9.2 KiB
Go XML Formatter
Synopsis
The Go XML Formatter, xmlfmt, will format the XML string in a readable way.
package main
import "github.com/go-xmlfmt/xmlfmt"
func main() {
xmlfmt.NL = "\n"
xml1 := `<root><this><is>a</is><test /><message><!-- with comment --><org><cn>Some org-or-other</cn><ph>Wouldnt you like to know</ph></org><contact><fn>Pat</fn><ln>Califia</ln></contact></message></this></root>`
x := xmlfmt.FormatXML(xml1, "\t", " ")
print(x)
// If the XML Comments have nested tags in them
xml1 = `<book> <author>Fred</author>
<!--
<price>20</price><currency>USD</currency>
-->
<isbn>23456</isbn> </book>`
x = xmlfmt.FormatXML(xml1, "", " ", true)
print(x)
}
Output:
<root>
<this>
<is>a
</is>
<test />
<message>
<!-- with comment -->
<org>
<cn>Some org-or-other
</cn>
<ph>Wouldnt you like to know
</ph>
</org>
<contact>
<fn>Pat
</fn>
<ln>Califia
</ln>
</contact>
</message>
</this>
</root>
<book>
<author>Fred
</author>
<!-- <price>20</price><currency>USD</currency> -->
<isbn>23456
</isbn>
</book>
There is no XML decoding and encoding involved, only pure regular expression matching and replacing. So it is much faster than going through decoding and encoding procedures. Moreover, the exact XML source string is preserved, instead of being changed by the encoder. This is why this package exists in the first place.
Note that
- the XML is mainly used in Windows environments, thus the default line ending is in Windows'
CRLFformat. To change the default line ending, see the above sample code (first line). - the case of XML comments nested within XML comments is not supported. Please avoid them or use any other tools to correct them before using this package.
- don't turn on the
nestedTagsInCommentsparameter blindly, as the code has become 10+ times more complicated because of it.
Command
To use it on command line, check out xmlfmt:
$ xmlfmt
XML Formatter
Version 1.1.0 built on 2021-12-06
Copyright (C) 2021, Antonio Sun
The xmlfmt will format the XML string without rewriting the document
Options:
-h, --help display help information
-f, --file *The xml file to read from (or stdin)
-p, --prefix each element begins on a new line and this prefix
-i, --indent[= ] indent string for nested elements
-n, --nested nested tags in comments
$ xmlfmt -f https://pastebin.com/raw/z3euQ5PR
<root>
<this>
<is>a
</is>
<test />
<message>
<!-- with comment -->
<org>
<cn>Some org-or-other
</cn>
<ph>Wouldnt you like to know
</ph>
</org>
<contact>
<fn>Pat
</fn>
<ln>Califia
</ln>
</contact>
</message>
</this>
</root>
$ xmlfmt -f https://pastebin.com/raw/Zs0qy0qz -n
<book>
<author>Fred
</author>
<!-- <price>20</price><currency>USD</currency> -->
<isbn>23456
</isbn>
</book>
Justification
The format
The Go XML Formatter is not called XML Beautifier because the result is not exactly as what people would expect -- some, but not all, closing tags stays on the same line, just as shown above. Having been looking at the result and thinking over it, I now think it is actually a better way to present it, as those closing tags on the same line are better stay that way in my opinion. I.e.,
When it comes to very big XML strings, which is what I’m dealing every day, saving spaces by not allowing those closing tags taking extra lines is plus instead of negative to me.
The alternative
To format it “properly”, i.e., as what people would normally see, is very hard using pure regular expression. In fact, according to Sam Whited from the go-nuts mlist,
Regular expression is, well, regular. This means that they can parse regular grammars, but can't parse context free grammars (like XML). It is actually impossible to use a regex to do this task; it will always be fragile, unfortunately.
So if the output format is so important to you, then unfortunately you have to go through decoding and encoding procedures. But there are some drawbacks as well, as put by James McGill, in http://stackoverflow.com/questions/21117161, besides such method being slow:
I like this solution, but am still in search of a Golang XML formatter/prettyprinter that doesn't rewrite the document (other than formatting whitespace). Marshalling or using the Encoder will change namespace declarations.
