julianfssen / tweetkit Goto Github PK

I just noticed you are still depending on faraday 1.x.

That makes it impossible to use in an app that has other dependencies on faraday 2.x. Faraday 2.0.0 was released about 18 months ago, Jan 2022.

Would it be possible to expand to allowing faraday 1.x-2.x, or just 2.x?

Are you interested in a PR? It doesn't look like there's CI running here (travis doesn't really provide free CI to open source anymore); it looks like there are some specs, I haven't looked to see how good the test coverage is, does it cover enough that if tests still pass after an upgrade to faraday 2 that's pretty reliable?

Thanks for the great work!

The latest version is July 16, 2021

Thank you!

Hi! Thanks for using this gem.

tweetkit is a Twitter v2 API wrapper inspired by the octokit gem and the current twitter gem. That said, tweetkit is currently very rudimentary and is nowhere near polished enough for reliable production usage.

With that in mind, I'm working on new features and improvements to make tweetkit suitable for serious usage, as shown below:

Upcoming Features

Flexible search DSL

The search DSL is already available in tweetkit. It is, however, quite basic and lacks documentation. In short, the search DSL allows you to easily create search statements based on Twitter's search operators.

Here's an example search code:

  client.search 'lionel messi' do
    contains 'psg', 'barcelona'
    without 'manchester united', 'manchester city'
    is_a :retweet
    is_not :verified
    has :media
  end

This searches for retweets containing the phrase lionel messi that also contains the words psg and barcelona. The matched retweets should not have any occurrences of manchester united and manchester city. The retweets should not be from verified accounts and it should contain a media object, such as a photo, GIF, or video, as determined by Twitter.

There is a lot to improve for this search DSL and I'm hoping to make it as easy-to-use and flexible as similar DSLs like RSpec, builder, and so on.

All v2 endpoints

The gem currently only supports fetching and searching for tweets. The plan is to cover all v2 endpoints, both existing and upcoming. I do not plan to add v1 endpoints at the moment as it is covered completely by the current twitter gem. It will be also be replaced by the v2 endpoints soon so they'll not be worked on unless there's demand from the community.

Improvements

Custom error handling

Create a custom error class to handle errors from the Twitter API. Currently, errors are handled by Ruby while some are returned as JSON responses from Twitter. This works OK-ish but ideally, I want to create a custom error class to improve error handling and improve debugging while using the gem.

Improved response object

Currently, the wrapper returns a custom Tweetkit::Response object. This response object contains a Tweets class, which is a collection of Tweet objects that contain the tweet content and id. Depending on the fields and expansions you included when calling the endpoints, you may also have access to a Resources and/or Meta class which contains tweet metadata, fields and expansions (read more on Twitter fields and expansions in the v2 endpoints).

The plan is to improve the response both usage and implementation-wise, so it'll be easy to work with it and also understand the code if needed.

Pagination

Currently, results are returned as is from Twitter's API, with pagination handled by passing in options when calling. This can be improved by introducing pagination to every (necessary) endpoint without the need for a page option. Instead, you can call response.next instead to get the next page or response.prev to go back.

Better client connection structure

Currently, the gem is relying on the internal Faraday module to fetch results with a simple get call. This may not work well for repeated calls in a short span of time or if concurrency is needed. I'm looking to make the connection module customizable so it can be tweaked for your desired HTTP library and specific use cases.

Documentation and tests

There is currently no documentation and tests for the gem. This will be added along the way as I work on the features and improvements.

These are the current tasks that are in my roadmap for this gem. Feel free to create issues or even submit PRs if you're interested. Thanks again for using this gem! :)

getting:

Authenticating with OAuth 2.0 Application-Only is forbidden for this endpoint

when calling the post_tweet method. What can I do to fix this? My credentials are all correct.

Interesting gem you have here 😉

I am wondering if you will be adding an explicit license to this project?

Thank you @julianfssen !

Unfortunately I'm getting an error updating to the latest version:

`
$ gem install tweetkit
ERROR: While executing gem ... (Gem::Resolver::Molinillo::CircularDependencyError)
There is a circular dependency between faraday-net_http and faraday

$ ruby -v
ruby 2.7.2p137 (2020-10-01 revision 5445e04352) [x86_64-darwin20]
`

Originally posted by @samuel-henry in #9 (comment)

Thanks for this code!

I got really confused trying to figure out how to post a tweet using v2 API with oauth -- after my previous code that had been using v1 API got shut off by twitter last month -- and I couldn't find much to do this for me or even demo how to do in ruby. Until I found this, and it works, I can tweet!

result = client.post_tweet(text: "some text")

But I am not finding any great way to determine if the result of that operation was successful or not though.

The actual HTTP status returned by twitter might be helpful (check is it 200?), but that information does not seem to actually be preserved by tweetdeck logic, it just gets through away?

All I find to look at is if it was succesful then result.response looks something like:

{"data"=>{"edit_history_tweet_ids"=>["999999999998"], "id"=>"9999999999999", "text"=>"some text"}}

BUT if it was NOT successful, then result.response looks something like:

{"title"=>"Unauthorized", "type"=>"about:blank", "status"=>401, "detail"=>"Unauthorized"}

OK, so in the error case I can check result.response["status"]... but if result.response["status"] is NOT present, I should assume success, maybe.... this is all seeming kind of fragile and unclear.

Can/should additional fields be added to the object returned from operations like post_tweet? An http_response_status (200, 401, etc)? Preserve and provide access to the entire faraday response that came back, in case the caller wants to look at it? A simple success? method, where tweetkit tries to figure out for me if the operation succeeded and just let me know? Other?

Thanks again for the code!

Hello, first off, great job so far. I am just getting started with the twitter api and had a hard time figuring it out. Seems like we're in a transition from oauth 1.a to 2 and v1 to v2. The documentation on twitter is a little all over the place. Anyway, your gem is saving me a lot of time, so thank you.

So when I first started to test this gem I noticed that if you follow the documentation and call

  response = client.tweet(1234567890)

you get an error "wrong number of arguments".

I then realized that the version "v0.1.4 " is missing a lot, so I switched to

gem 'tweetkit', git: '[email protected]:julianfssen/tweetkit.git', ref: '0284331' #allows posting to twitter

The error persisted though. Looked into the code a bit and realized the options were not optional and I had to pass in a empty {}

response = client.tweet(1234567890, *{})

I know this gem is in dev, but figured it might help to point this out.

1. Probably should make the options optional or specify they are't optional in the documentation. The syntax of having to *{} isn't ideal.

2. Does it make sense to make a new tag with the latest functionality? What I needed was the .post_tweet functionality so I switched to that specific ref.