jpopesculian / twitter-v2-rs Goto Github PK

Title. The get_user_bookmarks method on TwitterApi only works if you have the users id, so I don't see why there can't be a method on TwitterApiWithUserCtx for that.

Could you include or point to some existing code that uses the PaginableApiResult to fetch a list, such as a user’s list of followers?

My first several attempts to write this worked, but are super ugly. I feel like I am missing an obvious way to write it that would be readable or even beautiful. I used the governor crate to rate–limit the requests, but I haven’t yet done anything to retry transient errors; I feel like it’s ugly enough already.

Hello there, I noticed that the in_reply_to_tweet_id method of the TweetBuilder struct accepts a user_id parameter rather than a tweet_id as the name of the method would suggest. Is this intentional, or does it actually accept a user_id meaning that the method could be renamed to in_reply_to_user_id?

Hi,

I had a question about v0.1.7 on crates.io as I don't see the v0.1.7 tag on Github.
Does it need to be pushed to the repo? I wanted to check before using the latest version.

Thanks,
Patrick

In the example on the readme it contains the line

let auth: Oauth2Token = serde_json::from_str(&stored_oauth2_token)?;

The variable for the json string was not in the example, so I looked at the Oauth2Token struct. However, I could not find out how to find the expires and scopes part of the token on the Twitter developer portal.

#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct Oauth2Token {
    access_token: AccessToken,
    refresh_token: Option<RefreshToken>,
    #[serde(with = "time::serde::rfc3339")]
    expires: OffsetDateTime,
    scopes: Vec<Scope>,
}

On the twitter developer portal I can only figure out how to create the access token and refresh token. I am not too familiar with the twitter api but I feel like it may be possible to put expires and scopes inside of an Option. The Oauth2Token struct should probably have the documentation for where you can find expires and scopes as well as a new() function for creating the token without deserializing from json.

I'm free to work on this issue this week, depending on which route you choose (documentation for finding expires and scopes or stuffing them in Options)

I’m surprised to discover this, because usually finalizing a builder consumes the builder and returns a wholly new object. As currently written, the builder cannot be a temporary; it has to be kept around as long as the request and response are so that they can refer to it.

Was this a deliberate choice? If so, can you document the reasoning for it? If not, can we change the send (and stream) methods to take Self instead?

In public data structures, could we use primitive types for identifiers instead of NumericId and StringId?

NumericId isn't semantic anyway like TweetId or UserId. I did see TweetId/UserId structs being generated by macros for use in request bodies, so maybe NumericId and StringId is useful behind the scenes. For the public API though, they make the IDs in common data types like Tweet and User less ergonomic to consume.

All fields on the Expansions struct are currently private. This makes it quite useless right now and makes it impossible to retrieve things like image URLs and usernames from the tweet API. They should be made public.

Twitter puts some information about your account’s rate limits in the response headers of every request. Most important is the time at which your rate limit is reset after a 429 response. See the documentation at https://developer.twitter.com/en/docs/twitter-api/rate-limits for all of the details.

I would like there to be some way to access this information, so that I can act on it. As a quick test, I tried sticking it on to the ApiError struct:

ApiError { title: "Too Many Requests", kind: "about:blank", status: 429, detail: "Too Many Requests", errors: [], reset: Some(1660051376) }

I added the simplest possible thing to api_error_for_status in api_result.rs:

let reset = self.headers().get("x-rate-limit-reset")
  .map(|v| v.to_str().unwrap().parse().unwrap());

and then added the resulting value to the ApiError that it creates.

This works, but it is not very elegant. It leaves two of the headers unavailable, and it would be rather nicer if it created a std::time::Duration for me, though that might be going a bit far. It also doesn’t do anything to make this information available on successful requests.

Do you prefer any particular way of implementing this?

To tweet with an image, we need to follow these steps I think.

1. upload the image as media
2. get the uploaded media id
3. call the TwitterApi::~~~~.post_tweet().add_media(my_media_id).send()

I do not know how to do step 1 - 2.

Thanks for your help.

This is probably because of the custom deserializer on duration but I can't figure why it's behaving that way.