[ngrams] fix for grouping function for ngrams

For nearby terms `["a", "b", "c"]` we concatenated them into one
`"a b c"`. Instead we should leave the singletons as well:
`["a", "b", "b", "a b", "a c", "b c", "a b c"]`.

src/Gargantext/Core/Text/Terms/Multi/Group.hs

@@ -23,7 +23,7 @@ import Gargantext.Prelude
 group2 :: POS -> POS -> [TokenTag] -> [TokenTag]
 group2 p1 p2 (x@(TokenTag _ _ (Just p1') _):y@(TokenTag _ _ (Just p2') _):z) =
   if (p1 == p1') && (p2 == p2')
-     then group2 p1 p2 (x<>y : z)
+     then (x : y : group2 p1 p2 (x<>y : z))
      else (x : group2 p1 p2 (y:z))
 group2 p1 p2 (x@(TokenTag _ _ Nothing _):y) = (x: group2 p1 p2 y)
 group2 _ _ [x@(TokenTag _ _ (Just _) _)] = [x]

src/Gargantext/Core/Types.hs

@@ -59,7 +59,7 @@ type Label = [Text]
 
 data Terms = Terms { _terms_label :: Label
                    , _terms_stem  :: Stems
-                   } deriving (Ord)
+                   } deriving (Ord, Show)
 
 instance Eq Terms where
   (==) (Terms _ s1) (Terms _ s2) = s1 == s2