Discussing API v1.1

jnylund · 2013-02-05T12:47:20.000Z

even if we ask nicely? Seriously really need those, been waiting on going to 1.1 hoping you will fix those 2…

indr4_winata · 2013-02-05T15:09:44.000Z

Hi… I want to ask something. I have an apps that gather all user tweets using a cronjob and get their statuses (follower count, statuses count). The script is run every 1 hour so I can get hourly report. It works fine for 1 twitter accounts. when I have 2 twitter accounts to be gathered, the second twitter account lost some of its tweets and information, so the hourly report lose some of information (e.g: 1:00, 2:00, 3:00, … , 16:00, 18:00. 17:00 is missing!!). this still continues to other accounts. The fifth account is the last account which can be gathered, but the hourly report only have 1 hour reported per day. There are no information gathered from sixth twitter account.
I think that my cron is error, so I try giving a condition to get data only for 1 specific account when the cron start for the next hour, and it works fine again.

Is there any IP address or application key limitation that limits how many accounts can use the statuses/user_timeline API at the same time (or maybe 3-4 seconds delay each request)?
Can anyone tell me why is this happening?

smk · 2013-02-06T03:19:08.000Z

Twitter devs, please bump up the API limit. I tried to add ~500 people to a list via the API and due to API limiting, this has become a ten minute process.

Erwan · 2013-02-06T13:40:05.000Z

Does the search API version 1 still exist after migration to 1.1 ?

I used the search API version 1, and i don’t know how to make a request using JSONP with API 1.1 ?

episod · 2013-02-06T16:09:31.000Z

The Search API at search.twitter.com is in fact leaving when API v1 is all retired.

We don’t recommend using JSON-P with OAuth 1.0A. You’re better off using a server-side solution to access the Search API instead.

episod · 2013-02-06T16:10:07.000Z

Write-based methods aren’t rate limited like reads are in the API.

What specifically was happening when you were attempting this process? Consider opening a new thread instead.

episod · 2013-02-06T21:14:39.000Z

A minor feature added today:

[node:10363] and [node:10364] now take an optional “count” parameter, with a maximal value of 5000. If you decide to use this parameter, the cursored collection (and resultant next_cursor values) will be served respective to that more incremental count. This can be useful in constrained environments where the collection of 5,000 IDs may be too great to consume.

tapbot_paul · 2013-02-07T22:53:00.000Z

I’d like to request a rate limit increase for /direct_messages/show. Though not as widely used as /statuses/show and /users/show it feels similar enough that it should (IMO) have the same rate limits.

Use case is as follows. User is on a mobile device connected via a WAN (so streaming not possible) and has push notifications turned on. Push packets aren’t big enough to push the entire Direct Message but are big enough to pass in the id of the DM. As each push notification comes in a call is made to /direct_messages/show to fetch the latest DM in the thread. This gives the user the appearance of a streaming DM conversation but on a non streaming connection.

I understand that this is somewhat an edge case but I don’t think increasing the limit will cause a big spike on server load and would allow for a much better experience when quick firing DMs on the road.

Thanks,

tapbot_paul · 2013-02-07T23:05:45.000Z

I’d like to hop on and request an increase of this as well. If limiting per list is technically difficult perhaps perhaps an increase of the limit to 30 or 45 would suffice.

I believe the number of users that use List is relatively low but they tend to be the more engaged Twitter users and I think the current limit might pose a bit of a hardship for those users. There’s currently no Streaming available for lists so not really a way to mitigate making multiple requests to this call.

Erwan · 2013-02-08T08:16:05.000Z

Thanks for your answer. But i use codeBird JS and i need to use JSON-P for my request. Do you have a simple way to retrieve tweets for users using with API 1.1 and javascript ? (other than codebird)

qnnnews · 2013-02-10T21:50:45.000Z

Hi, could u clarify this

“direct_messages”:{"/direct_messages/show":{“limit”:15,“remaining”:15,“reset”:1360532306},"/direct_messages/sent_and_received":{“limit”:15,“remaining”:15,“reset”:1360532306},"/direct_messages/sent":{“limit”:15,“remaining”:15,“reset”:1360532306},"/direct_messages":{“limit”:15,“remaining”:15,“reset”:1360532306}}

Which rate control should i use to control direct messages calls and when :
direct_messages to control dm queries of messages received by active user
direct_messages/send to control dm sent by active user
direct_messages/send and received when ?, both ? 15 / window ? so only half quantum ?

Thanks (this send and received is new ?)

berkoguy · 2013-02-12T15:26:16.000Z

Hi, I’m migrating all my v1 API calls to 1.1 so thought I’d switch from the search api to /1.1/search/tweets.json at the same time. The response format includes a lot more user info which I’d like to suppress as I just need the user ID. I’ve tried using trim_user but this strips the user info completely. Is this a bug or do we need to cope with the more elaborate response?

Cheers,
Guy

episod · 2013-02-12T15:36:18.000Z

That search/tweets is omitting the user user/id & user/screen_name fields when trim_user is used is definitely a bug – I’ll make sure that gets fixed – until then, keep using the full response.

sferik · 2013-02-12T15:47:19.000Z

@episod There’s no bug here. The trim_user parameter works as it’s supposed to. The only issue is that it’s not documented on https://dev.twitter.com/docs/api/1.1/get/search/tweets.

@berkoguy This is the result of a long-overdue unification of the REST and Search APIs.

mgokhanbakal · 2013-02-13T05:26:17.000Z

I am trying to generate a retweet network graph, so I need to get retweets of a particular tweet from a person. As far as I understand from the API functions, there is no way to do that, am I right? Could you please help me about that?

P.S: Also, I am using twitter4j framework on visual studio.

Thanks.

FlockOfBirds · 2013-02-13T06:00:05.000Z

Hi @episod, I used the form to request elevated access, and got the same text back as @nivs. For my answer to that I got nothing back. How can I request elevated access so that I can still provide network analysis without making the user wait for hours or even days?

gkanters · 2013-02-13T10:07:00.000Z

I am having difficulty reading user_timelines this is the call I am using and it returns an error 215 (Bad authentication)

https://api.twitter.com/1.1/statuses/user_timeline.json?screen_name=gkanters&count=2

ezimir · 2013-02-13T11:05:37.000Z

Hi! One of our apps is currently in “shadow role” for Streaming API.

Will this elevated access be transfered to 1.1 or are we back to 5000 users using filter=follow?

Thanks.

tushar7saxena · 2013-02-13T11:11:59.000Z

I am currently in the process of developing an app for a major Marathon event for iPhone. The app has a twitter sheet which currently displays the tweets that are present on the timeline of the said event. Currently we are using v1.0 Twitter API. After the black out test on March 5th, is there a chance of Twitter reverting back to v1.0 if any problem arises or are they going to stick with v1.1 in any case. Please revert back asap as we are currently nearing completion of development and on the verge of launching the app on the app store.

episod · 2013-02-13T15:20:12.000Z

Are you using a signed OAuth request complete with access token?