<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE chapter SYSTEM "chapter.dtd">
<chapter>
<header>
<copyright>
<year>2004</year><year>2013</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
The contents of this file are subject to the Erlang Public License,
Version 1.1, (the "License"); you may not use this file except in
compliance with the License. You should have received a copy of the
Erlang Public License along with this software. If not, it can be
retrieved online at http://www.erlang.org/.
Software distributed under the License is distributed on an "AS IS"
basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
the License for the specific language governing rights and limitations
under the License.
</legalnotice>
<title>xmerl</title>
<prepared>UKH/L Bertil Karlsson</prepared>
<docno></docno>
<date>2004-06-16</date>
<rev>D</rev>
<file>xmerl_ug.xml</file>
</header>
<section>
<title>Introduction</title>
<section>
<title>Features</title>
<p>The <em>xmerl</em> XML parser is able to parse XML documents
according to the XML 1.0 standard. As default it performs
well-formed parsing, (syntax checks and checks of well-formed
constraints). Optionally one can also use xmerl as a validating
parser, (validate according to referenced DTD and validating
constraints). By means of for example the xmerl_xs module it is
possible to transform the parsed result to other formats,
e.g. text, HTML, XML etc.</p>
</section>
<section>
<title>Overview</title>
<p>This document does not give an introduction to XML. There
are a lot of books available that describe XML from
different views. At the <url href="http://www.w3.org">www.W3.org</url> site you will find
the <url href="http://www.w3.org/TR/REC-xml/">XML 1.0 specification</url> and other related specs. One site were
you can find tutorials on XML and related specs is <url href="http://www.zvon.org">ZVON.org</url>.</p>
<p>However, here you will find some examples of how to use
and to what you can use xmerl. A detailed description of the
user interface can be found in the reference manual.</p>
<p>There are two known shortcomings in xmerl:</p>
<list type="bulleted">
<item>It cannot retrieve external entities on the Internet
by a URL reference, only resources in the local file
system.</item>
<item>xmerl can parse Unicode encoded data. But, it fails
on tag names, attribute names and other mark-up names that
are encoded Unicode characters not mapping on ASCII.</item>
</list>
<p>By parsing an XML document you will get a record,
displaying the structure of the document, as return
value. The record also holds the data of the document. xmerl
is convenient to use in for instance the following scenarios:</p>
<p>You need to retrieve data from XML documents. Your
Erlang software can handle information from the XML document
by extracting data from the data structure received by
parsing.</p>
<p>It is also possible to do further processing of parsed
XML with xmerl. If you want to change format of the XML
document to for instance HTML, text or other XML format you
can transform it. There is support for such transformations
in xmerl.</p>
<p>One may also convert arbitrary data to XML. So it for
instance is easy to make it readable by humans. In this case
you first create xmerl data structures out of your data, then
transform it to XML. </p>
<p>You can find examples of these three examples of usage
below.</p>
</section>
</section>
<section>
<title>xmerl User Interface Data Structure</title>
<p>The following records used by xmerl to save the parsed
data are defined in <c>xmerl.hrl</c></p>
<p>The result of a successful parsing is a tuple
<c>{DataStructure,M}</c>. <c>M</c> is the XML production Misc,
which is the mark-up that comes after the element of the
document. It is returned "as is". <c>DataStructure</c> is an
<c>xmlElement</c> record, that among others have the fields
<c>name</c>, <c>parents</c>, <c>attributes</c> and
<c>content</c> like:</p>
<pre>
#xmlElement{name=Name,
...
parents=Parents,
...
attributes=Attrs,
content=Content,
...} </pre>
<p>The name of the element is found in the <c>name</c>
field. In the <c>parents</c> field is the names of the parent
elements saved. Parents is a list of tuples where the first
element in each tuple is the name of the parent element. The
list is in reverse order.</p>
<p>The record <c>xmlAttribute</c> holds the name and value of
an attribute in the fields <c>name</c> and <c>value</c>. All
attributes of an element is a list of xmlAttribute in the
field <c>attributes</c> of the xmlElement record.
</p>
<p>The <c>content</c> field of the top element is a list of
records that shows the structure and data of the document. If
it is a simple document like: </p>
<pre>
<?xml version="1.0"?>
<dog>
Grand Danois
</dog> </pre>
<p>The parse result will be:</p>
<pre>
#xmlElement{name = dog,
...
parents = [],
...
attributes = [],
content = [{xmlText,[{dog,1}],1,[],"\
Grand Danois\
",text}],
...
