Config::Model::Loader

NAME

Config::Model::Loader - Load serialized data into config tree

SYNOPSIS

 use Config::Model;
 # define configuration tree object
 my $model = Config::Model->new;
  $model->create_config_class(
    name    => "Foo",
    element => [
        [qw/foo bar/] => {
            type       => 'leaf',
            value_type => 'string'
        },
    ]
 );
 $model ->create_config_class (
    name => "MyClass",
    element => [
        [qw/foo bar/] => {
            type       => 'leaf',
            value_type => 'string'
        },
        hash_of_nodes => {
            type       => 'hash',     # hash id
            index_type => 'string',
            cargo      => {
                type              => 'node',
                config_class_name => 'Foo'
            },
        },
        [qw/lista listb/] => {
                              type => 'list',
                              cargo =>  {type => 'leaf',
                                         value_type => 'string'
                                        }
                              },
    ],
 ) ;
 my $inst = $model->instance(root_class_name => 'MyClass' );
 my $root = $inst->config_root ;
 # put data
 my $steps = 'foo=FOO hash_of_nodes:fr foo=bonjour -
   hash_of_nodes:en foo=hello
   ! lista=foo,bar lista:2=baz
     listb:0=foo listb:1=baz';
 $root->load( steps => $steps );
 print $root->describe,"\n" ;
 # name         value        type         comment
 # foo          FOO          string
 # bar          [undef]      string
 # hash_of_nodes <Foo>        node hash    keys: "en" "fr"
 # lista        foo,bar,baz  list
 # listb        foo,baz      list
 # delete some data
 $root->load( steps => 'lista~2' );
 print $root->describe(element => 'lista'),"\n" ;
 # name         value        type         comment
 # lista        foo,bar      list
 # append some data
 $root->load( steps => q!hash_of_nodes:en foo.=" world"! );
 print $root->grab('hash_of_nodes:en')->describe(element => 'foo'),"\n" ;
 # name         value        type         comment
 # foo          "hello world" string

DESCRIPTION

This module is used directly by Config::Model::Node to load serialized configuration data into the configuration tree.

Serialized data can be written by the user or produced by Config::Model::Dumper while dumping data from a configuration tree.

CONSTRUCTOR

new

The constructor should be used only by Config::Model::Node.

Parameters:

start_node: node ref of the root of the tree (of sub-root) to start the load from. Stored as a weak reference.

load string syntax

The string is made of the following items (also called "actions") separated by spaces. These actions can be divided in 4 groups:

•: navigation: moving up and down the configuration tree.

•: list and hash operation: select, add or delete hash or list item (also known as "id" items)

•: leaf operation: select, modify or delecte leaf value

•: annotation: modify or delete configuration annotation (aka comment)

-: Go up one node

!: Go to the root node of the configuration tree.

xxx: Go down using "xxx" element. (For "node" type element)

/xxx: Go up until the element "xxx" is found. This search can be combined with one of the command specified below, e.g "/a_string="foo bar""

list and hash operation

xxx:yy: Go down using "xxx" element and id "yy" (For "hash" or "list" element with "node" cargo_type). Literal "\n" are replaced by real "\n" (LF in Unix).

xxx:.foreach_match(yy) or xxx:~yy

Go down using "xxx" element and loop over the ids that match the regex specified by "yy". (For "hash"). For instance, with "OpenSsh" model, you could do

 Host:~"/.*.debian.org/" user='foo-guest'

to set "foo-user" users for all your debian accounts. The leading and trailing '/' may be omitted. Be sure to surround the regexp with double quote if space are embedded in the regex. Note that the loop ends when the load command goes above the element where the loop is executed. For instance, the instruction below tries to execute "DX=BV" and "int_v=9" for all elements of "std_id" hash:

 std_id:~/^\w+$/ DX=Bv int_v=9

In the examples below only "DX=BV" is executed by the loop:

 std_id:~/^\w+$/ DX=Bv - int_v=9
 std_id:~/^\w+$/ DX=Bv ! int_v=9

The loop is done on all elements of the hash when no value is passed after "":~"" (mnemonic: an empty regexp matches any value).

xxx:.rm(yy) or xxx:-yy: Delete item referenced by "xxx" element and id "yy". For a list, this is equivalent to "splice xxx,yy,1". This command does not go down in the tree (since it has just deleted the element). I.e. a '"-"' is generally not needed afterwards.

xxx:.rm_value(yy) or xxx:-=yy: Remove the element whose value is "yy". For list or hash of leaves. Does not not complain if the value to delete is not found.

xxx:..rm_match(yy) or xxx:-~/yy/: Remove the element whose value matches "yy". For list or hash of leaves. Does not not complain if no value were deleted.

xxx:.substitute(/yy/zz/) or xxx:=~s/yy/zz/: Substitute a value with another. Perl switches can be used(e.g. "xxx:=~s/yy/zz/gi")

xxx:<yy or xxx:.push(yy): Push "yy" value on "xxx" list

xxx:>yy or xxx:.unshift(yy): Unshift "yy" value on "xxx" list

xxx:@ or xxx:.sort: Sort the list

xxx:.insert_at(yy,zz): Insert "zz" value on "xxx" list before index "yy".

xxx:.insert_before(yy,zz): Insert "zz" value on "xxx" list before value "yy".

xxx:.insert_before(/yy/,zz): Insert "zz" value on "xxx" list before value matching "yy".

xxx:.insort(zz): Insert "zz" value on "xxx" list so that existing alphanumeric order is preserved.

