(|a, b| b.total_cmp(a));
|
||||
let mut en = ENGLISH_LETTER_FREQ;
|
||||
en.sort_unstable_by(|a, b| b.total_cmp(a));
|
||||
|
||||
let dot: f64 = obs.iter().zip(en).map(|(o, e)| o * e).sum();
|
||||
let norm_obs = obs.iter().map(|o| o * o).sum::<f64>().sqrt();
|
||||
let norm_en = en.iter().map(|e| e * e).sum::<f64>().sqrt();
|
||||
dot / (norm_obs * norm_en)
|
||||
}
|
||||
|
||||
/// Thresholds validated against the 380-document pdf-evals snapshot
|
||||
/// corpus (0 false positives) and the garbled ParseBench `att10k` page
|
||||
/// (vowel ratio 0.245, case-shift rate 0.225, cosine 0.532). Closest
|
||||
/// legitimate document on each axis: vowel ratio 0.264 (circuit
|
||||
/// schematic), case-shift rate 0.021, cosine 0.801.
|
||||
fn looks_garbled(&self) -> bool {
|
||||
// Need a statistically meaningful, Latin-dominant sample.
|
||||
if self.ascii_letters < 200
|
||||
|| self.non_latin_letters > self.ascii_letters + self.latin_ext_letters
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
// Real Latin-script text keeps vowels above ~30% of letters even in
|
||||
// acronym- and part-number-heavy documents; shifted text starves them.
|
||||
let vowel_ratio = self.ascii_vowels as f64 / self.ascii_letters as f64;
|
||||
if vowel_ratio > 0.30 {
|
||||
return false;
|
||||
}
|
||||
|
||||
// Signal 1: lowercase→uppercase transitions inside words. A shifted
|
||||
// lowercase alphabet straddles the ASCII uppercase block ('i'→'Z',
|
||||
// 't'→'e'), so garbled words flip case constantly. Real documents
|
||||
// stay ≤ 0.02 even with camelCase identifiers.
|
||||
let case_shifts = self.letter_bigrams >= 100
|
||||
&& self.case_shift_bigrams as f64 >= self.letter_bigrams as f64 * 0.10;
|
||||
|
||||
// Signal 2: the histogram is a permutation of natural language — an
|
||||
// English-like frequency SHAPE (sorted cosine high) but with letters
|
||||
// in the wrong POSITIONS (unsorted cosine low). This is the signature
|
||||
// of a substitution cipher and is case-independent, so it catches
|
||||
// all-lowercase and all-uppercase shifts as well as case-straddling
|
||||
// ones. Genuinely non-linguistic ASCII that is merely "unlike English"
|
||||
// fails one of the two halves: DNA/hex dumps have too steep a profile
|
||||
// (shape cosine < 0.90), while protein sequences, ticker symbols and
|
||||
// base64 are not sufficiently unlike English in position (unsorted
|
||||
// cosine ≥ 0.60) — so none of them are routed to OCR.
|
||||
let permuted_language = self.english_cosine() < 0.60 && self.english_shape_cosine() >= 0.90;
|
||||
|
||||
case_shifts || permuted_language
|
||||
}
|
||||
}
|
||||
|
||||
#[derive(Debug, Default)]
|
||||
pub(crate) struct TextQualityReport {
|
||||
pub(crate) pages_needing_ocr: Vec<u32>,
|
||||
pub(crate) has_encoding_issues: bool,
|
||||
pub(crate) reasons_by_page: BTreeMap<u32, Vec<String>>,
|
||||
}
|
||||
|
||||
#[derive(Debug, Default)]
|
||||
struct PageTextQualityEvidence {
|
||||
chars: usize,
|
||||
replacement_chars: usize,
|
||||
replacement_spans: usize,
|
||||
longest_replacement_run: usize,
|
||||
cipher_garble: CipherGarbleStats,
|
||||
}
|
||||
|
||||
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
|
||||
enum TextSpanIssueKind {
|
||||
Replacement,
|
||||
Strong,
|
||||
}
|
||||
|
||||
pub(crate) fn analyze_text_quality(items: &[TextItem]) -> TextQualityReport {
|
||||
let mut reasons_by_page = BTreeMap::new();
