Awesome
About SKDF
This directory contains a skeleton Keyword-Driven Framework project based on ashokkhape/automation_framework_selenium and ashokkhape/automation_framework_selenium and selenium-webdriver-software-testing/keyword-driven-framework
The project builds two runnable jars (com.github.sergueik.jprotractor
and com.github.sergueik.ngwebdriver
group id namespaces):
cp TestCase.xls ~/Desktop
pushd jprotractor
mvn -Pdevelop -DskipTests -Dmaven.test.skip=true install
REM java -jar target/skdf_jprotractor-develop.jar
REM alternatively
java -cp target\skdf_jprotractor-develop.jar;target\lib\* com.github.sergueik.jprotractor.Launcher
popd
pushd ngwebdriver
mvn clean install
REM java -jar target/skdf_ngwebdriver-develop.jar
java -cp target\skdf_ngwebdriver-develop.jar;target\lib\* com.github.sergueik.ngwebdriver.Launcher
popd
The launcher uses reflection to associate keywords with class methods
private static Map<String, String> methodTable = new HashMap<>();
static {
methodTable.put("CLICK", "clickButton");
methodTable.put("CLICK_BUTTON", "clickButton");
...
- a single method may have several keywords pointing to it;
String methodName = methodTable.get(keyword);
try {
Class<?> _class = Class.forName("com.github.sergueik.ngwebdriver.KeywordLibrary");
Method _method = _class.getMethod(methodName, Map.class);
System.out.println(keyword + " call method: " + methodName + " with "
+ String.join(",", params.values()));
// for static methods
_method.invoke(null, params);
// or when using instances methods
Object _object = _class.newInstance();
_method.invoke(_object, params);
Similar approach is used to define Selector strategies, most of which are static methods of the appropriate core Selenium class:
locatorTable.put("className",
By.class.getMethod("className", String.class));
locatorTable.put("css", By.class.getMethod("cssSelector", String.class));
locatorTable.put("id", By.class.getMethod("id", String.class));
locatorTable.put("linkText",
By.class.getMethod("linkText", String.class));
locatorTable.put("name", By.class.getMethod("name", String.class));
locatorTable.put("tagName", By.class.getMethod("tagName", String.class));
locatorTable.put("xpath", By.class.getMethod("xpath", String.class));
with few "synthetic" strategies, notably the TEXT:
// put synthetic selectors explicitly
locatorTable.put("text", methodMissing);
implemented e.g. through xpath
:
String amendedSelectorValue = String.format(
"//%s[contains(normalize-space(text()),'%s')]",
(selectorTagName != null) ? selectorTagName : "*", selectorValue);
_element = _wait.until(new ExpectedCondition<WebElement>() {
@Override
public WebElement apply(WebDriver d) {
return d.findElements(By.xpath(amendedSelectorValue)).stream()
.filter(o -> {
return (Boolean) (o.getText().contains(selectorValue));
}).findFirst().get();
}
});
The test step arguments are passed as hash of parameters. That is done so one does not care about the method signature.
Also the AngularJS introduced NgBy
locators which fequently require (multiple) additional arguments like e.g.
@FindBy(how = How.REPEATER_COLUMN, using = "row in rows", column = "name")
private List<WebElement> friendNames;
The step status is returned via params["status"]
entry, the step result (if any) is returned via params["result"]
Adding Tests to the Spreadsheet
To add a test case, put its name into the Index
sheet and mark it with Yes
to be executed
Next, add the steps. Making the cell border visible will ensure the blank cells are not getting skipped:
Adding jProtractor
One can explore additional selectors with jProtractor.
jProtractor is not available in maven central, therefore to use it with framework one needs do build it from source and
install it into current user's .m2
repo:
git clone https://github.com/sergueik/jProtractor.git
pushd jProtractor
mvn -Dmaven.test.skip=true clean install
which will install the jar:
[INFO] Installing C:\developer\sergueik\selenium_java\protractor\target\jprotractor-1.2-SNAPSHOT.jar to C:\Users\Serguei\.m2\repository\com\jprotractor\jprotractor\1.2-SNAPSHOT\jprotractor-1.2-SNAPSHOT.jar
[INFO] Installing C:\developer\sergueik\selenium_java\protractor\pom.xml to C:\Users\Serguei\.m2\repository\com\jprotractor\jprotractor\1.2-SNAPSHOT\jprotractor-1.2-SNAPSHOT.pom
This will add about 10 AngularJS-specific NgBy
locators:
options
input
selectedOption
repeater
model
binding
repeaterElement
repeaterRows
buttonText
repeaterColumn
cssContainingText
selectedRepeaterOption
partialButtonText
Adding ngWebDriver
Another implementation of Protractor selectors is ngWebDriver.
