NAME
DotCloud::Environment - easy handling of environment in dotcloud
VERSION
version 0.9.4
SYNOPSIS
# Most typical usage, suppose you have a shared 'lib' directory
# under the root of your dotCloud directory hierarchy
use DotCloud::Environment 'path_for';
use lib path_for('lib');
use My::Shared::Module; # in your project-root/lib directory
# Most typical usage when you set a default environment.json file
# in the root of your project and you need to access the variables
# of the 'redis' service
use DotCloud::Environment 'dotvars';
my $redis_vars = dotvars('redis');
# Not-very-typical usage examples from now on!
# get an object, fallback to $path if not in dotCloud deploy
my $dcenv = DotCloud::Environment->new(fallback_file => $path);
# you should now which services make part of your stack!
my $nosqldb_conf = $dcenv->service('nosqldb');
my $type = $nosqldb_conf->{type}; # e.g. mysql, redis, etc.
my $vars = $nosqldb_conf->{vars}; # e.g. login, password, host...
# suppose your nosqldb service is redis...
require Redis;
my $redis = Redis->new(server => "$vars->{host}:$vars->{port}");
$redis->auth($vars->{password});
# another service, similar approach
my $conf = $dcenv->service('database');
die 'not MySQL?!?' unless $conf->{type} eq 'mysql';
my ($host, $port, $user, $pass)
= @{$conf->{vars}}{qw< host port login password >}
require DBI;
my $dbh = DBI->connect("dbi:mysql:host=$host;port=$port;database=db",
$user, $pass, {RaiseError => 1});
DESCRIPTION
DotCloud::Environment is useful when you design applications to be
deployed in the dotCloud platform. It is assumed that you know what
dotCloud is (anyway, see ).
In general you will have multiple services in your application, and when
you are in one instance inside dotCloud you can access the configuration
of the relevant ones reading either /home/dotcloud/environment.yml or
/home/dotcloud/environment.json. For example, this lets your frontend or
backend applications know where the data services are, e.g. a Redis
database or a MySQL one.
This modules serves to two main goals:
* it reads either file to load the configuration of each service, so
that you can access this configuration easily as a hash of hashes;
* it lets you abstract from the assumption that you're actually in a
dotCloud instance, allowing you to use the same interface also in
your development environment.
With respect to the second goal, it should be observed that most of the
times in your development environment you don't have the same exact
situation as in dotCloud, e.g. it's improbable that you have a
/home/dotcloud directory around. With this module you can set a fallback
to be used in different ways, e.g.:
* providing a fallback file path to be loaded if
"/home/dotcloud/environment.json" is not found;
* setting up the "DOTCLOUD_ENVIRONMENT_FILE" environment variable to
point to the file to be used.
NOTE: as of version 0.9.1 this variable substitutes
"DOTCLOUD_ENVIRONMENT", which DotCloud started using for its own
purposes.
A Note On Available Data
Data about DotCloud services is organized according to the structure of
the variables set in the relevant files. There are four significant
parts:
application
there can be multiple applications you're loading variables from,
and DotCloud::Environment lets you distinguish them apart
service
this is the name of a service in DotCloud sense. For example, if you
have application whatever like this:
$ dotcloud list whatever
whatever (flavor: legacy):
- nosqldb (type: redis; instances: 1)
- sqldb (type: mysql; instances: 1)
- www (type: perl; instances: 1)
- backend (type: perl-worker; instances: 1)
you have four services defined: "nosqldb", "sqldb", "www" and
"backend"
subservice
this represents a subgroup of variables in a service. You should
always find two subservices: one is named "ssh", the other one has
the same name as the service type (e.g. "redis", "mysql",...).
It makes sense to consider "ssh" some kind of accessory information
and the other subservice as the "real" service.
variable name
this is the name of the variable, which is associated to a
subservice.
Values are assigned to variable names.
Suggested/Typical Usage
In order to keep your code clean, you will probably be dividing it
depending on the functional block that will be deployed as a service in
dotCloud. Suppose that you have a frontend service, a backend service
and a database; you probably have the following directory layout:
project
+- dotcloud.yml
+- backend
| | ...
| +- lib
| +- Backend.pm
+- frontend
| | ...
| +- lib
| +- FrontEnd.pm
+- lib
+- Shared.pm
Each service is put into a separate directory and all the code that they
both use (e.g. functions to connect to databases) is put in a common
"lib" directory.
How should you use DotCloud::Environment?
