Skip to main content

Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.

Jump to: navigation, search

Scout/Tutorial/Reorganize the tree and add a webservice

< Scout‎ | Tutorial
Revision as of 08:52, 19 October 2010 by Asc.bsiag.com (Talk | contribs) (Make it available to Scout)

Note.png
Scout Tutorial
This page belongs to the The Scout documentation has been moved to https://eclipsescout.github.io/.. It explains how reorganize the little application consisting of two pages such that you can add a webservice. You need that little The Scout documentation has been moved to https://eclipsescout.github.io/. in order to continue.


We started out with an application having two table pages. Notice how the list of persons comes just beneath the list of companies. If we want to show different named folders such as "persons" and "search resuilts" for a company, we need to add a page with nodes.

 Standard Outline
  │
  ├─Company Table Page
  │  │
  │  └─Person Table Page
  │
  └─Person Table Page
 Standard Outline
  │
  ├─Company Table Page
  │  │
  │  └─Company Details Node Page ← new
  │     │
  │     ├─Person Table Page
  │     │
  │     └─Bing Table Page ← new
  │
  └─Person Table Page

Add a node page

In a first step, we're going to aim for the following intermediary structure:

 Standard Outline
  │
  ├─Company Table Page
  │  │
  │  └─Company Details Node Page ← new
  │     │
  │     └─Person Table Page
  │
  └─Person Table Page

Return to the CompanyTablePage and click on the red cross () next to Exec Create Child Page in order to delete it.

Delete Create Child Page Method in Scout Minicrm.png

Once you have deleted the method, you can right-click on the Child Page folder and pick New Page...

Use AbstractTableWithNodes as your template. There's no need to give it a name. Use CompanyDetailsNodePage as the type name.

Note.png
No name?
Indeed, this particular node page does not need a name. That's because the name of a page is only shown if its parent is an outline or a page with nodes. If the parent is a page with table, the selected row replaces the name of the child table.


Now go to the newly created CompanyDetailsNodePage, click through to the Child Pages folder, right-click and pick Add Page... Pick the PersonTablePage from the list and click Finish.

Note.png
Child Page or Child Pages?
Note how there is only child page for a page with table where as there are multiple child pages for a page with nodes.


Fix data flow

If you attempt to test your application, you'll notice a problem: Every person is listed under every company!

Why is that?

We interrupted "the flow data": When the user picks a company from the CompanyTablePage, the appropriate child page is created. When we created The Scout documentation has been moved to https://eclipsescout.github.io/., we made sure to pass the value of the CompanyNrColumn along. The newly introduced CompanyDetailsNodePage needs to be fixed!

Return to the CompanyDetailsNodePage and click through to Variables. Pick Create New Property Bean... from the context menu, use companyNr as the name and Long as the bean type.

Click on the Exec Create Child Pages link on the Properties view of the CompanyDetailsNodePage. Change the code as follows:

@Override
protected void execCreateChildPages(Collection<IPage> pageList) throws ProcessingException {
PersonTablePage personTablePage = new PersonTablePage();
  personTablePage.setCompanyNr(getCompanyNr());
  pageList.add(personTablePage);
 
}

Return to the CompanyTablePage and click on the Exec Create Child Page link on the Properties view. Change the code as follows:

@Override
protected IPage execCreateChildPage(ITableRow row) throws ProcessingException {
  CompanyDetailsNodePage childPage=new CompanyDetailsNodePage();
  childPage.setCompanyNr(getTable().getCompanyNrColumn().getValue(row));
  return childPage;
}

We reached our goal! This is the new structure, now:

 Standard Outline
  │
  ├─Company Table Page
  │  │
  │  └─Company Details Node Page
  │     │
  │     └─Person Table Page
  │
  └─Person Table Page

Let's review how the data flows back and forth:

  1. CompanyTablePage calls execLoadTableData
  2. execLoadTableData calls the getCompanyTableData method on the StandardOutlineService
  3. StandardOutlineService returns tabular data including the primary key for every row
  4. user picks a company and clicks through
  5. the CompanyDetailsNodePage is created by execCreateChildPage; the value of the current CompanyNrColumn is copied to the node page's companyNr variable
  6. the user picks Persons and clicks through
  7. the PersonTablePage is created by execCreateChildPages; the value of the companyNr variable is copied to the table page's companyNr variable
  8. PersonTablePage calls execLoadTableData
  9. execLoadTableData calls the getPersonTableData method on the StandardOutlineService
  10. StandardOutlineService determines that the companyNr is not null, runs a SELECT statement that only selects appropriate persons from the database and returns tabular data

Create Axis web service consumers

We need to create a web service consumer. We will use Bing as an example web service. See their developers page for more information.

On the server side, right-click on Axis Web Service Consumers and pick New Webservice consumer… from the context menu. Paste http://api.search.live.net/search.wsdl?AppID=F1DFAE1859C73D8B105D99A78F0B027B62C64814 into the WSDL URL field and click OK.

Accept the import of all the generated classes.

Warning2.png
Troubleshooting
If we can access the WSDL in the browser, but receive a //Connection timed out// exception using the Scout SDK, we need to save the WSDL file locally and provide the generator with a file URL to the WSDL (e.g. file:///C:/Users/bornn/Desktop/search.wsdl). This problem arise because Eclipse does not necessarily share the browser's proxy settings.


Later, when deploying the application, we need to provide the proxy settings in the server's config.ini file:

 -Dhttp.proxyHost=proxy.example.com
 -Dhttp.proxyPort=3128
 -Dhttp.nonProxyHosts="localhost"
 -Dhttp.proxyUser=johndoe
 -Dhttp.proxyPassword=*****

Make it available to Scout

We need to add a method to the StandardOutlineService which will use the classes we just created.

Return to the StandardOutlineService and pick New Service Operation... from the context menu. Use getLiveSearchTableData as the name, Object[][] as the return type, query as the first argument's name, and String as it's type.

Scout New Service Operation.png

Use the following implementation:

public Object[][] getLiveSearchTableData(String query) throws ProcessingException {
  // use the ID from the URL
  String appId = "F1DFAE1859C73D8B105D99A78F0B027B62C64814";
  LiveSearchServiceLocator soapClient = new LiveSearchServiceLocator();
  SearchRequest request = new SearchRequest();
  request.setAppId(appId);
  request.setQuery(query);
  SourceType[] sources = {SourceType.Web};
  request.setSources(sources);
  SearchResponse response;
  try {
    //no idea why the indirection via SearchRequestType1
    //and SearchResponseType0 is necessary
    response = soapClient.getLiveSearchPort().search(new SearchRequestType1(request)).getParameters();
  }
  catch (Exception e) {
    throw new ProcessingException(e.getMessage(), e);
  }
  WebResult[] result = response.getWeb().getResults();
  Object[][] data = new Object[result.length][3];
  for (int row = 0; row < data.length; row++) {
    data[row][0] = result[row].getUrl();
    data[row][1] = result[row].getTitle();
    data[row][2] = result[row].getDescription();
  }
  return data;
}

Back to the top