Monday, August 28, 2006

Groovy: Parsing an XML file with Groovy

Groovy comes with a handy XML parser. With this parser you can load an XML file quickly and navigate it easy.

Currently I'm working with WebTest (http://webtest.canoo.com/) for testing functional aspects of an web application.

WebTest comes with a nice overview to present the results of the test runs. The report itself is transformed out of an XML file. For statistics reason, I'm using this XML file to extract specific data. This extraction is done with a little Groovy script.

To get an idea of the WebTest report XML file, have a look a the fragment below:

WebTestReport.xml:








...










name="resultFilename" value="response_1156746643278_invoke.html"/>
















...



This file can be parsed with the following code:


import java.text.SimpleDateFormat

FormatUS = new SimpleDateFormat('EEE MMM dd hh:mm:ss z yyyy', Locale.US)
FormatDE = new SimpleDateFormat('dd.MM.yyyy hh:mm:ss', Locale.GERMAN)

def folder = args?.size() ? args[0] : "data"
println "reading files from directory '$folder'"
def basedir = new File(folder)

// result files of the current run
files = basedir.listFiles().grep(~/.*xml$/)

List resultLines() {
List result = []
for (currentFile in files) {
println " processing $currentFile"
def summary = new XmlParser().parse(currentFile)

// List with all 'testresult' elements
def testresults = summary.testresult


testresults.each { testcase ->
def entry = [:]
entry.name = testcase.'@testspecname'.split(/\W/)[0]
entry.time = testcase.'@starttime'
entry.time = FormatUS.parse(entry.time)
entry.time = FormatDE.format(entry.time)
entry.ok = testcase.'@successful'

def rt = testcase.depthFirst().findAll { element ->
if (element.name() != 'parameter') return false
if (element.'@name' != 'taskName') return false
def roundtripTags = 'clickLink clickButton clickElement'
roundtripTags.tokenize().contains(element.'@value')
}
entry.roundtrips = rt.size()

def fields = testcase.depthFirst().findAll { element ->
if (element.name() != 'parameter') return false
if (element.'@name' != 'taskName') return false
def fieldsTags = 'selectForm setCheckbox setFileField setInputField setRadioButton'
fieldsTags.tokenize().contains(element.'@value')
}
entry.fields = fields.size()

def verifies = testcase.depthFirst().findAll { element ->
if (element.name() != 'parameter') return false
if (element.'@name' != 'taskName') return false
def fieldsTags = 'verifyCheckbox verifyCookie verifyDocumentURL verifyElement'
fieldsTags.tokenize().contains(element.'@value')
}
entry.verifies = verifies.size()

if (testcase.results.error) {
entry.resultsError = testcase.results.error.'@exception'[0]
} else {
entry.resultsError = ''
}

/*
if (testcase.results.failure) {
entry.resultsFailure = testcase.results.failure.'@message'[0].replaceAll(/\W/, " ")
println "####" + entry.resultsFailure
} else {
entry.resultsFailure = ''
}
*/

def groupSteps = testcase.results.step.findAll { step ->
step.parameter.any { param ->
param.'@name' == 'taskName' && param.'@value' == 'group'
}
}

def goodSteps = groupSteps.findAll{ it.result?.completed }

// assumed: there is always at least one group step, size() never 0
if (groupSteps.size() != 0) {
entry.complete_pct = goodSteps.size() / groupSteps.size()
} else {
entry.complete_pct = 0
}

if (goodSteps) {
entry.description = goodSteps[-1].parameter.findAll{it.'@name'=='description'}.'@value'[0]
} else {
entry.description = ''
}

result <<>
writer <<
"Testfall\t" <<
"Datum\t" <<
"Erfolgreich\t" <<
"Prozent\t" <<
"Error\t" <<
"Server Roundtrips\t" <<
"Number of Fields\t" <<
"Verifications\t" <<
"Beschreibung\n"
for (entry in resultLines()) {
writer <<
entry.name << "\t" <<
entry.time << "\t" <<
entry.ok << "\t" <<
entry.complete_pct << "\t" <<
entry.resultsError << "\t" <<
entry.roundtrips << "\t" <<
entry.fields << "\t" <<
entry.verifies << "\t" <<
entry.description << "\n"
}
}


Now you get a tab seperated file with the content of your result files from the web testing runs. I
import this file into Excel and create a pivot table to understand the results better.

I hope this helps :)

Friday, August 25, 2006

Groovy: Get all XML files from a directory tree

Currently I'm working in a project, where Groovy (a dynamic scripting language for the JVM) is used. One of the project members is Dierk König, which is a Groovy commiter and the author of Groovy in Action. This gives a real insight in the beauty of Groovy :)

Below you can see a nice solution for finding all XML files in a directory tree:


// pass a directory or use the current directory
def folder = args?.size() ? args[0] : "."
println "reading files from directory '$folder'"
def basedir = new File(folder)

// result files of the current run
files = basedir.listFiles().grep(~/.*xml$/)

List resultLines() {
List result = []
for (currentFile in files) {
println " processing $currentFile"
}
return result
}

Monday, August 14, 2006

GWT Widget Library

Google released a new version of GWT today, version 1.1.0. There were some API changes made that broke a few things in the GWT-WL, namely the FormPanel and SVG support. This release is to make the GWT-WL compatible with the latest GWT release. Below are a list of changes...

[from the site]

Read more at gwt-widget.sourceforge....

Friday, August 11, 2006

Eclipse Summit Europe 2006

The Eclipse Summit Europe will take place in Esslingen, Germany from October 11-12, 2006

Read more at www.eclipsecon.org/summ...

Eclipse 3.3 M1 - New and Noteworthy

It's amazing, how fast the Eclipse people work! Thanks to all of them :)

Read more at download.eclipse.org/ec...

Tuesday, August 08, 2006

Lint4j Maven Plugin - Lint4j

Lint4j comes with a Maven Plugin that simplifies generation of reports significantly.

I'll have to check this plugin with WebTestClipse ...

Read more at www.jutils.com/maven-pl...

dzone.com

An interesting site for developers. It's similar to del.icio.us and stumbleupon , but more focused on developers.

I like these community driven sites, because it allows me to find high quality sites.

Read more at www.dzone.com/

Thursday, August 03, 2006

SourceForge.net: WebTestClipse

I added WebTestClipse as a project to Sourceforge.net ( http://sourceforge.net/projects/webtestclipse/ ). The idea is, to use Sourceforge.net as a portal and to redirect to http://webtestclipse.canoo.com/ where the project is hosted and developed...

Read more at sourceforge.net/project...