Migrate to 4.0.0

Here are a changelog / migration guide from 3.x.x to 4.0.0 version of twitter-archive-reader.

Changed

TwitterArchive

constructor

The TwitterArchive constructor now takes two parameters: file and options.

Options available are described by the TwitterArchiveLoadOptions interface:

interface TwitterArchiveLoadOptions {
  load_images_in_zip?: boolean,
  build_ad_archive?: boolean,
}

• load_images_in_zip have the same effect as before
• New parameter build_ad_archive let you choose if you want to parse advertiser data. As those kind of data might be heavy, this is not enabled by default.

Option keep_loaded is removed. You can release the ZIP at any time with new method .releaseZip() and check if its loaded with .is_zip_loaded.

Option build_extended_gdpr is removed. "Extended" data is now systematically built.

Properties

Most of the data that was before in .extended_gdpr property is moved to accessors in TwitterArchive instances.

• .extended_gdpr.moments => .moments: Access to Twitter moments
• .extended_gdpr.lists => .lists: Access to created and subscribed lists
• .extended_gdpr.followers => .followers: Set of followers IDs
• .extended_gdpr.followings => .followings: Set of user IDs following the archive owner
• .extended_gdpr.mutes => .mutes: Set of muted user IDs
• .extended_gdpr.blocks => .blocks: Set of blocked user IDs
• .extended_gdpr.favorites => .favorites: New FavoriteArchive instance. See new features for more details

If these accessors are invoked with a classic archive, containers will be empty (empty array, or empty set).

Property .extended_gdpr is removed.

Remaining old properties are now stored in the new UserData instance (see new features).

Other changes on properties:

• .requires_dm_image_load => !.is_dm_images_available: Property is inverted and renamed for more consistency.
• .owner => .user.id
• .owner_screen_name => .user.screen_name: All user-related data is now stored in .user property

New

UserData

Store all of archive owner related user data.

Accessible through .user property of a TwitterArchive instance.

Contains screen name, name, user ID, screen name history, protected history... See related documentation to know more.

AdArchive

Store data about Twitter seen/engaged ads.

Accessible through .ads property of a TwitterArchive instance.

Initialization

This container is initialized only if build_ad_archive constructor option is set to true. If you haven't init this at construct time, you can still load it (if the ZIP is still loaded) with .loadArchivePart({ current_ad_archive: true }):

if (archive.is_zip_loaded) {
  await archive.loadArchivePart({ current_ad_archive: true });
}

FavoriteArchive

Store data about favorited tweets.

Accessible through .favorites property of a TwitterArchive instance.

Since 2019, favorites inside archives might be associated with favorited tweet text. Data type of a favorite now follows this interface:

interface PartialFavorite {
  tweetId: string;
  /** Text of the favorited tweet. Defined only if archive creation > around June 2019. */
  fullText?: string;
  /** URL to the tweet. Defined only if archive creation > around June 2019. */
  expandedUrl?: string;
}

This container allows quick access to those kind of data:

• .all: PartialFavorite[]: Get all favorited tweets informations
• .has(tweet_id: string): boolean: Check if a tweet is favorited
• .get(tweet_id: string): PartialFavorite: If favorite exists, get information
• .registred: Set<string>: Set of favorited tweet IDs (as present in old .extended_gdpr)
• .length: number: Favorite count
• .has_extended_favorites: boolean: True if this FavoriteArchive has favorites with fullText and expandedUrl properties

This object is iterable.

DMArchive

New method .conversationOf(dm_id: string) allows you to get the related conversation of a direct message.