The main goal is to let it find the right "environment.json" (or,
equivalently, "environment.yml") depending on the environment you are
into. If you are in dotCloud there is actually no problem, because by
default the *right* "/home/dotcloud/environment.json" file is selected;
for your local development the best thing to do is to put the
configuration file in the project's root directory, which becomes like
this:
project
+- dotcloud.yml
+- backend
| | ...
| +- lib
| +- Backend.pm
+- frontend
| | ...
| +- lib
| +- FrontEnd.pm
+- lib
| +- Shared.pm
|
+- environment.json
Putting the file in that position lets DotCloud::Environment find it by
default when no "/home/dotcloud/environment.json" file (or the
equivalent YAML file) is found in the system. Which hopefully is the
case of your development environment.
In this case, you would have this in each service:
# -- in BackEnd.pm and FrontEnd.pm --
use DotCloud::Environment 'path_for';
use lib path_for('lib');
use Shared ...;
The function "path_for" helps you to set up the right path in @INC so
that the module can find the shared code.
In the shared module you can do this:
# -- in Shared.pm --
use DotCloud::Environment 'dotenv';
# ... when you need it...
my $service = dotenv()->service('service-name');
# ... now you have a hash ref which should have at least two
# elements: ssh and the real subservice type, e.g. mysql, redis, ...
my $redis_host = $service->{redis}{host};
Most of the time all you need is to access the variables related to a
specific service, so there's a shortcut for this:
use DotCloud::Environment 'dotvars';
my %vars = dotvars('service-name');
The "dotvars" shortcut tries its best to DWIM, i.e. it lets you specify
either the name of a service or the name of a subservice.
For example, suppose that you want to implement a function to connect to
a Redis service called "redisdb":
sub get_redis {
my %vars = dotvars('redisdb');
# it could also be:
#
# my %vars = dotvars('redis'); # name of service type
#
# if there is only one service of type redis
require Redis;
my $redis = Redis->new(server => "$vars{host}:$vars{port}");
$redis->auth($vars{password});
return $redis;
}
Of course you can use "dotenv"/"dotvars" directly in "FrontEnd.pm" and
"BackEnd.pm", but you will probably benefit from refactoring your common
code to avoid duplications.
METHODS
new
$dcenv = DotCloud::Environment->new(%params);
$dcenv = DotCloud::Environment->new({%params});
Create a new object. Parameters are:
no_load
don't attempt to load the configuration
environment_string
unconditionally use the provided string, ignoring everything else;
environment_file
unconditionally use the provided file, ignoring everything else;
fallback_string
use the provided string if other methods fail;
fallback_file
use the provided file if other methods fail.
backtrack
if nothing works and no fallback is set, look for suitable files in
filesystem. This option is activated by default, so you can use it
to *disable* it (e.g. with "backtrack => 0").
Unless "no_load" is passed and set to true, the object creation also
calls the "load" method.
Returns the new object or "croak"s if errors occur.
load
$dcenv->load(%params);
$dcenv->load({%params});
Load the configuration for an application. The accepted parameters are
"environment_string", "environment_file", "fallback_string",
"fallback_file" and "backtrack" with the same meaning as in the
constructor (see "new").
The sequence to get the configuration string is the following:
environment_string
from parameter passed to the method
environment_file
from parameter passed to the method
environment_string
from parameter set in the constructor
environment_file
from parameter set in the constructor
DOTCLOUD_ENVIRONMENT_FILE
environment variable (i.e. $ENV{DOTCLOUD_ENVIRONMENT_FILE}). Note
that this was formerly $ENV{DOTCLOUD_ENVIRONMENT} but due to
DotCloud starting using this variable it is no longer available.
$DotCloud::Environment::main_file_path
which defaults to /home/dotcloud/environment.json (you SHOULD NOT
change this variable unless you really know what you're doing)
fallback_string
from parameter passed to the method
fallback_file
from parameter passed to the method
fallback_string
from parameter set in the constructor
fallback_file
from parameter set in the constructor
If none of the above works there's still some hope in case there is
option "backtrack" (or it was specified to the constructor). In this
case, either file is searched recursively starting from the following
directories:
* the one returned by "find_code_dir" (but as if it were called by the
caller of "load", i.e. with a value of "n" equal to 1)
* the current working directory
* the directory of the file that called us.
Actually, option "backtrack" is enabled by default, so if you do not
want the behaviour above you have to explicitly disable it (e.g. passing
"backtrack => 0" in the constructor).
It is possible to load multiple configuration files from multiple
applications.
Returns a reference to the object itself.
as_json
%json_for = $dcenv->as_json();
$json_for = $dcenv->as_json();
Rebuild the JSON representations of all the applications.