|
||||
let mut evidence_by_page = BTreeMap::<u32, PageTextQualityEvidence>::new();
|
||||
|
||||
for item in items {
|
||||
if !matches!(item.item_type, crate::types::ItemType::Text) {
|
||||
continue;
|
||||
}
|
||||
|
||||
let evidence = evidence_by_page.entry(item.page).or_default();
|
||||
evidence.chars += item.text.chars().filter(|ch| !ch.is_whitespace()).count();
|
||||
evidence.cipher_garble.add_text(&item.text);
|
||||
|
||||
match text_span_decoding_issue_kind(&item.text) {
|
||||
Some(TextSpanIssueKind::Strong) => {
|
||||
add_ocr_reason(
|
||||
&mut reasons_by_page,
|
||||
item.page,
|
||||
OCR_REASON_SUSPECTED_GARBLED_TEXT,
|
||||
);
|
||||
}
|
||||
Some(TextSpanIssueKind::Replacement) => {
|
||||
let stats = replacement_text_stats(&item.text);
|
||||
evidence.replacement_chars += stats.0;
|
||||
evidence.replacement_spans += 1;
|
||||
evidence.longest_replacement_run = evidence.longest_replacement_run.max(stats.1);
|
||||
}
|
||||
None => {}
|
||||
}
|
||||
}
|
||||
|
||||
for (page, evidence) in evidence_by_page {
|
||||
if reasons_by_page.contains_key(&page) {
|
||||
continue;
|
||||
}
|
||||
if page_replacement_evidence_needs_ocr(&evidence) || evidence.cipher_garble.looks_garbled()
|
||||
{
|
||||
add_ocr_reason(
|
||||
&mut reasons_by_page,
|
||||
page,
|
||||
OCR_REASON_SUSPECTED_GARBLED_TEXT,
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
let pages_needing_ocr: Vec<u32> = reasons_by_page.keys().copied().collect();
|
||||
TextQualityReport {
|
||||
has_encoding_issues: !pages_needing_ocr.is_empty(),
|
||||
pages_needing_ocr,
|
||||
reasons_by_page,
|
||||
}
|
||||
}
|
||||
|
||||
pub(crate) fn region_items_have_decoding_issue(items: &[TextItem]) -> bool {
|
||||
items.iter().any(|item| {
|
||||
matches!(item.item_type, crate::types::ItemType::Text)
|
||||
&& text_span_has_decoding_issue(&item.text)
|
||||
})
|
||||
}
|
||||
|
||||
fn text_span_has_decoding_issue(text: &str) -> bool {
|
||||
text_span_decoding_issue_kind(text).is_some()
|
||||
}
|
||||
|
||||
fn text_span_decoding_issue_kind(text: &str) -> Option<TextSpanIssueKind> {
|
||||
let text = text.trim();
|
||||
if text.is_empty() {
|
||||
return None;
|
||||
}
|
||||
|
||||
if has_dollar_as_space_pattern(text)
|
||||
|| has_private_use_text_run(text)
|
||||
|| is_cid_garbage(text)
|
||||
|| has_cid_control_token(text)
|
||||
{
|
||||
return Some(TextSpanIssueKind::Strong);
|
||||
}
|
||||
|
||||
if has_replacement_text_run(text) {
|
||||
return Some(TextSpanIssueKind::Replacement);
|
||||
}
|
||||
|
||||
None
|
||||
}
|
||||
|
||||
fn replacement_text_stats(text: &str) -> (usize, usize) {
|
||||
let mut replacement = 0usize;
|
||||
let mut current_run = 0usize;
|
||||
let mut longest_run = 0usize;
|
||||
|
||||
for ch in text.chars() {
|
||||
if ch == '\u{FFFD}' {
|
||||
replacement += 1;
|
||||
current_run += 1;
|
||||
longest_run = longest_run.max(current_run);
|
||||
} else {
|
||||
current_run = 0;
|
||||
}
|
||||
}
|
||||
|
||||
(replacement, longest_run)
|
||||
}
|
||||
|
||||
fn page_replacement_evidence_needs_ocr(evidence: &PageTextQualityEvidence) -> bool {
|
||||
if evidence.replacement_chars == 0 || evidence.chars == 0 {
|
||||
return false;
|
||||
}
|
||||
|
||||
// If the entire page is only a short broken text layer, even a short
|
||||
// replacement run is enough evidence. On otherwise text-heavy pages,
|
||||
// require density so math formulas do not force full-page OCR.
