Trying to parse xml with nokogiri and ruby

Question

I am trying to parse the xml below to get the email address out. I can get the messageid but I think having the a: in front is enabling me to use xpath. Not sure how to pull out the email address. I am trying

xml.xpath("//s:Body/Discover/request/EmailAddress").children.text.to_s

and

xml.xpath("//s:Body/Discover/EmailAddress").children.text.to_s

if i do xml.xpath("//s:Body").children.text.to_s i get the email and the version with all the newlines and tabs but i do not want to parse the email out if i do not have to.

<s:Envelope xmlns:a="http://www.w3.org/2005/08/addressing" xmlns:s="http://www.w3.org/2003/05/soap-envelope"> 
    <s:Header> 
        <a:Action s:mustUnderstand="1">test url</a:Action> 
        <a:MessageID>mid</a:MessageID> 
        <a:ReplyTo> 
            <a:Address>test url</a:Address> 
        </a:ReplyTo> 
        <a:To s:mustUnderstand="1">test url</a:To> 
    </s:Header> 
    <s:Body> 
        <Discover xmlns="test url"> 
            <request xmlns:i="http://www.w3.org/2001/XMLSchema-instance"> 
                <EmailAddress>bob@xml.com</EmailAddress> 
                <RequestVersion>1.0</RequestVersion> 
            </request> 
        </Discover> 
    </s:Body> 
</s:Envelope>

Answer 1

The test url is preventing Nokogiri's Xpath from catching on to your namespacing within s:Body. Try simply

 email = xml.xpath("//s:Body").first.to_xml.scan(/<EmailAddress>([^<]+)/)[0][0]

Answer 2

The Discover element (and its children) are in a different namespace, and you need to specify this in your query. The second argument to the xpath method is a hash where you can associate prefixes used in the query with namespace urls. Have a look at the section on namespaces in the Nokogiri tutorial.

With Nokogiri, if you don’t specify a namespace hash it will automatically register any namespaces defined on the root node for you. In this case that is the a prefix for http://www.w3.org/2005/08/addressing and the s prefix for http://www.w3.org/2003/05/soap-envelope. This is why your query for //s:Body works. The namespace declaration for Discover isn’t on the root, so you have to register it yourself.

When you provide your own namespace hash Nokogiri doesn’t add those defined on the root, so you will also need to include any of those used in your query.

In your case the following will find the EmailAddress node. The actual prefix you used doesn’t matter (here I’ve chosen t) as long as the URI matches).

xml.xpath('//s:Body/t:Discover/t:request/t:EmailAddress',
    's' => "http://www.w3.org/2003/05/soap-envelope",
    't' => "test url")