Returns a hash (in list context) or an anonymous hash (in scalar
context) with each application name pointing to the relevant JSON
string.
as_yaml
%yaml_for = $dcenv->as_yaml();
$yaml_for = $dcenv->as_yaml();
Rebuild the YAML representations of all the applications.
Returns a hash (in list context) or an anonymous hash (in scalar
context) with each application name pointing to the relevant YAML
string.
merge_json
$dcenv->merge_json($json_string);
Add (or replace) the configuration of an application, provided as JSON
string. You should not need to do this explicitly, because this does the
same for you with autodetection of the format:
$dcenv->load(environment_string => $json_or_yaml_string);
Return a reference to the object itself.
merge_yaml
$dcenv->merge_yaml($yaml_string);
Add (or replace) the configuration of an application, provided as YAML
string. You should not need to do this explicitly, because this does the
same for you with autodetection of the format:
$dcenv->load(environment_string => $json_or_yaml_string);
application_names
my @names = $dcenv->application_names();
Returns the names of the applications loaded. Generally only one
application will be available, i.e. the one of the stack you're working
with.
applications
my %conf_for = $dcenv->applications();
my $conf_for = $dcenv->applications();
Get a hash (in list context) or anonymous hash (in scalar context) with
the relevant data of all the applications. Example:
{
app1 => {
project => 'app1',
environment => 'default',
service_id => 0,
service_name => 'www',
services => {
nosqldb => {
redis => {
login => 'redis',
password => 'wafadsfsdfdsfdas',
host => 'data.app1.dotcloud.com',
port => '12345',
}
}
sqldb => {
mysql => {
login => 'mysql',
password => 'wafadsfsdfdsfdas',
host => 'data.app1.dotcloud.com',
port => '54321',
}
}
}
},
app2 => {
# ...
}
}
application
my %conf_for = $dcenv->application($appname);
my $conf_for = $dcenv->application($appname);
Get a hash (in list context) or anonymous hash (in scalar context) with
the relevant data for the requested application. Example:
{
project => 'app1',
environment => 'default',
service_id => 0,
service_name => 'www',
services => {
nosqldb => {
redis => {
login => 'redis',
password => 'wafadsfsdfdsfdas',
host => 'data.app1.dotcloud.com',
port => '12345',
}
}
sqldb => {
mysql => {
login => 'mysql',
password => 'wafadsfsdfdsfdas',
host => 'data.app1.dotcloud.com',
port => '54321',
}
}
}
}
service
my %conf_for = $dcenv->service(%params); # also with \%params
my $conf_for = $dcenv->service(%params); # also with \%params
Get a hash (in list context) or anonymous hash (in scalar context) with
the relevant data for the requested service. Example:
{
ssh => {
host => 'data.app1.dotcloud.com',
port => '12345',
url => 'ssh://data.app1.dotcloud.com:12345/',
},
redis => {
login => 'redis',
password => 'wafadsfsdfdsfdas',
host => 'data.app1.dotcloud.com',
port => '12345',
}
}
The parameters are the following:
service
(Required) the name of the service.
application
(Optional) the name of the application.
The name of the application is optional because in most cases it can be
omitted, e.g. because there is only one application. The name can be
also provided in the service name, in line with what normally happens in
dotCloud where the complete name of a service is something like
"application.service".
This is the algorithm:
* if the name of the service is of the form "application.service", the
name is split into the two components;
* otherwise, if the application parameter is present it is used
* otherwise the service is searched among all the services of all the
applications.
If exactly one service is found it is returned, otherwise this method
"croak"s.
subservice
my %conf_for = $dcenv->subservice($subservice_name);
my %conf_for = $dcenv->subservice(%params); # also with \%params
my $conf_for = $dcenv->subservice(%params); # also with \%params
Get a hash (in list context) or anonymous hash (in scalar context) with
the relevant data for the requested subservice. Example:
redis => {
login => 'redis',
password => 'wafadsfsdfdsfdas',
host => 'data.app1.dotcloud.com',
port => '12345',
}
It can be called with a single non-reference scalar that represents the
subservice to look for. Otherwise it accepts the following parameters in
a hash or a reference to a hash:
subservice
the name of the subservice
service
the name of the service, see "service"
application
the name of the application, see "application"
with obvious meanings.
The application and the service name can also be specified in the
subservice name with separating dots like in the following examples:
application.service.subservice
service.subservice
These configurations in the subservice name override parameters of the
same name (e.g. specifying "service.subservice" overrides the
$params{service} input parameters).