|
||||
if evidence.chars <= 80 && evidence.longest_replacement_run >= 2 {
|
||||
return true;
|
||||
}
|
||||
|
||||
let replacement_density_bps = evidence.replacement_chars * 10_000 / evidence.chars;
|
||||
let enough_bad_text = evidence.replacement_chars >= 12 && replacement_density_bps >= 500;
|
||||
let repeated_bad_spans = evidence.replacement_spans >= 3 && replacement_density_bps >= 250;
|
||||
let long_bad_run = evidence.longest_replacement_run >= 8 && replacement_density_bps >= 250;
|
||||
|
||||
enough_bad_text || repeated_bad_spans || long_bad_run
|
||||
}
|
||||
|
||||
fn has_replacement_text_run(text: &str) -> bool {
|
||||
let (replacement, longest_run) = replacement_text_stats(text);
|
||||
longest_run >= 2 || replacement >= 3
|
||||
}
|
||||
|
||||
fn has_private_use_text_run(text: &str) -> bool {
|
||||
let mut total = 0usize;
|
||||
let mut private_use = 0usize;
|
||||
let mut current_run = 0usize;
|
||||
let mut longest_run = 0usize;
|
||||
|
||||
for ch in text.chars() {
|
||||
if ch.is_whitespace() {
|
||||
current_run = 0;
|
||||
continue;
|
||||
}
|
||||
total += 1;
|
||||
if is_private_use_char(ch) {
|
||||
private_use += 1;
|
||||
current_run += 1;
|
||||
longest_run = longest_run.max(current_run);
|
||||
} else {
|
||||
current_run = 0;
|
||||
}
|
||||
}
|
||||
|
||||
if private_use == 0 {
|
||||
return false;
|
||||
}
|
||||
|
||||
longest_run >= 3 || (total >= 5 && private_use >= 2 && private_use * 2 >= total)
|
||||
}
|
||||
|
||||
fn has_cid_control_token(text: &str) -> bool {
|
||||
text.split_whitespace().any(token_has_cid_control)
|
||||
}
|
||||
|
||||
fn token_has_cid_control(token: &str) -> bool {
|
||||
let mut total = 0usize;
|
||||
let mut c1_control = 0usize;
|
||||
|
||||
for ch in token.chars() {
|
||||
total += 1;
|
||||
if ('\u{0080}'..='\u{009F}').contains(&ch) {
|
||||
c1_control += 1;
|
||||
}
|
||||
}
|
||||
|
||||
total >= 5 && c1_control >= 2 && c1_control * 20 >= total
|
||||
}
|
||||
|
||||
fn is_private_use_char(ch: char) -> bool {
|
||||
matches!(
|
||||
ch as u32,
|
||||
0xE000..=0xF8FF | 0xF0000..=0xFFFFD | 0x100000..=0x10FFFD
|
||||
)
|
||||
}
|
||||
|
||||
/// Check if extracted text is predominantly garbage (non-alphanumeric).
|
||||
///
|
||||
/// Broken font encodings produce text like "----1-.-.-.___ --.-. .._ I_---."
|
||||
/// where most characters are punctuation/symbols. Real text in any language
|
||||
/// has >50% alphanumeric characters.
|
||||
pub(crate) fn is_garbage_text(markdown: &str) -> bool {
|
||||
let mut alphanum = 0usize;
|
||||
let mut non_alphanum = 0usize;
|
||||
|
||||
let chars: Vec<char> = markdown.chars().collect();
|
||||
let mut i = 0usize;
|
||||
while i < chars.len() {
|
||||
let ch = chars[i];
|
||||
let mut run_end = i + 1;
|
||||
while run_end < chars.len() && chars[run_end] == ch {
|
||||
run_end += 1;
|
||||
}
|
||||
|
||||
let is_decorative_leader = matches!(ch, '.' | '_' | '·') && run_end - i >= 3;
|
||||
if !is_decorative_leader {
|
||||
for &run_ch in &chars[i..run_end] {
|
||||
if run_ch.is_whitespace() {
|
||||
continue;
|
||||
}
|
||||
// Skip markdown syntax chars that we add (not from the PDF)
|
||||
if matches!(run_ch, '#' | '*' | '|' | '-' | '\n') {
|
||||
continue;
|
||||
}
|
||||
if run_ch.is_alphanumeric() {
|
||||
alphanum += 1;
|
||||
} else {
|
||||
non_alphanum += 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
i = run_end;
|
||||
}
|
||||
|
||||
let total = alphanum + non_alphanum;
|
||||
total >= 50 && alphanum * 2 < total
|
||||
}
|
||||
|
||||
/// Detect garbage from failed CID-to-Unicode mapping on Identity-H fonts.
