From 72cb3c2157c6c95f282acf1db2db7a135132bce8 Mon Sep 17 00:00:00 2001
From: Pasquale Pinto <pasquale@driebit.nl>
Date: Fri, 11 Oct 2024 11:22:35 +0200
Subject: [PATCH] CVN-141 Add ability to replace html from values

---
 .../mod_ginger_rdf/models/m_rdf_triple.erl    | 37 ++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

diff --git a/modules/mod_ginger_rdf/models/m_rdf_triple.erl b/modules/mod_ginger_rdf/models/m_rdf_triple.erl
index b99ba957..72259b6e 100644
--- a/modules/mod_ginger_rdf/models/m_rdf_triple.erl
+++ b/modules/mod_ginger_rdf/models/m_rdf_triple.erl
@@ -9,7 +9,9 @@
     m_to_list/2,
     m_value/2,
 
-    remove_empty/1
+    remove_empty/1,
+    replace_html/1,
+    replace_html/2
 ]).
 
 -include_lib("zotonic.hrl").
@@ -167,3 +169,36 @@ is_empty_object(#rdf_value{value = Value}) ->
     is_empty_object(Value);
 is_empty_object(_) ->
     false.
+
+% Like 'replace_html/2' with 'z_html:strip/1' being the chosen modifier.
+replace_html(In) -> replace_html(In, fun z_html:strip/1).
+
+% Replaces all html values with the given modifier function, recursively, from
+% an 'rdf_value', an 'rdf_resource', a 'triple' or a list of either;
+% returns any other value unaltered.
+% Note: the type of the output is always the same as the input's.
+replace_html(List, ReplaceFun) when is_list(List) ->
+    lists:map(fun(Item) -> replace_html(Item, ReplaceFun) end, List);
+replace_html(Triple, ReplaceFun) when is_record(Triple, triple) ->
+    NewObject = replace_html(Triple#triple.object, ReplaceFun),
+    Triple#triple{object=NewObject};
+replace_html(Resource, ReplaceFun) when is_record(Resource, rdf_resource) ->
+    NewTriples = replace_html(Resource#rdf_resource.triples, ReplaceFun),
+    Resource#rdf_resource{triples=NewTriples};
+replace_html(Value, ReplaceFun) when is_record(Value, rdf_value) ->
+    ValueTerm = Value#rdf_value.value,
+    NewValueTerm =case is_html_value(ValueTerm) of
+        true -> ReplaceFun(ValueTerm);
+        false -> replace_html(ValueTerm, ReplaceFun)
+    end,
+    Value#rdf_value{value=NewValueTerm};
+replace_html(Other, _ReplaceFun) ->
+    Other.
+
+% Returns 'true' iff the input is a binary string of html, 'false' otherwise.
+is_html_value(Term) when is_binary(Term) ->
+    case z_html_parse:parse(Term) of
+        {ok, _HtmlNode} -> true;
+        {error, nohtml} -> false
+    end;
+is_html_value(_) -> false.