Friday, July 22, 2011

[twitter-dev] Digest for twitter-development-talk@googlegroups.com - 5 Messages in 5 Topics

Group: http://groups.google.com/group/twitter-development-talk/topics

    RogersBlant <rogerstenning@gmail.com> Jul 22 01:52AM -0700 ^
     
    Morning, all - first post on here!
     
    This is posibly a non-starter, but let's see...
     
    In addition to my twitter account, I also run a blogspot blog, at
    http://rogersblant.blogspot.com/ . I use the Profile Widget to enable
    my twitters to be retweeted on my blog. The colours customisation is
    fine, as is the addition of the scrollbar and automatic resizing of
    the widget.
     
    However, the problem I have is that the font and the size of the font,
    are neither in keeping with, or match, my blog. I have tried using css
    to change both, but to no success.
     
    Given that the scripting supplied by Twitter utilises CSS within the
    script, I find that result to be surprising; I am therefore wondering
    if the widget can be modified either locally by me using CSS, or by
    Twitter adding more settings, to allow for different font families and
    sizes to be used?
    if anyone knows how this might be achieved locally by me, then I'm all
    ears; if not, then I suspect it'll be down to Twitter, as and when
    (if) they decide that it's a feature worth adding!
     
    Cheers for any help,
     
    Roger

     

    defthym <dimefthim@gmail.com> Jul 22 03:28AM -0700 ^
     
    Hi,
     
    I'm trying to parse the google:location from this url:
    http://search.twitter.com/search.atom?q=morning&rpp=100&geocode=48.853873,2.340088,2500km&page=1
     
    I am using the "xml" package in R language (but the solution can be
    similar in other languages) and my core (to parse the "title" and not
    the "google:location") is:
     
    twitter_url = paste('http://search.twitter.com/search.atom?
    q=morning&rpp=100&geocode=48.853873,2.340088,2500km1, sep='')
    mydata.xml <- xmlParseDoc(twitter_url, asText=F)
    mydata.vector <- xpathSApply(mydata.xml, '//s:entry/s:title, xmlValue,
    namespaces =c('s'='http://www.w3.org/2005/Atom'))
    mydata.vectors <- c(mydata.vector, mydata.vectors)
     
    when I replace "title" with "google:location" there is an error,
    probably because of not using the Google Base API (http://
    base.google.com/ns/1.0), because it is not available any more (but I
    am not absolutely sure about that)
     
    How can I parse the location? Any suggestions?
     
    Thanks,
    Dimitris

     

    CT <crowdtwist@gmail.com> Jul 21 01:01PM -0700 ^
     
    Hi.
     
    I've been using the UserStreams API for quite some time using the
    Phirehose patches. We are trying to get SiteStreams to work and made
    the necessary modifications to the UserStreams code to work with
    SiteStreams. Unfortunately, I am always receiving this error when I
    try to connect and get data from the stream:
     
    HTTP ERROR 400: Bad Request (45Must specify some followings to use
    sitestreams, e.g. follow=1,2,3,4)
     
    The following is output for how I'm trying to connect and what I'm
    sending:
     
    + Connecting to twitter stream: https://sitestream.twitter.com/2b/site.json
    with params: array ( 'delimited' => 'length', 'replies' => 'all',
    'with' => 'user', 'follow' => '17448575',)
    + Resolved host sitestream.twitter.com to 199.59.148.137
    + Connecting to 199.59.148.137
    + Connection established to 199.59.148.137
    Array
    (
    [oauth_consumer_key] => [Surpressed]
    [oauth_nonce] => 17c00ae9913dd68d5d9c493c1474807c
    [oauth_signature_method] => HMAC-SHA1
    [oauth_timestamp] => 1311277876
    [oauth_token] => [Surpressed]
    [oauth_version] => 1.0
    )
    string(307) "POST&https%3A%2F%2Fsitestream.twitter.com%2F2b
    %2Fsite.json&oauth_consumer_key%3D[Surpressed]%26oauth_nonce
    %3D17c00ae9913dd68d5d9c493c1474807c%26oauth_signature_method%3DHMAC-
    SHA1%26oauth_timestamp%3D1311277876%26oauth_token%3D[Surpressed]
    %26oauth_version%3D1.0"
    + POST /2b/site.json HTTP/1.1
    + Host: sitestream.twitter.com:443
    + Authorization: OAuth realm="https://sitestream.twitter.com/2b/
    site.json", oauth_consumer_key="[Surpressed]",
    oauth_token="[Surpressed]",
    oauth_nonce="17c00ae9913dd68d5d9c493c1474807c",
    oauth_timestamp="1311277876", oauth_signature_method="HMAC-SHA1",
    oauth_version="1.0", oauth_signature="%2FKI0z4%2BMKXZVPQig0tt3YMMtBEk
    %3D"
    +
    + delimited=length&replies=all&with=user&follow=17448575
    +
    + HTTP failure 1 of 20 connecting to stream: HTTP ERROR 400: Bad
    Request (45Must specify some followings to use sitestreams, e.g.
    follow=1,2,3,4). Sleeping for 10 seconds.
     
    I was wondering if anyone had any ideas as to why we would receive
    this message. Any help would be greatly appreciated, we're struggling
    with this issue.
     
    Thanks!

     

    Craig Walls <habuma@gmail.com> Jul 22 07:59AM -0700 ^
     
    I fully understand how followers/ids (or friends/ids) can be used with
    users/lookup to fetch profiles of users who follow or are friends of a user.
    Makes perfect sense, except for a bit of dissonance between the two
    resources.
     
    If I use followers/ids without setting the cursor parameter, I might get all
    of a user's followers, but if the user has a lot of followers (Charlie Sheen
    as an extreme example), I get an error. Okay...the documentation warned me
    that might happen. So, I set cursor to -1 to get the first 5000 of Charlie's
    followers. But then users/lookup only lets me fetch 100 profiles at a time.
    So I have to chunk the 5000 IDs I get from followers/ids into groups of 100
    and make 50 calls to users/lookup to get all 5000 profiles. Aside from the
    rate limit concerns, this takes quite some time to complete.
     
    Okay, so maybe I shouldn't fetch 5000 profiles at a time. Maybe I should
    only fetch 100 at a time and page through them as/if needed. Well, that's
    fine and even desirable...but now I have two page-oriented things to keep
    track of: The cursor for the given set of 5000 IDs and another index into
    that 5000 for the 100 IDs I want to fetch. It's not a big deal, I suppose,
    but it wouldn't be necessary if I could get followers/ids to give me only
    100 (or 50...or 25...or some number I specify) IDs at a time.
     
    Honestly, I don't like cursoring...there, I've said it and I do feel better.
    It seems to leak out an internal detail through the API. But also, I'm left
    with less control over which set and how many entries I get. I prefer the
    page and count/per_page approach taken in other API resources (which, btw,
    why the inconsistency with count on some resources and per_page on others?).
     
     
    What is recommended here? Keep track of the cursor and the index into that
    cursor for the 100 ID "page" I need? Do-able, I suppose, but it'd be so much
    easier if I could even get a cursor'd chunk of 100 IDs...or better yet
    paging.

     

    Fabien Penso <fabienpenso@gmail.com> Jul 22 10:17AM +0200 ^
     
    Am I the only one having this issue?
     

     

--
Have you visited the Developer Discussions feature on https://dev.twitter.com/discussions yet?
 
Twitter developer links:
Documentation and resources: https://dev.twitter.com/docs
API updates via Twitter: https://twitter.com/twitterapi
 
Unsubscribe or change your group membership settings: http://groups.google.com/group/twitter-development-talk/subscribe
Posted by at

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)