} </pre>
<p>Where the content of the top element is:
<c>[{xmlText,[{dog,1}],1,[],"\ Grand Danois\ ",text}]</c>. Text will be returned in <c>xmlText</c> records. Though,
usually documents are more complex, and the content of the top
element will in that case be a nested structure with
xmlElement records that in turn may have complex content. All of
this reflects the structure of the XML document.</p>
<p>Space characters between mark-up as <c>space</c>,
<c>tab</c> and <c>line feed</c> are normalized and returned as
xmlText records.</p>
<section>
<title>Errors</title>
<p>An unsuccessful parse results in an error, which may be a
tuple <c>{error,Reason}</c> or an exit:
<c>{'EXIT',Reason}</c>. According to the XML 1.0 standard
there are <c>fatal error</c> and <c>error</c> situations. The
fatal errors <em>must</em> be detected by a conforming parser
while an error <em>may</em> be detected. Both categories of
errors are reported as fatal errors by this version of xmerl,
most often as an exit.</p>
</section>
</section>
<section>
<title>Getting Started</title>
<p>In the following examples we use the XML file
"motorcycles.xml" and the corresponding DTD
"motorcycles.dtd". motorcycles.xml looks like:
<marker id="motorcyclesxml"></marker>
</p>
<codeinclude file="motorcycles.txt" tag="" type="none"></codeinclude>
<p>and motorcycles.dtd looks like: </p>
<codeinclude file="motorcycles_dtd.txt" tag="" type="none"></codeinclude>
<p>If you want to parse the XML file motorcycles.xml you run
it in the Erlang shell like:</p>
<pre>
3> {ParsResult,Misc}=xmerl_scan:file("motorcycles.xml").
{{xmlElement,motorcycles,
motorcycles,
[],
{xmlNamespace,[],[]},
[],
1,
[],
[{xmlText,[{motorcycles,1}],1,[],"\
",text},
{xmlElement,bike,
bike,
[],
{xmlNamespace,[],[]},
[{motorcycles,1}],
2,
[{xmlAttribute,year,[],[],[],[]|...},
{xmlAttribute,color,[],[],[]|...}],
[{xmlText,[{bike,2},{motorcycles|...}],
1,
[]|...},
{xmlElement,name,name,[]|...},
{xmlText,[{...}|...],3|...},
{xmlElement,engine|...},
{xmlText|...},
{...}|...],
[],
".",
undeclared},
...
],
[],
".",
undeclared},
[]}
4> </pre>
<p>If you instead receives the XML doc as a string you can
parse it by <c>xmerl_scan:string/1</c>. Both file/2 and string/2
exists where the second argument is a list of options to the
parser, see the <seealso marker="xmerl_scan">reference manual</seealso>.</p>
</section>
<section>
<title>Example: Extracting Data From XML Content</title>
<p>In this example consider the situation where you want to
examine a particular data in the XML file. For instance, you
want to check for how long each motorcycle have been recorded.</p>
<p>Take a look at the DTD and observe that the structure of an
XML document that is conformant to this DTD must have one
motorcycles element (the root element). The motorcycles element
must have at least one bike element. After each bike element it
may be a date element. The content of the date element is
#PCDATA (Parsed Character DATA), i.e. raw text. Observe that if
#PCDATA must have a <c><![CDATA["<"]]></c> or a <c><![CDATA["&"]]></c> character it must
be written as <c><![CDATA["<"]]></c> and <c><![CDATA["&"]]></c>
respectively. Also other character entities exists similar to
the ones in HTML and SGML.</p>
<p>If you successfully parse the XML file with the validation
on as in:
<c>xmerl_scan:file('motorcycles.xml',[{validation,true}])</c>
you know that the XML document is valid and has the structure
according to the DTD.</p>
<p>Thus, knowing the allowed structure it is easy to write a
program that traverses the data structure and picks the
information in the xmlElements records with name date.</p>
<p>Observe that white space: each space, tab or line feed,
between mark-up results in an xmlText record.</p>
<p></p>
</section>
<section>
<title>Example: Create XML Out Of Arbitrary Data</title>
<p>For this task there are more than one way to go. The "brute
force" method is to create the records you need and feed your
data in the content and attribute fields of the appropriate
element.</p>
<p>There is support for this in xmerl by the "simple-form"
format. You can put your data in a simple-form data structure