Most Protractor-specific locators are the same (the class is ByAngular
):
options
repeater
model
binding
buttonText
cssContainingText
partialButtonText
Few method signatures of are different:
repeaterCell
repeaterRow
repeaterColumn
The implementation of the correspondent keyword methods is through chaining the ngWebDriver methods e.g.
case "repeaterColumn":
ngDriver.waitForAngularRequestsToFinish();
ByAngularRepeater _elementRepeater = ByAngular.repeater(selectorValue);
ByAngularRepeaterColumn _elementRepeaterColumn = _elementRepeater.column(selectorColumn);
_element = driver.findElement(_elementRepeaterColumn);
return _element;
Some methods are unique to jProtracror, they do not exist in ngWebDriver:
selectedOption
selectedRepeaterOption
In addition the "text" keyword is recognized with the usual implementation through xpath
locator:
String.format("//%s[contains(normalize-space(text()), '%s')]", (selectorTagName != null) ? selectorTagName : "*", selectorValue);
alternatively if the selectorTagName
is provided, the Java 8 stream based impplementation is possible:
_element = driver.findElements(By.tagName(selectorTagName)).stream()
.filter(o -> o.getText().contains(selectorValue)).findFirst().get();
Introduction
The original Keyword Driven Framework suggests identifying few columns (not necessarily exactly four, though).
- Summary: brief description of the step
- Target: name of the Web Page object/element, like "Link" or "Input"
- Action: name of the action, which will be performed on Target Element such as click, open browser, input text etc.
- Data: any value which is needed by the Object to perform any action, like text value for input field.
It is very likely inspired by
Selenium IDE Firefox Add-On Command, Target, Value columns, with significanylty narrowed choice of commands (Keywords). The advantages from taking such approach are discussed many times:
- Less Technical Expertise: manual testers or non technical testers can easily write test scripts for automation using the Framework than code straight.
- Easy To Understand: With no coding is exposed, the test flow is easy to read and understand. Keywords & actions are descriptive.
- Early Start: One can start building Keyword Driven test cases immediately deferring more challenging tasks like Page Object model to a later stage. Keyword steps are quick to identify from requirements documentation or manual test.
- Re-usability: With implementing modularization in Keyword Driven, Re-usability can be further increased. Equipped with a stable and powerful Execution Engine in Keyword Driven Framework, it encourage extreme code re-usability.
- Automation: Excel file is (a lot) easier to produce by a recording tool than a full blown program.
Work in Progress
Adding
- ahajamit/chrome-devtools-webdriver-integration Selenium 3.x compatible Chrome Devtool Selenium extension project that exposes whole set of Chrome DevTools Protocol automation API to Selenium in similar fashion Selenium 4 is doing (Selenum 4 is currently in alpha) as cdp_integration
- sukgu/shadow-automation-selenium shadow ROOT DOM automation javascript API wrapper project, which offrs custom API as shadow_automation
No keywords defined for any of those yet.
Note
the urls http://www.seleniumeasy.com/test/basic-first-form-demo.html
and https://www.seleniumeasy.com/test/input-form-demo.html
no longer exist on http://www.seleniumeasy.com
though the former is still referenced in documentation for Python testers.
These pages were apparently mirrored on https://demo.anhtester.com
The directory http://www.seleniumeasy.com/test/
is also gone and there is no mirror
The update of test material is a work in progress. The sheets KeywordFramework
are curently disabled.
The Alert
sheet is updated and enabled (it had a typo in the custom method name for processing the alert popup)
See Also
- qaf
- NPOI
- ExcelDataReader (.net)
- LinqToExcel
- save-selenium-webdriver-testng-result-excel
- DataProvider - Data Driven Testing with Selenium and TestNG
- Excel Java poi helpers
- Fast Excel poi-clean Java implementation
License
This project is licensed under the terms of the MIT license.