The ¡°new.irb¡± robot we have created can do
literature search and Web navigations automatically. Our purpose, however, is to extract the
literature information and save them into a database. We will achieve this by extracting Web
information during robot navigation, defining variables to store the
information, and saving the variable data into local databases. We will define
some variables to store the abstract author, title and text in this section.
From the ¡°pubmed¡±
task created previously (shown below) we will define four variables: AbstractURL, Title, Author, and Abstract from
the 3, 4, 5 and 6th actions, respectively.

Right-click the forth action: ¡°Extract data like ¡®Analysis of ***¡¯¡±,
and select the ¡°Name Variable ¡¡± menu item, as below.

It shows a dialog box as the following. Give it a name ¡°Title¡±, select a data
transformation option: ¡°Exclude all tags¡±, and press [OK].

Similarly, define Author and
Abstract for the 5 and 6th action. The ¡°pubmed¡±
task becomes:

Finally, we are going to define the AbstractURL
from the 3rd action, which is more complex. Bring up a result page
including a list of entries from the pubmed search in
the browser panel like:

Then right-click on the 3rd action ¡°Get a list of links like
***¡±, and select the ¡°Name Variables ¡¡± menu item. It shows:
As you can see, the search result page serves as an example page for
data columns. Put a name ¡°AbstractURL¡± after the ¡°COLUMN3¡±, select a transformation
¡°Absolute url of hyperlink¡± after it, and click the [Update]. It shows:

Notice that URLs are shown in the sample table in the lower half of the
page. Now click [Return]. We have finished defining the four
variables.