helmetroo / fetch-twitter-bookmarks Goto Github PK

• Look at zeit/pkg

Check should actually fetch the tweet first if id given

Or is there a way to determine how much to scroll by when Twitter automatically removes out of sight tweets from the DOM?

Another approach might be to scroll down a ways past the last element caught when selecting article elements.

Should be a mock version of cli-progress

We can probably use immutable-js and store tweets in an OrderedSet

• https://mobile.twitter.com/thecreativeindp/status/1138507461487411205

Code quality only.
https://github.com/typescript-eslint/typescript-eslint

Create a ScrollState interface to represent scroll progress to help break down extractTweets method.

CommandLineInterface.run() should be broken up.

Need to also firmly decide which behavior most makes sense as the default.

Argument should decide if we

• Fetch from the very beginning
• Fetch from a specific cursor
• Fetch from the last cursor (current working default)
• Allow for tweet ID or timestamp as a starting point (if a means through API can be figured out)

Should be a CLI option

Decide if we go with monorepo or separate repos for modules.

Or by default, use STDOUT if filename not provided.

i.e. name=tiny. Full size image is what we want.

Quits in Powershell but gives an error

Should probably have a separate class that maintains the browser and bookmarks page for the extractor to extract tweets from. Also want to make Puppeteer classes like Browser and Page mockable for testing

Ideas:

• Download a working copy of the bookmarks page that can work without logging in, replace network requests with test data

Tests:

• Ensure all tweets are downloaded if maxLimit not set
• Ensure only maxLimit tweets are fetched if maxLimit set
• CLI should error if invalid parameters are given
• ...

Instead of a file, tweets could be saved to a working database, instead of saving the current ones in the in-memory OrderedSet.

XML/CSV/others...?

Formatting only. See https://prettier.io/docs/en/comparison.html