|
||||
///
|
||||
/// When CID values don't correspond to Unicode codepoints, the raw bytes often
|
||||
/// produce characters in the C1 control range (U+0080–U+009F) or Private Use
|
||||
/// Area, mixed with random Latin Extended characters. Valid text in any
|
||||
/// language almost never contains C1 controls. We also fall back to the
|
||||
/// general `is_garbage_text` check for non-alphanumeric-heavy patterns.
|
||||
pub(crate) fn is_cid_garbage(text: &str) -> bool {
|
||||
if is_garbage_text(text) {
|
||||
return true;
|
||||
}
|
||||
let mut total = 0usize;
|
||||
let mut c1_control = 0usize;
|
||||
let mut high_latin = 0usize;
|
||||
for ch in text.chars() {
|
||||
if ch.is_whitespace() {
|
||||
continue;
|
||||
}
|
||||
total += 1;
|
||||
// C1 control characters (U+0080–U+009F) — almost never in real text
|
||||
if ch == '·' {
|
||||
continue;
|
||||
}
|
||||
if ('\u{0080}'..='\u{009F}').contains(&ch) {
|
||||
c1_control += 1;
|
||||
}
|
||||
// High Latin-1 (U+00A0–U+00FF) — legitimate in Western European text
|
||||
// but when combined with ASCII in CID passthrough, indicates mojibake
|
||||
// from CID values being misinterpreted as Latin-1 characters.
|
||||
if ('\u{00A0}'..='\u{00FF}').contains(&ch) {
|
||||
high_latin += 1;
|
||||
}
|
||||
}
|
||||
if total < 5 {
|
||||
return false;
|
||||
}
|
||||
// If ≥5% of non-whitespace chars are C1 controls, it's garbage
|
||||
if c1_control >= 2 && c1_control * 20 >= total {
|
||||
return true;
|
||||
}
|
||||
// If ≥40% of non-whitespace chars are high Latin-1 AND the text has few
|
||||
// ASCII letters, it's likely CID-as-Latin-1 mojibake (Japanese/CJK PDFs
|
||||
// where CID values 0x80-0xFF become accented Latin characters). Keep a
|
||||
// minimum length so short math tokens like "2×()×" do not route a clean
|
||||
// page to OCR.
|
||||
let ascii_letters = text.chars().filter(|c| c.is_ascii_alphabetic()).count();
|
||||
total >= 20 && high_latin * 5 >= total * 2 && ascii_letters * 3 < total
|
||||
}
|
||||
@@ -48,7 +48,7 @@ tips of directly from customers received other employees paid tips rec’d. entr
|
||||
|
||||
**Page 3**
|
||||
|
||||
27 28 29 30 31 **Subtotals** **from pages** **1, 2, and 3** **Totals**
|
||||
27 28 29 30 31 **Subtotals from pages** **1, 2, and 3** **Totals**
|
||||
|
||||
**1.** Report total cash tips (col. **a**) on Form 4070, line **1.**
|
||||
**2.** Report total credit card tips (col. **b**) on Form 4070, line **2.**
|
||||
@@ -76,7 +76,6 @@ forms simpler, we would be happy to hear from you. You can write to the Tax Form
|
||||
|
||||
**Unreported Tips.—**If you received tips of $20 or more for any month while working for one employer but did not report them to your employer, you must figure and pay social security and Medicare taxes on the unreported tips when you file your tax return. If you have unreported tips, you **must** use Form 1040 and **Form 4137,** Social Security and Medicare Tax on Unreported Tip Income, to report them. You may **not** use Form 1040A or 1040EZ. Employees subject to the Railroad Retirement Tax Act **cannot** use Form 4137 to pay railroad retirement tax on unreported tips. To get railroad retirement credit, you must report tips to your employer. If you do not report tips to your employer as required, you may be charged a penalty of 50% of the social security and Medicare taxes (or railroad retirement tax) due on the unreported tips unless there was reasonable cause for not reporting them. **Additional Information.—**Get **Pub. 531,** Reporting Tip Income, and Form 4137 for more information on tips. If you are an employee of certain large food or beverage establishments, see Pub. 531 for tip allocation rules. **Recordkeeping.—**If you do not keep a daily record of tips, you must keep other reliable proof of the tip income you received. This proof includes copies of restaurant bills and credit card charges that show amounts customers added as tips. Keep your tip income records for as long as the information on them may be needed in the administration of any Internal Revenue law.
|
||||
|
||||
**Instructions** *(continued)*
|
||||
### Instructions (continued)
|
||||
|
||||
Use this space to total your tips for the year
|
||||
|
||||
|
||||
Reference in New Issue
Block a user