Did you know ... Search Documentation:
Predicate load_html/3
 load_html(+Input, -DOM, +Options) is det
Load HTML text from Input and unify the resulting DOM structure with DOM. Options are passed to load_structure/3, after adding the following default options:
dtd(DTD)
Pass the DTD for HTML as obtained using dtd(html, DTD).
dialect(Dialect)
Current dialect from the Prolog flag html_dialect
max_errors(-1)
syntax_errors(quiet)
Most HTML encountered in the wild contains errors. Even in the context of errors, the resulting DOM term is often a reasonable guess at the intent of the author.

You may also want to use the library(http/http_open) to support loading from HTTP and HTTPS URLs. For example:

:- use_module(library(http/http_open)).
:- use_module(library(sgml)).

load_html_url(URL, DOM) :-
    load_html(URL, DOM, []).
Tags are associated to your profile if you are logged in
Tags:
Jan Wielemaker said (2014-10-20T12:26:59):0 upvotes 0 0 downvotes
Picture of user Jan Wielemaker.

HTML is often illegal. The SWI-Prolog parser will emit warnings, but often creates the intended parse tree. Typically, add the following options for processing arbitrary HTML if you are not interested in (minor) syntactical problems with the input.

  • syntax_errors(quiet)
  • max_errors(-1)