echos-ocr/src/lib/pdf.ts
Randall Stillwell 7e423a3e50 Build Echo OCR app: full-stack response card scanner
- Next.js 15 + Tailwind v4 + shadcn/ui + TanStack Table
- Prisma + PostgreSQL schema for ResponseCard, ProcessingJob, AppSettings
- Ollama vision integration with structured extraction prompts
- PDF-to-image pipeline with pdf2pic + sharp
- MinIO S3 storage for uploaded files and extracted images
- Chokidar-based folder monitoring for automatic processing
- REST API (9 endpoints) ready for Monday.com integration
- Dashboard with filterable data table, chip filters, bulk actions, CSV export
- Card detail page with editable fields and side-by-side scanned images
- Upload page with drag-and-drop and real-time processing queue
- Settings page for Ollama, folder watch, and Monday.com config
- Dockerfile (multi-stage) + docker-compose for local dev
- Configured for Coolify deployment at echo.stillwell.cloud

Made-with: Cursor
2026-03-10 07:17:45 -05:00

60 lines
1.5 KiB
TypeScript

import { fromBuffer } from "pdf2pic";
import sharp from "sharp";
export interface PageImage {
page: number;
buffer: Buffer;
width: number;
height: number;
}
export async function pdfToImages(pdfBuffer: Buffer): Promise<PageImage[]> {
const converter = fromBuffer(pdfBuffer, {
density: 200,
format: "jpeg",
width: 1600,
height: 2200,
quality: 90,
});
const pageCount = await getPdfPageCount(pdfBuffer);
const images: PageImage[] = [];
for (let i = 1; i <= pageCount; i++) {
const result = await converter(i, { responseType: "buffer" });
if (result.buffer) {
const metadata = await sharp(result.buffer).metadata();
images.push({
page: i,
buffer: result.buffer as Buffer,
width: metadata.width || 1600,
height: metadata.height || 2200,
});
}
}
return images;
}
async function getPdfPageCount(pdfBuffer: Buffer): Promise<number> {
const text = pdfBuffer.toString("latin1");
const matches = text.match(/\/Type\s*\/Page(?!s)/g);
return matches ? matches.length : 2;
}
export async function imageToBase64(buffer: Buffer): Promise<string> {
const processed = await sharp(buffer)
.resize(1200, undefined, { withoutEnlargement: true })
.jpeg({ quality: 85 })
.toBuffer();
return processed.toString("base64");
}
export async function processUploadedImage(buffer: Buffer): Promise<Buffer> {
return sharp(buffer)
.resize(1600, undefined, { withoutEnlargement: true })
.normalize()
.sharpen()
.jpeg({ quality: 90 })
.toBuffer();
}