# discord-json-2021-01 ## Manual edits * images -> placeholders * `accounts/avatar.png` * manually scrub folder names * `account/applications/0000000000000` ## Notes about files * `activity/` * All the .json are NDJSON so some json tools don't like them * _Massive_ files. They hang scrub.ts for a long long time (had to run these piecemeal) * These files also have an _incredible_ amount of shapes and variance. * Instead of outputing all the shapes I made a sort of "super-object" to capture the shape with `jq -n '[inputs] | add' events-2021-00000-of-00001.json.tmp > unique_shape.json` and then scrubbing `unique_shape.json` * `messages/` * I hand did these to keep all the ids the same * There are multiple types of chats. DMs, guild channels, etc * I hand did the csvs as I have no scrubber for that * These are only **THE EXPORTING USERS MESSAGES**, no other user, just fyi * Ids in `messages.csv` are just the id of the message, not of any user * There is the potential to derive missing info from a channel via `@` tags sent or possibly via attachments. Maybe... * `11111111111111111` * This one has a shorter id (it's an older one) * Has `type: 0` but there's no guild information in `channel.json` * The user name was `null` in `index.json` * It's a really odd one * `222222222222222222` * This was a dm channel (said `direct message with xxx#7777` in index.json) * Has `type: 1` and there are two recipients (just the ids) in `channel.json` * Unfortunately that's all the info in the export * `333333333333333333` * This was a normal guild channel * `type: 0` and there's guild information in `channel.json` * I kept a good set of messages around from this one to show how attachements and other stuff works * The last message seemed to be a link not as an attachment. Links just seem to be normal text * `programs/` * was empty... * `servers/`` * Info about _some_ of the guilds we have ids for * guild.json didn't really contain anything except the name * I kept around the only guild I noticed an audit-log.json with info in it