This answer is purely about inventive step in general:
Let's take claim 1 from the question as an example: (please note: I didn't read the specification so this is purely based on the wording of the claim, it might not concur with the correct interpretation of the actual patent claim which needs to be interpreted in light of the specification!)
- A method for on- line filtering a presentation comprising:
a) identifying a structure in the presentation;
b) detecting an undesired content in said structure;
c) determining a domination of said structure by said undesired content, and
d) disabling a portion of said structure according to a result of said determining.
Now, suppose you have as prior art the html sanitization you mentioned in a comment. According to my understanding, html sanitization does the following:
So, what does that mean for claim 1 above?
We seem to fulfill a) by identifying a structure starting and ending with html tags.
We also seem to fulfill d) by removing the structure completly because "completly" is also "a portion". However, completly is only a small subset of "a portion", so there might be an easy way of getting around some prior art like that by amending the claim to say "a portion less than 100%" or something like that. It might even be that it says "a portion is less than 100% of the structure" in the description, invalidating this prior art.
But suppose we actually fulfill steps a) and d) completly. Then the prior art "html sanitization" is not destroying novelty because it does not teach all the elements from the claim (namely: b) and c) are missing).
So, inventive step. We need another document that allows the person skilled in the art to reach the invention.
I'll make one up. Suppose we have a publication about identifying ads in html code. It identifies html structs like
<embed> </embed>, searches the content for words like viagra (greetings to the spam detecting bots at SE ;)) and determines if the structure is mainly filled with said undesired content by, say, language analysis. It then determines that the structure is an ad.
This reference teaches a), b) and c). It identifies a structure (html tags), detects an undesired content ("viagra") and determines domination of the structure by said content (checks if the text in the structure relates mainly to viagra).
Combining the references the person skilled in the art could see the advantage of removing html tags with a greater precision by detecting if they contain ads. The person skilled in the art could then reach the invention by combining the two documents, meaning the claim is not inventive over the prior art.
Disclaimer: One could argue against that by saying that html sanitization and ad blocking don't serve the same goal because removing all possibly problematic structures is more secure than removing only some so the person skilled in the art wouldn't combine the references. That argument might or might not be enough to prove inventive step, but it would at least pose some danger to the argumentation. So this isn't a clear cut case.
The above used goes d'accord with pretty much all standard methods for determing (missing) inventive step in the US, Europe and many other places. You need two combinable documents teaching every single step in a method or every single limitation in a system.
You can also use basic knowledge of a person skilled in the art (e.g. web developer), but you need to prove that knowledge. So if you have a textbook saying "Identified ads should be removed from a html site for user convenience." (step d)) you might use that to say my made-up source (steps a) - c)) in combination with the knowledge of the person skilled in the art leads to the claimed method making it obvious / non inventive.