NAME

Net::Twitter::Manual::MigratingToV1_1 - Migrating from Twitter API v1 to v1.1

VERSION

version 4.01043

SYNOPSIS

    use Net::Twitter
    my $nt = Net::Twitter->new(
        traits              => [qw/API::RESTv1_1/],
        consumer_key        => $consumer_key,
        consumer_secret     => $consumer_secret,
        access_token        => $access_token,
        access_token_secret => $access_token_secret,
    );

DESCRIPTION

Net::Twitter prior to version 4.0 used Twitter API version 1. Twitter API v1.1 introduced changes that are not entirely backwards compatible, requiring some changes to calling code.

Net::Twitter attempts to provided backwards compatibility where possible. This document describes the changes required to your existing application code, using Net::Twitter, for use with Twitter's API v1.1.

FIRST

EXCEPTIONS

Trait RateLimit incompatible with API::RESTv1_1

The "RateLimit" trait is incompatible with Twitter API v1.1. Rate limiting is one of the most extensive changes in v1.1. In v1 there were two hourly rate limits, one per IP address for unauthenticated calls, and one per-user/application for authenticated calls. In v1.1, all calls must be authenticated, and each API endpoint (i.e., each method) has it's own rate limit. Rather than hourly, the new rate limits operate on a 15 minute window.

If your code currently uses the "RateLimit" role, you'll need to write some custom code provide equivalent functionality.

rate_limit_status

The return value for "rate_limit_status" is entirely different. See Twitter's API rate_limit_status <https://dev.twitter.com/docs/api/1.1/get/application/rate_limit_status> documentation for details.

blocking

blocking_ids

friends

followers

With API v1.1, these methods use cursor based paging. If you do not pass a "cursor" parameter, Twitter assumes "cursor => -1">. Existing code that expects an arrayref return value must be modified to expect a hashref and dereference the "users" slot:

    # With API v1
    my $r = $nt->friends;
    my @friends = @$r;
    # With API v1.1
    my $r = $nt->friends;
    my @friends = @{$r->{users}};
    

search

The "search" method semantics and return value are substantially different between Twitter API v1 and v1.1. In v1, "search" was provided by the "API::Search" trait. In v1.1, "search" is included in the "API::RESTv1_1" trait.

So, first, drop "API::Search" from your calls to "new". The "API::Search" trait is incompatible with "API::RESTv1_1".

In v1, Twitter returned a hashref with several keys containing meta data. The actual array of results were contained in the "results" slot:

    # With Twitter API v1
    my $nt = Net::Twitter->new(traits => [qw/API::REST API::Search/]);
    my $r = $nt->search('perl hacker');
    for my $status ( @{$r->{results} ) {
        # process each status...
    }
    

In v1.1, Twitter returns a hash with 2 slots: "search_metadata" and "statuses".

    # With Twitter API v1.1
    my $nt = Net::Twitter->new(traits => [qw/API::RESTv1_1/], %oauth_credentials);
    my $r = $nt->search('perl hacker');
    for my $status ( @{$r->{statuses} ) {
        # process each status...
    }
    

Paging through search results works differently in v1.1. In v1, Twitter offered a "page" parameter:

    # With Twitter API v1
    for ( my $page = 1; $page <= 10; ++$page ) {
        my $r = $nt->search({ q => $query, page => $page, rpp => 100 });
        last unless @{$r->{results}};
        # process a page of results...
    }
    

In v1.1, use "max_id" and "count" to get paged results:

    # With Twitter API v1.1
    for ( my %args = ( q => $query, count => 100 ), my $n = 0; $n < 1000; ) {
        my $r = $nt->search({ %args });
        last unless @{$r->{statuses}};
        $args{max_id} = $r->{statuses}[-1]{id} - 1;
        $n += @{$r->{statuses}};
        # process a page of results...
    }