license: MIT
Convert Ruby Hashes to XML
A hash is meant to be a structured set of data. So is XML. The two are very similar in that they have the capability of nesting information within a tree structure. With XML you have nodes. With Hashes, you have key/value pairs. The value of an XML node is referenced by its parent's name. A hash value is referenced by its key. This basic lesson tells the majority of what you need to know about creating XML via Hashes in Ruby using the XmlFu gem.
Add this line to your application's Gemfile:
gem 'xml-fu'
And then execute:
$ bundle
Or install it yourself as:
$ gem install xml-fu
Configuration was reworked to be more flexible and provide a centralized object for configuration. As such,
the configuration options have been moved off of the XmlFu
module into the XmlFu.config
object. The XmlFu.configure
method still works the same for setting configurations, but for reading configuration variables, you need to go
through the XmlFu.config
object.
- ADDED
- support for OpenStruct conversion
- REMOVED
infer_simple_value_nodes
configuration optionXmlFu::Array.infer_node
XmlFu.parse()
- XmlFu is for generation of XML, not parsing.
- MOVED
- configuration options moved to
XmlFu.config
- configuration options moved to
Hash keys are translated into XML nodes (whether it be a document node or attribute node).
With Ruby, a hash key may be a string or a symbol. Strings will be preserved as they may
contain node namespacing (<foo:Bar>
would need preserved rather than converted). There are some
exceptions to this rule (especially with special key syntax discussed in Types of Nodes and the
like). Symbols will be converted into an XML safe name by lower camel-casing them. So :foo_bar
will become fooBar
. You may change the conversion algorithm to your liking by setting the
XmlFu.config.symbol_conversion_algorithm
to a lambda or proc of your liking.
For a complete list, reference XmlFu::Configuration::ALGORITHMS
:camelcase
:downcase
:lower_camelcase
(default):none
(result of:sym.to_s
):upcase
# Default Algorithm (:lower_camelcase)
XmlFu.xml( :FooBar => "bang" ) #=> "<fooBar>bang</fooBar>"
XmlFu.xml( :foo_bar => "bang" ) #=> "<fooBar>bang</fooBar>"
# Built-in Algorithms (:camelcase)
XmlFu.config.symbol_conversion_algorithm = :camelcase
XmlFu.xml( :Foo_Bar => "bang" ) #=> "<FooBar>bang</FooBar>"
XmlFu.xml( :foo_bar => "bang" ) #=> "<FooBar>bang</FooBar>"
# Built-in Algorithms (:downcase)
XmlFu.config.symbol_conversion_algorithm = :downcase
XmlFu.xml( :foo_bar => "bang" ) #=> "<foo_bar>bang</foo_bar>"
XmlFu.xml( :Foo_Bar => "bang" ) #=> "<foo_bar>bang</foo_bar>"
XmlFu.xml( :FOO => "bar" ) #=> "<foo>bar</foo>"
# Built-in Algorithms (:upcase)
XmlFu.config.symbol_conversion_algorithm = :upcase
XmlFu.xml( :foo_bar => "bang" ) #=> "<FOO_BAR>bang</FOO_BAR>"
XmlFu.xml( :Foo_Bar => "bang" ) #=> "<FOO_BAR>bang</FOO_BAR>"
XmlFu.xml( :foo => "bar" ) #=> "<FOO>bar</FOO>"
# Custom Algorithm
XmlFu.config.symbol_conversion_algorithm = lambda {|sym| sym.do_something_special }
Because there are multiple types of XML nodes, there are also multiple types of keys to denote them.
By default, XmlFu assumes that all XML nodes will contain closing tags. However, if you want to explicitly create a self-closing node, use the following syntax when you define the key.
XmlFu.xml("foo/" => "bar") #=> <foo/>
One thing to take note of this syntax is that XmlFu will ignore ANY value you throw at it if the key syntax denotes a self-closing tag. This is because a self-closing tag cannot have any contents (hence the use for a self-closing tag).
By default, if you pass a pure string as a value, special characters will be escaped to keep the XML compliant. If you know that the string is valid XML and can be trusted, you can add the exclamation point to the end of the key name to denote that XmlFu should NOT escape special characters in the value.
XmlFu.xml("foo" => "<bar/>") #=> <foo><bar/></foo>
XmlFu.xml("foo!" => "<bar/>") #=> <foo><bar/></foo>
Yes, the attributes of an XML node are nodes themselves, so we need a way of defining them. Since XPath syntax
uses @
to denote an attribute, so does XmlFu.
Note: Keep in mind, because of the way that XML::Builder
accepts an attributes hash and the way that
Ruby stores and retrieves values from a Hash, attributes won't always be generated in the same order.
We can only guarantee that the attributes will be present, not in any specific order.
XmlFu.xml(:agent => {
"@id" => "007",
"FirstName" => "James",
"LastName" => "Bond"
})
<agent id="007">
<FirstName>James</FirstName>
<LastName>Bond</LastName>
</agent>
The value in a key/value pair describes the key/node. Different value types determine the extent of this description.
Simple value types describe the contents of the XML node.
XmlFu.xml( :foo => "bar" ) #=> <foo>bar</foo>
XmlFu.xml( "foo" => "bar" ) #=> <foo>bar</foo>
XmlFu.xml( :foo => 0 ) #=> <foo>0</foo>
XmlFu.xml( :pi => 3.14159 ) #=> <pi>3.14159</pi>
XmlFu.xml( :foo => nil ) #=> <foo xsi:nil="true"/>
Hash are parsed for their translated values prior to returning a XmlFu value.
