pdf2md

You can integrate pdf2md into your Node.js application to convert PDF buffers into Markdown text. The library accepts a PDF buffer and returns a Promise that resolves with the converted Markdown string.

Note: The example usage requires a callbacks argument in the function signature, though the implementation details of these callbacks are not specified in the README.

const path = require('path');
const fs = require('fs')
const pdf2md = require('@opendocsg/pdf2md')

const pdfBuffer = fs.readFileSync(filePath)
pdf2md(pdfBuffer, callbacks)
  .then(text => {
    let outputFile = allOutputPaths[i] + '.md'
    console.log(`Writing to ${outputFile}...`)
    fs.writeFileSync(path.resolve(outputFile), text)
    console.log('Done.')
  })
  .catch(err => {
    console.error(err)
  })

When using the pdf2md library, a transformation or PDF parse operation returns a ParseResult object. This object contains the processed content and metadata used for subsequent transformations or debugging.

Key properties of a ParseResult instance:

• pages: An array of Page objects representing the parsed content of the PDF.
• globals: An object containing properties accessible to all subsequent transformations (primarily used in debug mode).
• messages: An array of messages intended for display during transformations (primarily used in debug mode).

The WordType class defines how different types of markdown elements (like links and footnotes) are formatted into text. It is an enumeration used by the library to determine the string representation of specific word elements during the PDF to Markdown conversion process.

Available types include:

• LINK: Formats a string as a standard markdown link: [text](url).
• FOOTNOTE_LINK: Formats a string as a footnote reference: ^text.
• FOOTNOTE: Formats a string as a footnote content: (^text).

// Note: This is a conceptual representation of the WordType enum values
// used internally by the library to format markdown output.

WordType.LINK // returns '[text](text)'
WordType.FOOTNOTE_LINK // returns '^text'
WordType.FOOTNOTE // returns '(^text)'

When performing recursive conversions on a large number of files, you may encounter the error Allocation failed - JavaScript heap out of memory. To resolve this, bypass npx and run the CLI script directly using node with the --max-old-space-size flag to increase the available memory.

$ node lib/pdf2md-cli.js --max-old-space-size=4096 --inputFolderPath=[your input folder path] --outputFolderPath=[your output folder path] --recursive

When using the CLI tool, you can provide the following options:

• --inputFolderPath: The path to the existing folder containing the PDF files to be converted.
• --outputFolderPath: The path to the existing folder where Markdown files will be saved.
• --recursive: If specified, the tool will convert all PDFs found in subfolders within the input folder. Omit this flag to only process the top-level directory.

The pdf2md CLI tool allows you to batch convert PDF files from an input directory to an output directory. You can run it via npx or directly via node for advanced memory management.

$ npx @opendocsg/pdf2md --inputFolderPath=[your input folder path] --outputFolderPath=[your output folder path] --recursive

The isHeadline(type) function returns true if the provided BlockType is a headline (i.e., its name starts with 'H' and is two characters long, such as H1, H2, etc.).

const { isHeadline } = require('./lib/models/markdown/BlockType');

const result = isHeadline(BlockType.H1); // true
const result2 = isHeadline(BlockType.PARAGRAPH); // false

The blockToText function converts a LineItemBlock into its Markdown string representation based on its assigned BlockType. If the block has no type assigned, it defaults to rendering the items as plain text (without inline formatting).

const { blockToText } = require('./lib/models/markdown/BlockType');

// Assuming 'block' is a valid LineItemBlock object
const markdownString = blockToText(block);

The headlineByLevel(level) function returns the corresponding BlockType for a given heading level (1 through 6).

• If level is between 1 and 6, it returns the matching H1-H6 member.
• If level is outside this range, it logs a warning and defaults to BlockType.H6.

const { headlineByLevel } = require('./lib/models/markdown/BlockType');

const h1 = headlineByLevel(1); // Returns BlockType.H1
const h9 = headlineByLevel(9); // Returns BlockType.H6 and logs a warning