AnyEvent-CouchDB

 view release on metacpan or  search on metacpan

lib/AnyEvent/CouchDB.pm  view on Meta::CPAN


  my $db = $couch->db('database');
  $db    = couchdb('database');
  $db    = couchdb('http://somewhere.com:7777/database/');

With authentication:

  # user is the username and s3cret is the password
  $db = couchdb('http://user:s3cret@somewhere.com:7777/database');

Work with individual CouchDB documents;

  my $user = $db->open_doc('~larry')->recv;
  $user->{name} = "larry";
  $db->save_doc($user)->recv;

Query a view:

  $db->view('users/all', { startkey => 'b', endkey => 'bZZZ' })->recv

Finally, an asynchronous example:

  # Calling cb allow you to set a callback that will run when results are available.
  $db->all_docs->cb(sub {
    my ($cv) = @_;
    print pp( $cv->recv ), "\n";
  });

  # However, you have to be in an event loop at some point in time.
  AnyEvent->condvar->recv;

=head1 DESCRIPTION

AnyEvent::CouchDB is a non-blocking CouchDB client implemented on top of the
L<AnyEvent> framework.  Using this library will give you the ability to run
many CouchDB requests asynchronously, and it was intended to be used within
a L<Coro>+L<AnyEvent> environment.  However, it can also be used synchronously
if you want.

Its API is based on jquery.couch.js, but we've adapted the API slightly so that
it makes sense in an asynchronous Perl environment.

=head2 AnyEvent condvars

The main thing you have to remember is that all the data retrieval methods
return an AnyEvent condvar, C<$cv>.  If you want the actual data from the
request, there are a few things you can do.

You may have noticed that many of the examples in the SYNOPSIS call C<recv>
on the condvar.  You're allowed to do this under 2 circumstances:

=over 4

=item Either you're in a main program,

Main programs are "allowed to call C<recv> blockingly", according to the
author of L<AnyEvent>.

=item or you're in a Coro + AnyEvent environment.

When you call C<recv> inside a coroutine, only that coroutine is blocked
while other coroutines remain active.  Thus, the program as a whole is
still responsive.

=back

If you're not using Coro, and you don't want your whole program to block,
what you should do is call C<cb> on the condvar, and give it a coderef to
execute when the results come back.  The coderef will be given a condvar
as a parameter, and it can call C<recv> on it to get the data.  The final
example in the SYNOPSIS gives a brief example of this.

Also note that C<recv> will throw an exception if the request fails, so be
prepared to catch exceptions where appropriate.

Please read the L<AnyEvent> documentation for more information on the proper
use of condvars.

=head2 The \%options Parameter

Many data retrieval methods will take an optional C<\%options> hashref.
Most of these options get turned into CGI query parameters.  The standard
CouchDB parameters are as follows:

=over 4

=item key=keyvalue

This lets you pick out one document with the specified key value.

=item startkey=keyvalue

This makes it so that lists start with a key value that is greater than or
equal to the specified key value.

=item startkey_docid=docid

This makes it so that lists start with a document with the specified docid.

=item endkey=keyvalue

This makes it so that lists end with a key value that is less than or
equal to the specified key value.

=item count=max_rows_to_return

This limits the number of results to the specified number or less.
If count is set to 0, you won't get any rows, but you I<will> get
the metadata for the request you made.

=item update=boolean

If you set C<update> to C<false>, CouchDB will skip doing any updating of a
view.  This will speed up the request, but you might not see all the latest
data.

=item descending=boolean

Views are sorted by their keys in ascending order.  However, if you set
C<descending> to C<true>, they'll come back in descending order.

=item skip=rows_to_skip



( run in 1.860 second using v1.01-cache-2.11-cpan-39bf76dae61 )