and feed it into
<c>xmerl:export_simple(Content,Callback,RootAttributes)</c>. Content
may be a mixture of simple-form and xmerl records as xmlElement
and xmlText.</p>
<p>The Types are:</p>
<list type="bulleted">
<item>Content = [Element]</item>
<item>Callback = atom()</item>
<item>RootAttributes = [Attributes]</item>
</list>
<p>Element is any of:</p>
<list type="bulleted">
<item>{Tag, Attributes, Content}</item>
<item>{Tag, Content}</item>
<item>Tag</item>
<item>IOString</item>
<item>#xmlText{}</item>
<item>#xmlElement{}</item>
<item>#xmlPI{}</item>
<item>#xmlComment{}</item>
<item>#xmlDecl{}</item>
</list>
<p>The simple-form structure is any of <c>{Tag, Attributes, Content}</c>, <c>{Tag, Content}</c> or <c>Tag</c> where:</p>
<p></p>
<list type="bulleted">
<item>Tag = atom()</item>
<item>Attributes = [{Name, Value}| #xmlAttribute{}]</item>
<item>Name = atom()</item>
<item>Value = IOString | atom() | integer()</item>
</list>
<p>See also reference manual for
<seealso marker="xmerl#export_simple-3">xmerl</seealso></p>
<p>If you want to add the information about a black Harley
Davidsson 1200 cc Sportster motorcycle from 2003 that is in
shape as new in the motorcycles.xml document you can put the
data in a simple-form data structure like:</p>
<pre>
Data =
{bike,
[{year,"2003"},{color,"black"},{condition,"new"}],
[{name,
[{manufacturer,["Harley Davidsson"]},
{brandName,["XL1200C"]},
{additionalName,["Sportster"]}]},
{engine,
["V-engine, 2-cylinders, 1200 cc"]},
{kind,["custom"]},
{drive,["belt"]}]} </pre>
<p>In order to append this data to the end of the
motorcycles.xml document you have to parse the file and add Data
to the end of the root element content.</p>
<pre>
{RootEl,Misc}=xmerl_scan:file('motorcycles.xml'),
#xmlElement{content=Content} = RootEl,
NewContent=Content++lists:flatten([Data]),
NewRootEl=RootEl#xmlElement{content=NewContent}, </pre>
<p>Then you can run it through the export_simple/2 function: </p>
<pre>
{ok,IOF}=file:open('new_motorcycles.xml',[write]),
Export=xmerl:export_simple([NewRootEl],xmerl_xml),
io:format(IOF,"~s~n",[lists:flatten(Export)]), </pre>
<marker id="new_motorcyclesxml"></marker>
<p>The result would be: </p>
<codeinclude file="new_motorcycles.txt" tag="" type="none"></codeinclude>
<p>If it is important to get similar indentation and newlines
as in the original document you have to add #xmlText{} records
with space and newline values in appropriate places. It may also
be necessary to keep the original prolog where the DTD is
referenced. If so, it is possible to pass a RootAttribute
<c>{prolog,Value}</c> to <c>export_simple/3</c>. The following
example code fixes those changes in the previous example:</p>
<pre>
Data =
[#xmlText{value=" "},
{bike,[{year,"2003"},{color,"black"},{condition,"new"}],
[#xmlText{value="\
"},
{name,[#xmlText{value="\
"},
{manufacturer,["Harley Davidsson"]},
#xmlText{value="\
"},
{brandName,["XL1200C"]},
#xmlText{value="\
"},
{additionalName,["Sportster"]},
#xmlText{value="\
"}]},
{engine,["V-engine, 2-cylinders, 1200 cc"]},
#xmlText{value="\
"},
{kind,["custom"]},
#xmlText{value="\
"},
{drive,["belt"]},
#xmlText{value="\
"}]},
#xmlText{value="\
"}],
...
NewContent=Content++lists:flatten([Data]),
NewRootEl=RootEl#xmlElement{content=NewContent},
...
Prolog = ["<?xml version=\\"1.0\\" encoding=\\"utf-8\\" ?>
<!DOCTYPE motorcycles SYSTEM \\"motorcycles.dtd\\">\
"],
Export=xmerl:export_simple([NewRootEl],xmerl_xml,[{prolog,Prolog}]),
... </pre>
<p>The result will be: </p>
<codeinclude file="new_motorcycles2.txt" tag="" type="none"></codeinclude>
</section>
<section>
<title>Example: Transforming XML To HTML</title>
<p>Assume that you want to transform the <seealso marker="#motorcyclesxml">motorcycles.xml</seealso> document to
HTML. If you want the same structure and tags of the resulting
HTML document as of the XML document then you can use the
<c>xmerl:export/2</c> function. The following:</p>
<pre>
2> {Doc,Misc}=xmerl_scan:file('motorcycles.xml').
{{xmlElement,motorcycles,
motorcycles,
[],
{xmlNamespace,[],[]},
[],
1,
[],
[{xmlText,[{motorcycles,1}],1,[],"\
",text},
{xmlElement,bike,
...
3> DocHtml=xmerl:export([Doc],xmerl_html).
