Tom Butcher
01b3cc1aa1
All checks were successful
farmcontrol/farmcontrol-server/pipeline/head This commit looks good
- Modify the Linux package build command in package.json to include the --no-bytecode flag for improved compatibility with ESM packages. - Refactor pdfUtils.js to handle dynamic imports more robustly, ensuring compatibility with pkg snapshots and improving error handling for ESM modules. - Simplify image processing logic by removing unnecessary default property access for the sharp module.
93 lines
2.4 KiB
JavaScript
93 lines
2.4 KiB
JavaScript
/**
|
|
* PDF conversion utilities loaded on demand so headless startup
|
|
* does not require ESM-only pdf-to-img at module load time.
|
|
*/
|
|
import { createRequire } from 'node:module'
|
|
import { pathToFileURL } from 'node:url'
|
|
|
|
const require = createRequire(import.meta.url)
|
|
|
|
// Dynamic import() from pkg snapshot bytecode throws "Invalid host defined options"
|
|
// (nodejs/node#43663). Use require() for CJS packages in pkg; ESM packages need
|
|
// --no-bytecode on the pkg build so file-URL import() works.
|
|
async function importPackage(name) {
|
|
const packagePath = require.resolve(name)
|
|
|
|
if (typeof process.pkg !== 'undefined') {
|
|
try {
|
|
return require(packagePath)
|
|
} catch (error) {
|
|
if (error?.code !== 'ERR_REQUIRE_ESM') {
|
|
throw error
|
|
}
|
|
}
|
|
}
|
|
|
|
return import(pathToFileURL(packagePath).href)
|
|
}
|
|
|
|
function resolveDefault(module) {
|
|
return module?.default ?? module
|
|
}
|
|
|
|
export async function convertPDFToImage(pdfInput, options = {}) {
|
|
const [pdfToImgModule, sharpModule] = await Promise.all([
|
|
importPackage('pdf-to-img'),
|
|
importPackage('sharp')
|
|
])
|
|
const { pdf } = pdfToImgModule
|
|
const sharp = resolveDefault(sharpModule)
|
|
|
|
try {
|
|
const { width, height, page_numbers, scale = 2, ...pdfOnlyOptions } =
|
|
options
|
|
|
|
const pdfOptions = {
|
|
scale,
|
|
...pdfOnlyOptions
|
|
}
|
|
|
|
const document = await pdf(pdfInput, pdfOptions)
|
|
const outputImages = []
|
|
|
|
if (
|
|
page_numbers &&
|
|
Array.isArray(page_numbers) &&
|
|
page_numbers.length > 0
|
|
) {
|
|
for (const pageNum of page_numbers) {
|
|
let image = await document.getPage(pageNum)
|
|
|
|
if (width || height) {
|
|
const resizeOptions = {}
|
|
if (width) resizeOptions.width = width
|
|
if (height) resizeOptions.height = height
|
|
image = await sharp(image).resize(resizeOptions).toBuffer()
|
|
}
|
|
|
|
outputImages.push(image)
|
|
}
|
|
} else {
|
|
for await (const image of document) {
|
|
let processedImage = image
|
|
|
|
if (width || height) {
|
|
const resizeOptions = {}
|
|
if (width) resizeOptions.width = width
|
|
if (height) resizeOptions.height = height
|
|
processedImage = await sharp(image)
|
|
.resize(resizeOptions)
|
|
.toBuffer()
|
|
}
|
|
|
|
outputImages.push(processedImage)
|
|
}
|
|
}
|
|
|
|
return outputImages
|
|
} catch (error) {
|
|
console.error('Error converting PDF to image:', error)
|
|
throw error
|
|
}
|
|
}
|