Add structure detection of news objects

[ivbeg]

Add structure detection and xpath reconstruction. Instead of dynamic news detection build pseudo-code to extract news from the page.

It should implement analysis logic that should detect:

• news list block container
• the type of news list: sub-blocks or mixed list
• headline tag
• text tag/tag-block
• date tag is exists
• links is exists
• images if exists

Is something missing?