fix(Util): flatten ignoring certain fields

[newracket]

Please describe the changes this PR makes and why it should be merged:
Fixes #7712

Improves clarity of toJSON functions throughout. Currently, some fields are ignored in toJSON if the value tied to that field is an object (as pointed out by issue #7712). This change fixes issue #7712 by making it so fields tied to an object don't get ignored.

Status and versioning classification:

• Code changes have been tested against the Discord API, or there are no code changes
• I know how to update typings and have done so, or typings don't need updating

[newracket]

This is what Message.toJSON() looks like before I made the change.

{
  "channelId": "723654712410374194",
  "guildId": "605969074178228224",
  "id": "963209574242979900",
  "createdTimestamp": 1649717458068,
  "type": 0,
  "system": false,
  "content": "",
  "authorId": "605975599609020437",
  "pinned": false,
  "tts": false,
  "nonce": null,
  "embeds": [
    {}
  ],
  "components": [],
  "attachments": [],
  "stickers": [],
  "editedTimestamp": null,
  "webhookId": null,
  "groupActivityApplicationId": null,
  "applicationId": null,
  "activity": null,
  "flags": 0,
  "reference": null,
  "interaction": null,
  "cleanContent": ""
}

This is what it looks like with the change. Both mentions and embeds were omitted previously, but with the change, they're both now included

{
  "channelId": "723654712410374194",
  "guildId": "605969074178228224",
  "id": "963209574242979900",
  "createdTimestamp": 1649717458068,
  "type": 0,
  "system": false,
  "content": "",
  "authorId": "605975599609020437",
  "pinned": false,
  "tts": false,
  "nonce": null,
  "embeds": [
    {
      "data": {
        "type": "rich",
        "author": {
          "name": "hey"
        }
      }
    }
  ],
  "components": [],
  "attachments": [],
  "stickers": [],
  "editedTimestamp": null,
  "mentions": {
    "everyone": false,
    "users": [],
    "roles": [],
    "crosspostedChannels": [],
    "repliedUser": null
  },
  "webhookId": null,
  "groupActivityApplicationId": null,
  "applicationId": null,
  "activity": null,
  "flags": 0,
  "reference": null,
  "interaction": null,
  "cleanContent": ""
}

[kyranet]

One of my concerns here is... shouldn't we try to use .toJSON where possible?

With the switch to raw data, we may have nested data, which is what happens here in your sample:

  "embeds": [
    {
+     "data": {
        "type": "rich",
        "author": {
          "name": "hey"
        }
+     }
    }
  ],

If .toJSON were to be preferred and bypass the JSON serialization from the utility, data would get dropped, and the data would be "flat".

[newracket]

@kyranet I made a second commit making it try to use .toJSON if it exists (both when the element's an object and when it's an array). The same Message.toJSON() from my previous commit now looks like this.

{
  "channelId": "723654712410374194",
  "guildId": "605969074178228224",
  "id": "963209574242979900",
  "createdTimestamp": 1649717458068,
  "type": 0,
  "system": false,
  "content": "",
  "authorId": "605975599609020437",
  "pinned": false,
  "tts": false,
  "nonce": null,
  "embeds": [
    {
      "type": "rich",
      "author": {
        "name": "hey"
      }
    }
  ],
  "components": [],
  "attachments": [],
  "stickers": [],
  "editedTimestamp": null,
  "mentions": {
    "everyone": false,
    "users": [],
    "roles": [],
    "crosspostedChannels": [],
    "repliedUser": null,
    "members": [],
    "channels": []
  },
  "webhookId": null,
  "groupActivityApplicationId": null,
  "applicationId": null,
  "activity": null,
  "flags": 0,
  "reference": null,
  "interaction": null,
  "cleanContent": ""
}