xxx:.insort(zz): For hash element containing nodes: creates a new hash element with "zz" key on "xxx" hash so that existing alphanumeric order of keys is preserved. Note that all keys are sorted once this instruction is called. Following instructions are applied on the created element. I.e. putting key order aside, "xxx:.insort(zz)" has the same effect as "xxx:zz" instruction.

xxx:.insort(zz,vv): For hash element containing leaves: creates a new hash element with "zz" key and assing value "vv" so that existing alphanumeric order of keys is preserved. Note that all keys are sorted once this instruction is called. Putting key order aside, "xxx:.insort(zz,vv)" has the same effect as "xxx:zz=vv" instruction.

xxx:.ensure(zz): Ensure that list "xxx" contains value "zz". If value "zz" is already stored in "xxx" list, this function does nothing. In the other case, value "zz" is inserted in alphabetical order.

xxx:=z1,z2,z3: Set list element "xxx" to list "z1,z2,z3". Use ",," for undef values, and "" for empty values. I.e, for a list "('a',undef,'','c')", use "a,,"",c".

xxx:yy=zz: For "hash" element containing "leaf" cargo_type. Set the leaf identified by key "yy" to value "zz". Using "xxx:~/yy/=zz" is also possible.

xxx:.copy(yy,zz): copy item "yy" in "zz" (hash or list).

xxx:.json("path/to/file.json/foo/bar"): Store "bar" content in array or hash. This should be used to store hash or list of values. You may store deep data structure. In this case, make sure that the structure of the loaded data matches the structure of the model. This won't happen by chance.

xxx:.clear: Clear the hash or list.

leaf operation

xxx=zz: Set element "xxx" to value "yy". load also accepts to set elements with a quoted string. (For "leaf" element) Literal "\n" are replaced by real "\n" (LF in Unix). Literal "\\" are replaced by "\". For instance "foo="a quoted string"" or "foo="\"bar\" and \"baz\""".

xxx=~s/foo/bar/

Apply the substitution to the value of xxx. "s/foo/bar/" is the standard Perl "s" substitution pattern. Patterns with white spaces must be surrounded by quotes:

  xxx=~"s/foo bar/bar baz/"

Perl pattern modifiers are accepted

  xxx=~s/FOO/bar/i

xxx~: Undef element "xxx"

xxx.=zzz: Appends "zzz" value to current value (valid for "leaf" elements).

xxx=.file(yyy): Store the content of file "yyy" in element "xxx". Store STDIn in value xxx when "yyy" is '-'.

xxx=.json(path/to/data.json/foo/bar)

Open file "data.json" and store value from JSON data extracted with "foo/bar" subpath. For instance, if "data.json" contains:

 {
    "foo": {
       "bar": 42
    }
 }

The instruction "baz=.json(data.json/foo/bar)" stores 42 in "baz" element.

xxx=.yaml(path/to/data.yaml/0/foo/bar)

Open file "data.yaml" and store value from YAML data extracted with "0/foo/bar" subpath. Since a YAML file can contain several documents (separated by "---" lines, the subpath must begin with a number to select the document containing the required value. For instance, if "data.yaml" contains:

  ---
  foo:
    bar: 42

The instruction "baz=.yaml(data.yaml/0/foo/bar)" stores 42 in "baz" element.

xxx=.env(yyy): Store the content of environment variable "yyy" in element "xxx".

annotation

xxx#zzz or xxx:yyy#zzz: Element annotation. Can be quoted or not quoted. Note that annotations are always placed at the end of an action item. I.e. "foo#comment", "foo:bar#comment" or "foo:bar=baz#comment" are valid. "foo#comment:bar" is not valid.

Quotes

You can surround indexes and values with double quotes. E.g.:

  a_string="\"foo\" and \"bar\""

Examples

You can use cme to modify configuration with "cme modify" command.

For instance, if Config::Model::Ssh is installed, you can run:

 cme modify ssh 'ControlMaster=auto ControlPath="~/.ssh/master-%r@%n:%p"'

To delete "Host *" entry:

 cme modify ssh 'Host:-"*"'

To specify 2 "Host" with a single command:

 cme modify ssh 'Host:"foo* bar*" ForwardX11=yes HostName="foo.com" - Host:baz HostName="baz.com"'

Note the '"-"' used to go up one node before ""Host:baz"". In this case, "up one node" leads to the "root node", so ""!"" could also be used instead of ""-"":

 cme modify ssh 'Host:"foo* bar*" ForwardX11=yes HostName="foo.com" ! Host:baz HostName="baz.com"'

Let's modify now the host name of using a ".org" domain instead of ".com". The ":~" operator uses a regexp to loop over several Host entries:

 cme modify ssh 'Host:~/ba[rz]/ HostName=~s/.com$/.org/'

Now that ssh config is mucked up with dummy entries, let's clean up:

 cme modify ssh 'Host:-"baz" Host:-"foo* bar*"'

Methods

load

Load data into the node tree (from the node passed with "node") and fill values as we go following the instructions passed with "steps". ("steps" can also be an array ref).

Parameters are:

steps (or step): A string or an array ref containing the steps to load. See "load string syntax" in above for a description of the string.

check: Whether to check values while loading. Either "yes" (default), "no" or "skip". Bad values are discarded when "check" is set to "skip".

caller_is_root: Change the target of the "!" command: when set, the "!" command go to caller node instead of going to root node. (default is false)

  The GNU Lesser General Public License, Version 2.1, February 1999

2022-07-28

perl v5.34.0