pdf2md returns an empty markdown for some files in Japanese

[dmtran-g]

Description

pdf2md returns an empty markdown for the following file: tax_template.pdf

However, https://pdf2md.morethan.io/ (https://github.com/jzillmann/pdf-to-markdown) manages to read the PDF file and convert it to markdown.

To Reproduce

import pdf2md from '@opendocsg/pdf2md';
import { readFileSync, writeFileSync } from 'fs';

const file = readFileSync('tax_template.pdf');
const markdownText = await pdf2md(file);
console.log('markdownText', markdownText);
writeFileSync('tax_template.md', markdownText);

Expected behavior

pdf2md returns a non empty markdown, like https://pdf2md.morethan.io/

Misc

OS: macOS 15.3.2
node.js v20.18.1
pdf2md 0.2.1

[dmtran-g]

FYI, I see that jzillmann@f34d8d5#diff-521901391348e18fa231e4ac8108d5bbafad0702f8bd13e3b6b5c59528a0bc13 (a commit that is present in pdf-to-markdown but not in pdf2md) introduces character maps, which seem to be used to handle various languages, including Japanese.