Update bot Harry's use of Wikipedia

Specify utf8 API parameter to avoid ascii escaping (\u00E9)
Can now handle more inputs such as "what is Athenian democracy?"
Fix double url encoding after jsonopen update in 7.4

Author: Andy Bruère

RAWDATA/HARRY/keywordless.top

@@ -1,4 +1,4 @@
-# this is the topic for anything that has no obvious topic. 
+# this is the topic for anything that has no obvious topic.
 # it is called directly from the control script when everything else fails.
 
 topic: ~keywordless nostay()
@@ -12,30 +12,32 @@ u: (< say _*) keep() repeat() '_0
 #! yes, if user just says OK or yes or continue, then move along and give control to bot
 u: ( < [~yesanswer ~more]>) ^keep() gambit($$currenttopic) # current topic gets to continue
 
-?: (what is a _*1 >) keep() $$tmp = _0
-	 $$url = ^"\"https://en.wikipedia.org/w/api.php?action=query&prop=extracts&exsentences=2&exintro=&explaintext=&format=json&titles=$$tmp\""
-	 $$user_agent = ^"[email protected] User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)"
-	 $$tmp = ^jsonopen(GET $$url "" $$user_agent)
+#! What is democracy?
+#! What is climate change?
+?: (what be {a} _*~2 >) ^keep() $$tmp = '_0
+	$$url = ^"\"https://en.wikipedia.org/w/api.php?action=query&prop=extracts&exsentences=2&exintro=&explaintext=&format=json&utf8=&redirects=&titles=$$tmp\""
+	# Url escaping is done by ^jsonopen
+	$$user_agent = ^"[email protected] User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)"
+	$$tmp = ^jsonopen(GET $$url "" $$user_agent)
 		# ^jsontree($$tmp)
 		# ^jsonpath(  ".query .pages .56333 .extract" $$tmp)
-		@0 = query(direct_v ? extract ? 1)
+		@0 = ^query(direct_v ? extract ? 1)
 		@0object
-			
-?: (who be _*1  >) keep()  $$tmp = substitute(character '_0 _ %20 )
-	 $$url = ^"\"https://en.wikipedia.org/w/api.php?action=query&prop=extracts&exsentences=2&exintro=&explaintext=&format=json&titles=$$tmp\""
-	 $$user_agent = ^"[email protected] User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)"
-	 $$tmp = ^jsonopen(GET $$url "" $$user_agent)
-		# ^jsontree($$tmp)
-		# ^jsonpath(  ".query .pages .56333 .extract" $$tmp)
-		@0 = query(direct_v ? extract ? 1)
-		@0object
-	
-?: (who be _*1 _*1 >) keep()  $$tmp = join('_0  %20  '_1)
-	 $$url = ^"\"https://en.wikipedia.org/w/api.php?action=query&prop=extracts&exsentences=2&exintro=&explaintext=&format=json&titles=$$tmp\""
-	 $$user_agent = ^"[email protected] User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)"
-	 $$tmp = ^jsonopen(GET $$url "" $$user_agent)
+
+#! Who is Gandhi
+#! Who is Lyndon Johnson # merged to Lyndon_Jonhnson
+#! Who is Athenian democracy # testing two non merged words
+?: (who be _*~2 >) ^keep()
+	$$tmp = ^pos(capitalize ^original(_0))
+	##<<
+		Using original instead of '_0 to skip Wordnet dictionnary, since Wikipedia has more searchable content.
+		and correct spelling is left to the user.  In return this avoids excessive spellcheck on proper nouns.
+		Wikipedia does not systematically redirect mis-cased requests. Proper names should always be capitalized
+	##>>
+	$$url = ^"\"https://en.wikipedia.org/w/api.php?action=query&prop=extracts&exsentences=2&exintro=&explaintext=&format=json&utf8=&redirects=&titles=$$tmp\""
+	$$user_agent = ^"[email protected] User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)"
+	$$tmp = jsonopen(GET $$url "" $$user_agent)
 		# ^jsontree($$tmp)
 		# ^jsonpath(  ".query .pages .56333 .extract" $$tmp)
-		@0 = query(direct_v ? extract ? 1)
+		@0 = ^query(direct_v ? extract ? 1)
 		@0object
-