template_fpi.txt

FPI Template design document:


<------------------Principles of FPI creation (TODO later):------------>

=>give examples of good case and bad case"
=>give reasons why exploiting content using contains: etc.
=>give examples of effectively using the child div. 
		--> look at delta example. 
	

<-----------------------------ATTRIBUTES-------------------------------->
1. Attribute "type" for <div> indicates if that div is scalar or vector.
   By default, every <div> is scalar.
   For fields that could have multiple values such as addresses, ccns etc,
   create a parent <div> element that has a selector that will collect
   all these element and with a - type='vector'. 
   e.g.
   <div name="addresses" type="vector">
   	<action type="fetch-dom-element">body table.address
	</action>
	
	<div name="address">
	     <action type="store">td:eq(2)
	     </action>	     
	</div>
   </div>

   
			<--------- <div> attributes ------------->   
2. Attribute "can_be_a_null" for a <div> tells if that field can be 
   empty. Each FPI should have at least one  - can_be_a_null='no'
   to indicate whether PI download was successful or not.
   Usually these fields are Name and/or Email.

   In case of vector elements, put this attribute (can_be_a_null='no')
   in the parent <div> element where you mention (type="vector") 
   (See point 1).
     

		<--------- <action "type" attributes ------------->   
3. Currently supported type-attributes for any <action> are:
   fetch-url: Fetches the actual URL. Store the resulting HTML page
   	      as a DOM document as current node's value.
   fetch-href: Applies the CSS selector in <action>'s text to
   	       the parent node's value. It reads the href of
	       resulting element.
	       Then fetches that href. Store the resulting HTML page
	       as a DOM document as current node's value.
   fetch-dom-element: Applies the CSS selector in <action>'s text to
   	       	      the parent node's value. Store the resulting 
		      element(s) as current node's value.
   simulate-click:  Applies the CSS selector in <action>'s text to
   	       	    the parent node's value. Simulate a mouse-click
		    on resulting element(s). Then read the HTML page.
		    Store that HTML page as DOM document in current
		    node's value.
   combine-n-store: Applies the CSS selector in <action>'s text to
   	       	    the parent node's value. Assuming that there are
		    multiple elements generated, it combines the 
		    text contained in those elements with a comma.
		    This is particularly useful for addresses.
		    One needs to combine Address line 1, line 2,
		    city, state, zip and country.
   store: This action can only be associated with a leaf node.
   	  It apply the CSS selector in <action>'s text to parent
	  node's value. Store the resulting element's text
	  as final value of the PI field. The PI field is 
	  mentioned in the "name" attribute of leaf <div>
   	  One important thing is the result of applying 
	  the CSS selector + filter for this leaf node
	  must be an HTML element which can be passed to
	  jQuery and can be applied a .text() method on.
	  (i.e. it should not be a string like "xyz").
	  
	  <--------- <action> attributes ------------->   
   
	combine_operator : current default is " ," but with this attribute this
		can be changed to anything such as "-", "" , etc. This useful for 
		feids such as phonenumber and/or ccn. Where the comma is not effective.