Quick-n-dirty posts export as CSV from BlueSky repo.car files. Just exports text and timestamp

unpack.ts

import { CarIndexedReader } from "@ipld/car";
import { decode } from "cbor2";
import { stringify as csvStringify } from 'csv-stringify/sync';

interface Subject {
  cid: string;
  uri: string;
}

interface EmbedImages {
  $type: 'app.bsky.embed.images';
  images: object[];
}

interface EmbedRecordWithMedia {
  $type: 'app.bsky.embed.recordWithMedia';
}

type Embed = EmbedImages | EmbedRecordWithMedia;

interface Like {
  $type:'app.bsky.feed.like';
  subject: Subject;
  createdAt: string;
}

interface Post {
  $type: 'app.bsky.feed.post';
  text: string;
  reply?: {
    root: Subject
    parent: Subject
  }
  langs?: string[];
  embed?: Embed;
  facets?: object[];
  createdAt: string;
}

interface Follow {
  $type: 'app.bsky.feed.follow';
  subject: string;
  createdAt: string;
}

interface Unknown {}

type Block = Like | Post | Follow | Unknown;

const [_node, scriptFile, path] = process.argv;
if (!path) {
  console.error(`Usage: ${scriptFile} [path-to-repo.car]`);
  process.exit(1);
}

const repo = await CarIndexedReader.fromFile(path);
const blocks = await repo.blocks();

process.stdout.write(csvStringify([['type', 'text', 'createdAt']]))

while (true) {
  const block = await blocks.next();
  if (block.done) {
    break;
  }

  const payload = decode(block.value.bytes) as Block;

  if ('$type' in payload && payload.$type === 'app.bsky.feed.post') {
    process.stdout.write(csvStringify([[payload.$type, payload.text, payload.createdAt]]))
  }
}