["<!DOCTYPE HTML PUBLIC \\"",
"-//W3C//DTD HTML 4.01 Transitional//EN",
"\\"",
[],
">\
",
[[["<","motorcycles",">"],
["\
",
[["<",
"bike",
[[" ","year","=\\"","2000","\\""],[" ","color","=\\"","black","\\""]],
">"],
... </pre>
<p>Will give the result <url href="result_export.html">result_export.html</url></p>
<p>Perhaps you want to do something more arranged for human
reading. Suppose that you want to list all different brands in
the beginning with links to each group of motorcycles. You also
want all motorcycles sorted by brand, then some flashy colors
on top of it. Thus you rearrange the order of the elements and
put in arbitrary HTML tags. This is possible to do by means of
the <url href="http://www.w3.org/Style/XSL/">XSL Transformation (XSLT)</url> like functionality in xmerl. </p>
<p>Even though the following example shows one way to transform data
from XML to HTML it also applies to transformations to other
formats.</p>
<p><c>xmerl_xs</c> does not implement the entire XSLT
specification but the basic functionality. For all details see
the <seealso marker="xmerl_xs">reference manual</seealso></p>
<p>First, some words about the xmerl_xs functionality:</p>
<p>You need to write template functions to be able to control
what kind of output you want. Thus if you want to encapsulate a
<c>bike</c> element in <p> tags you simply write a
function:</p>
<pre>
template(E = #xmlElement{name='bike'}) ->
["<p>",xslapply(fun template/1,E),"</p>"]; </pre>
<p>With <c>xslapply</c> you tell the XSLT processor in which
order it should traverse the XML structure. By default it goes
in preorder traversal, but with the following we make a
deliberate choice to break that order:</p>
<pre>
template(E = #xmlElement{name='bike'}) ->
["<p>",xslapply(fun template/1,select("bike/name/manufacturer")),"</p>"]; </pre>
<p>If you want to output the content of an XML element or an attribute you will get the value as a string by the <c>value_of</c> function:</p>
<pre>
template(E = #xmlElement{name='motorcycles'}) ->
["<p>",value_of(select("bike/name/manufacturer",E),"</p>"]; </pre>
<p>In the xmerl_xs functions you can provide a select(String)
call, which is an <url href="http://www.w3.org/TR/xpath">XPath</url>
functionality. For more details see the xmerl_xs <url href="xmerl_xs_examples.html">tutorial</url>.</p>
<p>Now, back to the example where we wanted to make the output
more arranged. With the template:</p>
<pre>
template(E = #xmlElement{name='motorcycles'}) ->
[ "<head>\
<title>motorcycles</title>\
</head>\
",
"<body>\
",
\011 "<h1>Used Motorcycles</h1>\
",
\011 "<ul>\
",
\011 remove_duplicates(value_of(select("bike/name/manufacturer",E))),
\011 "\
</ul>\
",
\011 sort_by_manufacturer(xslapply(fun template/1, E)),
"</body>\
",
\011 "</html>\
"]; </pre>
<p>We match on the top element and embed the inner parts in an
HTML body. Then we extract the string values of all motorcycle
brands, sort them and removes duplicates by
<c>remove_duplicates(value_of(select("bike/name/manufacturer", E)))</c>. We also process the substructure of the top element
and pass it to a function that sorts all motorcycle information
by brand according to the task formulation in the beginning of
this example.</p>
<p>The next template matches on the <c>bike</c> element:</p>
<pre>
template(E = #xmlElement{name='bike'}) ->
{value_of(select("name/manufacturer",E)),["<dt>",xslapply(fun template/1,select("name",E)),"</dt>",
"<dd><ul>\
",
"<li style="color:green">Manufacturing year: ",xslapply(fun template/1,select("@year",E)),"</li>\
",
"<li style="color:red">Color: ",xslapply(fun template/1,select("@color",E)),"</li>\
",
"<li style="color:blue">Shape : ",xslapply(fun template/1,select("@condition",E)),"</li>\
",
"</ul></dd>\
"]}; </pre>
<p>This creates a tuple with the brand of the motorcycle and
the output format. We use the brand name only for sorting
purpose. We have to end the template function with the "built
in clause" <c>template(E) -> built_in_rules(fun template/1, E).</c></p>
<p>The entire program is motorcycles2html.erl:</p>
<codeinclude file="motorcycles2html.erl" tag="" type="erl"></codeinclude>
<p>If we run it like this:
<c>motorcycles2html:process_to_file('result_xs.html', 'motorcycles2.xml').</c> The result will be <url href="result_xs.html">result_xs.html</url>. When the
input file is of the same structure as the previous
"motorcycles" XML files but it has a little more 'bike'
elements and the 'manufacturer' elements are not in order.</p>
</section>
</chapter>