For example an element like "< ns1:Element />" will be translated to something like '< Element xmlns="http://bla...bla/ns1" >< /Element >' which seems harmless enough except when the intent is to not alter the xml other than formatting. -- James McGill Nov 12 '15
Using Sam's code as an example,
https://play.golang.org/p/JUqQY3WpW5
The above code formats the following XML
<soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:ns="http://example.com/ns">
<soapenv:Header/>
<soapenv:Body>
<ns:request>
<ns:customer>
<ns:id>123</ns:id>
<ns:name type="NCHZ">John Brown</ns:name>
</ns:customer>
</ns:request>
</soapenv:Body>
</soapenv:Envelope>
into this:
<Envelope xmlns="http://schemas.xmlsoap.org/soap/envelope/" xmlns:_xmlns="xmlns" _xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" _xmlns:ns="http://example.com/ns">
<Header xmlns="http://schemas.xmlsoap.org/soap/envelope/"></Header>
<Body xmlns="http://schemas.xmlsoap.org/soap/envelope/">
<request xmlns="http://example.com/ns">
<customer xmlns="http://example.com/ns">
<id xmlns="http://example.com/ns">123</id>
<name xmlns="http://example.com/ns" type="NCHZ">John Brown</name>
</customer>
</request>
</Body>
</Envelope>
I know they are syntactically the same, however the problem is that they look totally different.
That's why there is this package, an XML Beautifier that doesn't rewrite the document.
Credit
The credit goes to diotalevi from his post at http://www.perlmonks.org/?node_id=261292.
However, it does not work for all cases. For example,
$ echo '<Envelope xmlns=http://schemas.xmlsoap.org/soap/envelope/ xmlns:_xmlns=xmlns _xmlns:soapenv=http://schemas.xmlsoap.org/soap/envelope/ _xmlns:ns=http://example.com/ns><Header xmlns=http://schemas.xmlsoap.org/soap/envelope/></Header><Body xmlns=http://schemas.xmlsoap.org/soap/envelope/><request xmlns=http://example.com/ns><customer xmlns=http://example.com/ns><id xmlns=http://example.com/ns>123</id><name xmlns=http://example.com/ns type=NCHZ>John Brown</name></customer></request></Body></Envelope>' | perl -pe 's/(?<=>)\s+(?=<)//g; s(<(/?)([^/>]+)(/?)>\s*(?=(</?))?)($indent+=$3?0:$1?-1:1;"<$1$2$3>".($1&&($4 eq"</")?"\n".(" "x$indent):$4?"\n".(" "x$indent):""))ge'
<Envelope xmlns=http://schemas.xmlsoap.org/soap/envelope/ xmlns:_xmlns=xmlns _xmlns:soapenv=http://schemas.xmlsoap.org/soap/envelope/ _xmlns:ns=http://example.com/ns><Header xmlns=http://schemas.xmlsoap.org/soap/envelope/></Header>
<Body xmlns=http://schemas.xmlsoap.org/soap/envelope/><request xmlns=http://example.com/ns><customer xmlns=http://example.com/ns><id xmlns=http://example.com/ns>123</id>
<name xmlns=http://example.com/ns type=NCHZ>John Brown</name>
</customer>
</request>
</Body>
</Envelope>
I simplified the algorithm, and now it should work for all cases:
echo '<Envelope xmlns="http://schemas.xmlsoap.org/soap/envelope/" xmlns:_xmlns="xmlns" _xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" _xmlns:ns="http://example.com/ns"><Header xmlns="http://schemas.xmlsoap.org/soap/envelope/"></Header><Body xmlns="http://schemas.xmlsoap.org/soap/envelope/"><request xmlns="http://example.com/ns"><customer xmlns="http://example.com/ns"><id xmlns="http://example.com/ns">123</id><name xmlns="http://example.com/ns" type="NCHZ">John Brown</name></customer></request></Body></Envelope>' | perl -pe 's/(?<=>)\s+(?=<)//g; s(<(/?)([^>]+)(/?)>)($indent+=$3?0:$1?-1:1;"<$1$2$3>"."\n".(" "x$indent))ge'
<Envelope xmlns="http://schemas.xmlsoap.org/soap/envelope/" xmlns:_xmlns="xmlns" _xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" _xmlns:ns="http://example.com/ns">
<Header xmlns="http://schemas.xmlsoap.org/soap/envelope/">
</Header>
<Body xmlns="http://schemas.xmlsoap.org/soap/envelope/">
<request xmlns="http://example.com/ns">
<customer xmlns="http://example.com/ns">
<id xmlns="http://example.com/ns">
123</id>
<name xmlns="http://example.com/ns" type="NCHZ">
John Brown</name>
</customer>
</request>
</Body>
</Envelope>
This package is a direct translate from above Perl code into Go, then further enhanced by @ruandao.