refactor: simplify Value builtin cost models with improved documentation

Simplify the R model definitions for Value-related builtins by replacing
custom linear model implementation with standard linearInY wrapper for
valueContains. This maintains the same statistical behavior while
improving code maintainability.

Add inline comments documenting the parameter wrapping strategy used
for each model (Logarithmic wrapping for lookupCoin/valueContains,
ValueTotalSize for contains operand, unwrapped for valueData/unValueData).

Clean up formatting inconsistencies in model definitions.

Author: Unisay

plutus-core/cost-model/data/models.R

@@ -806,29 +806,25 @@ modelFun <- function(path) {
         mk.result(m, "exp_mod_cost")
     }
 
-    dropListModel   <- linearInX     ("DropList")
+    dropListModel <- linearInX ("DropList")
 
     ## Arrays
     lengthOfArrayModel        <- constantModel ("LengthOfArray")
     listToArrayModel          <- linearInX ("ListToArray")
     indexArrayModel           <- constantModel ("IndexArray")
 
     ## Values
-    lookupCoinModel <- linearInZ ("LookupCoin")
 
-    ## ValueContains is O(n₂ × log max(m₁, k₁)) where n₂ is the total size of the second Value
-    ## We model this as linear in the sum of sizes, which is conservative
-    valueContainsModel <- {
-        fname <- "ValueContains"
-        filtered <- data %>%
-            filter.and.check.nonempty(fname) %>%
-            discard.upper.outliers()
-        m <- lm(t ~ I(x_mem + y_mem), filtered)
-        mk.result(m, "added_sizes")
-    }
+    # Z wrapped with `Logarithmic . ValueOuterOrMaxInner`
+    lookupCoinModel           <- linearInZ ("LookupCoin")    
+
+    # X wrapped with `Logarithmic . ValueOuterOrMaxInner`
+    # Y wrapped with `ValueTotalSize`
+    valueContainsModel        <- linearInY("ValueContains")         
 
-    valueDataModel <- constantModel ("ValueData")
-    unValueDataModel <- linearInX ("UnValueData")
+    # Sizes of parameters are used as is (unwrapped):
+    valueDataModel            <- constantModel ("ValueData")
+    unValueDataModel          <- linearInX ("UnValueData")
 
     ##### Models to be returned to Haskell #####