XmlFu.xml(:foo => {
:bar => {
:biz => "bang"
}
})
<foo>
<bar>
<biz>bang</biz>
</bar>
</foo>
Should you require setting node attributes as well as setting the value of the XML node, you may use the =
key in a nested hash to denote explicit content.
XmlFu.xml(:agent => {
"@id" => "007",
"=" => "James Bond"
})
<agent id="007">James Bond</agent>
This key will not get around the self-closing node rule. The only nodes that will be used in this case will be attribute nodes (additional content will be ignored).
XmlFu.xml("foo/" => {
"@id" => "123",
"=" => "You can't see me." # ignored (node is self-closing)
})
<foo id="123"/>
Since version 0.2.0, support has been added for converting an OpenStruct object. OpenStruct objects behave
similar to Hashes, but they do not allow the flexibility that Hashes provide when naming keys/methods. As such,
the advanced naming capabilities are not available with OpenStruct objects and key conversion will go through
XmlFu.config.symbol_conversion_algorithm
.
Since the value in a key/value pair is (for the most part) used as the contents of a node, there are some assumptions that XmlFu makes when dealing with Array values.
- For a typical key, the contents of the array are considered to be nodes to be contained within the
<key>
node.
XmlFu.xml( "SecretAgents" => [
{ "agent/" => { "@id"=>"006", "@name"=>"Alec Trevelyan" } },
{ "agent/" => { "@id"=>"007", "@name"=>"James Bond" } }
])
<SecretAgents>
<agent name="Alec Trevelyan" id="006"/>
<agent name="James Bond" id="007"/>
</SecretAgents>
There comes a time that you may want to declare the contents of an array as a collection of items denoted by the key name. Using the asterisk (also known for multiplication --- hence multiple keys) we denote that we want a collection of <key> nodes.
XmlFu.xml( "person*" => ["Bob", "Sheila"] )
<person>Bob</person>
<person>Sheila</person>
In this case, the value of "person*" is an array of two names. These names are to be the contents of multiple
<person>
nodes and the result is a set of sibling XML nodes with no parent.
How about a more complex example:
XmlFu.xml(
"person*" => {
"@foo" => "bar",
"=" => [
{
"@foo" => "nope", # override default
"=" => "Bob"
},
"Sheila"
]
}
)
<person foo="nope">Bob</person>
<person foo="bar">Sheila</person>
This is getting interesting, isn't it? In this example, we are setting a default "foo" attribute on each of the items in the collection of <person> nodes. However, you'll notice that we overwrote the default "foo" with Bob.
Array values are flattened prior to translation, to reduce the need to iterate over nested arrays.
XmlFu.xml(
:foo => [
[{"a/" => nil}, {"b/" => nil}],
{"c/" => nil},
[
[{"d/" => nil}, {"e/" => nil}],
{"f/" => nil}
]
]
)
<foo>
<a/>
<b/>
<c/>
<d/>
<e/>
<f/>
</foo>
:foo
in this case, is the parent node of its contents
Since with simple values, you cannot infer the value of their node container purely on their value, simple values are currently ignored in arrays and only Hashes are translated.
XmlFu.xml("foo" => [
{:bar => "biz"},
nil, # ignored
true, # ignored
false, # ignored
42, # ignored
3.14, # ignored
"simple string", # ignored
['another','array','of','values'] # ignored
])
<foo>
<bar>biz</bar>
</foo>
The following options are available to pass to XmlFu.xml(obj, options)
.
:instruct
- if
true
:- Adds
<xml version="1.0" encoding="UTF-8"?>
to generated XML
- Adds
- This will be overridden by
XmlFu.config.include_xml_declaration
- if
XmlFu.configure do |config|
config.symbol_conversion_algorithm = :default # (:lower_camelcase)
config.fail_on_invalid_construct = false # (false)
config.include_xml_declaration = nil # (nil)
end
This is used to convert symbol keys in a Hash to Strings.
When an unsupported object is passed to XmlFu.xml()
, the default action is to return nil
as
the result. When fail_on_invalid_construct
is enabled, XmlFu.xml()
will raise an ArgumentError
to denote that the passed object is not supported rather than fail silently.
Deals with adding/excluding <?xml version="1.0" encoding="UTF-8"?>
to generated XML
- When
true
, ALWAYS adds declaration to XML - When
false
, NEVER adds declaration to XML - When
nil
, control is given to:instruct
option ofXmlFu.xml()
(default)
- if key denotes self-closing node (key/)
- attributes are preserved with Hash values
- value and
=
values are ignored
- if key denotes collection (key*) with Array value
- Array is flattened
- Only Hash and Simple values are translated
- Hashes may override default attributes set by parent
- (applies to Array values only)
- if key denotes contents (key) with Array value
- Array is flattened
- Only Hash items in array are translated
- if value is Hash:
@
keys are node attributes=
key can be used to specify node content
- if value is simple value:
- it is the node content
- unless: key denotes a self-closing node
- Fork it
- Create your feature branch (
git checkout -b my-new-feature
) - Commit your changes (
git commit -am 'Added some feature'
) - Push to the branch (
git push origin my-new-feature
) - Create new Pull Request