Croaks if more than one subservice with the given name is found.
subservice_vars
my %vars = $dcenv->subservice_vars('subservice-name');
my $vars = $dcenv->subservice_vars('subservice-name');
my %vars = $dcenv->subservice_vars(%params); # also \%params
my $vars = $dcenv->subservice_vars(%params); # also \%params
my @values = $dcenv->subservice_vars(%params); # also \%params
my $values = $dcenv->subservice_vars(%params); # also \%params
Shorthand to get the configuration variables of a single subservice.
The input parameter list can be a single string with the name of the
subservice, or a hash/anonymous hash with parameters. Depending on the
input, the return value might be structured like a hash or like an
array:
subservice
the name of the subservice, see "subservice"
service
the name of the service, see "service"
application
the name of the application, see "application"
list
(Optional) if a list is provided, then the values corresponding to
each item in order is returned. This allows writing things like
this:
my ($host, $port, $password) = $dcenv->service_list(
service => 'nosqldb',
list => [ qw< host port password > ],
);
and get directly the values to put into variables. In this case, the
return value can be a list of values or an anonymous array with the
values.
If this parameter is not present, the whole name/value hash is
returned, either as a list or as an anonymous hash depending on the
context.
service_vars
my %vars = $dcenv->service_vars('service-name');
my $vars = $dcenv->service_vars('service-name');
my %vars = $dcenv->service_vars(%params); # also \%params
my $vars = $dcenv->service_vars(%params); # also \%params
my @values = $dcenv->service_vars(%params); # also \%params
my $values = $dcenv->service_vars(%params); # also \%params
Shorthand to get the configuration variables of a single service. This
assumes that a *main* subservice can be found in the requested service,
according to the following algorithm:
* first of all, a service is found with "service"
* the *background* service "ssh" is ignored
* if only one subservice remains in the service, it is assumed to be
the *main* subservice.
After this, the method behaves as if "subservice_vars" with the *main*
subservice were called.
FUNCTIONS
Nothing is exported by default, but you can import the following
functions. If you need both, you can use the ":all" tag, e.g.:
use DotCloud::Environment ':all';
This module uses Sub::Exporter under the hood; this means that if you're
not happy with the name of the imported subroutines you can provide your
own names, e.g.:
use DotCloud::Environment
dotvars => { -as => 'dotcloud_variables_for' };
my $vars = dotcloud_variables_for('my-service');
dotenv
my $singleton = dotenv();
This function returns a default instance of DotCloud::Environment that
should suit the needs for the typical/suggested usage. Subsequent calls
to the function always return the same object.
It can be useful if you don't want a global variable in your code, e.g.:
my @application_names = dotenv()->application_names();
# ...
my $vars = dotenv()->service_vars('my-sql-db');
dotvars
my $vars = dotvars('service-name-or-subservice-name');
This function gets the configuration variables for the provided service
using the default singleton instance. Most of the time this is exactly
what you want, and nothing more.
This function actually calls "service_vars" or "subservice_vars" behind
the scenes, you can pass all the parameters that the method accepts.
find_code_dir
my $code_directory = find_code_dir(%params);
This function tries to find the file dotcloud.yml that describes the
application backtracking from the current working directory and from the
directory containing the file that called us (i.e. what happens to be
"(caller($n))[1]").
Parameters:
n an integer, defaulting to 0, that tells how to call "caller()". You
shouldn't need to set it, anyway.
unix
when set, the name of the directory will be returned in Unix format,
so that you can use it with "use lib". By default the format is the
same as the system.
This should be useful if you want to put a default configuration file
there or if you want to set up a shared library directory. If you are
interested into this feature, anyway, look at "path_for" which is easier
to use.
path_for
use lib path_for('lib');
This function produces a list of paths that are suitable for "use lib".
It uses "find_code_dir" internally, see it for details.
You should pass a list of subdirectories which will be rebased using the
result of "find_code_dir" as a parent directory. If you are actually in
the dotCloud enviroment, the example above produces the path
"/home/dotcloud/code/lib".
Returns a list of Unix paths, one element for each input directory.
AUTHOR
Flavio Poletti
COPYRIGHT AND LICENSE
Copyright (C) 2011 by Flavio Poletti polettix@cpan.org.
This module is free software. You can redistribute it and/or modify it
under the terms of the Artistic License 2.0.
This program is distributed in the hope that it will be useful, but
without any warranty; without even the implied warranty of
merchantability or fitness for a particular purpose.