Graphical Creation of New Tagsets¶
From the main window of the ITPilot Wrapper Generation Tool, click on the “Tools” menu of the menu bar, and then click on the “Scanner & TagSet configuration” submenu option. A new window will be open, such as the one shown in Scanner and Tag Set Generation Tool.
This tool is divided into three vertical areas, where each one contains information on the scanners, tagsets and specific tags that currently exist in the ITPilot installation you are working with.
In the upper part of the window, the user may specify if the complete set of scanners and tagsets are shown, if only those generated manually will be shown, or if only the ones automatically generated in the web extraction process will be shown (see section Configuration of the Extractor Component). In case of selecting the automatically generated, a “Remove Unused Scanners” button will be available: this button will remove all the scanners that are not being used in a wrapper.
By selecting a scanner with the mouse you will be able to see their internal characteristics: “lexer” type (see section Lexer Types) and, most importantly, the tagsets included. The central area shows all the existing tagsets and the tags in each one. Lastly, the right-hand area indicates the available tags. There is a tag defined for every HTML tag.
To illustrate the creation of new tagsets, we will create a new set used to extract data from table cells which values are text fragments, which may contain new lines in the middle (by using the <BR /> or the <P> tags) and links. These values cannot be extracted with the default tagset all4_6, since it assumes that <BR />, <P> and <A> tags act as data separators. The new tagset will only contain the HTML tags used to define tables: <TABLE>, </TABLE>, <TR>, </TR>, <TD> abd </TD>.
Therefore, create a new tagset called “myTextTagSet”, by clicking the button in the central “TagSets” area. Once is created, edit it by selecting it in the list of tagsets and click the button. The arrows between the “TagSets” and “Tags” areas will became enabled; you can then select any tags to be included in the tagset and click the . For this example, the required tags are TABLE, ENDTABLE, TR, ENDTR, TD and ENDTD.
The “TagSets” area provides also the option “HTML tags not removed from texts” which is used to describe which HTML elements from the ones that do not match the elements used in the set will not be deleted from the processed text; this is useful when it is required that certain HTML structure is kept on the extracted data (for instance, in order to show them directly in some kind of web application or portlet). In order to use this option, a regular expression must be written. This expression must be an OR (“|”) of the expressions which define each of the desired tags (they can be consulted in “Tag Value” text area). Note that, if the tags used in the regular expression are included in the tagset they will not be returned as tags. For example if you want to return the P and ENDP tags as text, then the regular expression should be the following one: “<P”[\n\r\t][^\>]*“>”|“<P>”|“</P”[\n\r\t][^\>]*“>”]|“</P>”, and then P and ENDP will not be returned as tags although they are included in the tagset.
In our example we will not use this option.
To complete the stage, save the tagset by clicking on . If you want to cancel the changes made to the tagset and return to the last saved version, simply click on the “Revert to Saved” button. Note that it is also possible to create a new tagset by copying an existing one. To do this click the button in the central “TagSets” area, check the “Copy existing TagSet” option in the “New TagSet” dialog and then select the desired tagset before clicking Ok.
Lastly, create a new scanner and link it to the recently created tagset. The operation is similar to the step indicated above, clicking on in the left-hand part of the scanner generation window and creating a new scanner, “myLexer”. Then, with the recently created scanner marked, add the “myTextTagSet” tagset to the scanner by clicking on the button of the scanner, selecting the tagset and then clicking on the arrow between both areas to include it in the scanner. The scanner generation window will be similar in appearance to that in Creating a new scanner, where the “Included TagSets” field of the scanner area displays the “myTextTagSet” tagset.
The last step of this process involves saving the scanner so that it can be used by any ITPilot application. To do so, simply click on in the scanner area. Besides, it will be necessary to install the scanner in the Wrapper Server when a wrapper using it is deployed (use the appropriate option in the deploy wrapper dialog). Please see section Deploying Wrappers to the Wrapper Server for more information about how to do it.