A deno script that downloads all files in the Slack export archive

download.ts

/**
 * A deno script that downloads all files in the Slack export archive
 * 2023-03-03
 *
 * Copyright (c) 2023 0b5vr
 * SPDX-License-Identifier: MIT
 *
 * Usage:
 *   - Install deno - https://deno.land/
 *   - `deno run --allow-read --allow-write --allow-net download.ts <export>`
 *
 * Ref: https://gist.github.com/marnitto/f9970c2bbf2798782c06bc2b9d58c0cc
 */

import * as fs from "https://deno.land/[email protected]/fs/mod.ts";
import * as path from "https://deno.land/[email protected]/path/mod.ts";

async function downloadFile(
  fileEntity: Record<string, unknown>,
  downloadDir: string,
) {
  const name = fileEntity["name"] as string | undefined;
  if (name == null) {
    console.log("🤔 Skipping a file - The file entry does not have a name");
    return;
  }

  const id = fileEntity["id"] as string | undefined;
  if (id == null) {
    console.log(`🤔 Skipping ${name} - The file entry does not have an id`);
    return;
  }

  const url = fileEntity["url_private_download"] as string | undefined;
  if (url == null) {
    console.log(`🤔 Skipping ${name} - The file entry does not have an URL`);
    return;
  }

  console.log(`📥 Downloading ${name}`);

  fs.ensureDir(downloadDir);

  const filepath = path.join(downloadDir, `${id}_${name}`);

  const res = await fetch(url);
  const dst = await Deno.open(filepath, { write: true, create: true });
  await res.body?.pipeTo(dst.writable);
}

const ignoreJsonList = new Set([
  "channels.json",
  "integration_logs.json",
  "users.json",
]);

async function main(): Promise<void> {
  const exportDir = Deno.args[0];

  if (exportDir == null) {
    throw new Error("Please specify an exported directory!");
  }

  const exportDirStat = await Deno.stat(exportDir);
  if (!exportDirStat.isDirectory) {
    throw new Error("The given path is not a directory");
  }

  const diaryEntryIter = fs.expandGlob(path.join(exportDir, "**/*.json"));
  for await (const diaryEntry of diaryEntryIter) {
    if (ignoreJsonList.has(diaryEntry.name)) {
      console.log(`⏩ Skipping ${diaryEntry.name}`);
      continue;
    }

    const diary = JSON.parse(await Deno.readTextFile(diaryEntry.path));

    if (!Array.isArray(diary)) {
      console.log(`🤔 Skipping ${diaryEntry.name} - The JSON is not an array`);
      continue;
    }

    const diaryPathSplit = diaryEntry.path.split("\\");
    const date = diaryPathSplit[diaryPathSplit.length - 1].split(".")[0];
    const channel = diaryPathSplit[diaryPathSplit.length - 2];

    const downloadDir = path.join(exportDir, channel, date);

    console.log(`👀 Processing ${downloadDir}`);

    for (const messages of diary as Record<string, unknown>[]) {
      const file = messages["file"] as Record<string, unknown> | undefined;
      if (file != null) {
        await downloadFile(file, downloadDir);
      }

      const files = messages["files"] as Record<string, unknown>[] | undefined;
      if (files != null) {
        for (const file of files) {
          await downloadFile(file, downloadDir);
        }
      }
    }
  }

  console.log("✅ Completed");
}

main();