combine a bunch of files into a single markdown file to paste into an AI or whatever

combine_files_to_markdown.ts

#!/usr/bin/env -S bun
import { readdir } from "fs/promises"

// Function to read and format file content
async function readFileContent(filePath: string): Promise<string> {
  const file = Bun.file(filePath)
  const content = await file.text()
  const extension = filePath.split(".").pop()
  if (extension === "md" || extension === "mdx") return content.replace(/^(#+) /gm, "#$1 ") // increase heading levels by 1
  const codeBlockSyntax = getSyntaxForExtension(extension)
  return content
    ? `\`\`\`${codeBlockSyntax} filename="${filePath}"\n${content.trim()}\n\`\`\``
    : "(Contents available upon request)"
}

// Function to get code block syntax based on file extension
function getSyntaxForExtension(extension?: string): string {
  switch (extension) {
    case "ts":
    case "tsx":
      return "typescript"
    case "js":
    case "jsx":
      return "javascript"
    case "json":
      return "json"
    default:
      return extension || ""
  }
}

// Function to determine if a file is binary based on its MIME type
async function isBinaryFile(filePath: string): Promise<boolean> {
  const file = Bun.file(filePath)
  const mimeType = file.type
  return !(
    (mimeType?.startsWith("text/") || mimeType === "application/json" || mimeType === "image/svg+xml") // Example of non-binary type you might still want to include
  )
}

// Recursively walk the target folder and collect file paths
async function walkFolder(folderPath: string): Promise<string[]> {
  const files: string[] = []

  const entries = await readdir(folderPath, { withFileTypes: true })
  for (const entry of entries) {
    const entryPath = `${folderPath}/${entry.name}`
    if (entry.isDirectory()) {
      files.push(...(await walkFolder(entryPath)))
    } else {
      // Check if the file is binary before adding
      const binary = await isBinaryFile(entryPath)
      if (binary) continue
      if (entryPath.includes("/.turbo/")) continue

      files.push(entryPath)
    }
  }
  return files
}

const filenameToSortKey = (filename: string) =>
  filename.toLowerCase().includes("/readme")
    ? `0${filename}` //
    : filename.toLowerCase().endsWith(".md")
      ? `1${filename}`
      : filename

const fileSorter = (a: string, b: string) => filenameToSortKey(b).localeCompare(filenameToSortKey(a))

// Main function
async function main() {
  // Get target folder and output file from command line arguments
  // eslint-disable-next-line @typescript-eslint/no-unused-vars
  const [$0, $1, ...sourceFiles] = process.argv

  if (sourceFiles.length === 0) {
    const commandName = __filename.replace(__dirname + "/", "")
    console.warn(`No source files provided`)
    console.warn(`\t${__filename}`)
    console.warn(`\t${commandName} $(rg -l 'search term') > ../output.md`)
    console.warn(
      `\t${commandName} $(rg -l 'prisma.*relay|relay.*prisma' | grep -vE 'json|yaml|svg|yml|snap') > ../prisma+relay.md`,
    )
    process.exit(1)
  }

  // Collect all file paths
  const filePaths = await Promise.all(sourceFiles.map(file => walkFolder(file).catch(() => file))).then(files =>
    files.flat().sort(fileSorter).reverse(),
  )

  const header = `
# Combined Files

This markdown file was generated by combining an entire folder structure into a single document.

## Files Included:

${filePaths.map(pathname => `* ${pathname.replace(process.env.HOME!, "~")}`).join("\n")}

---
`

  // Read and format each file content
  const fileContents = await Promise.all(
    filePaths.map(async filePath => {
      const fileName = filePath.replace(`${sourceFiles}/`, "")
      const content = await readFileContent(filePath)
      return `\n<details><summary>${fileName}</summary>\n\n${content.trim()}\n</details>\n`
    }),
  )

  // Combine all content into a single markdown string
  const markdownContent = header + fileContents.join("\n\n") + "\n"
  console.log(markdownContent)

  // Write the markdown content to the output file
  // await Bun.write(outputFile, markdownContent)

  // console.log(`Combined files into ${outputFile}`)
}

// Run the main function
// @ts-ignore - ignore top-level await error
await main()

usage:

combine_files_to_markdown.ts $(rg -l '@RelayResolver') | pbcopy

uses ripgrep to find files that mention @RelayResolver and then copy the combined markdown to your clipboard on macOS