use larger set of words in words.txt dataset

README.md

@@ -78,4 +78,5 @@ The reason for implementing a new trie specialised to strings rather than using
 
 # Credits
 
-The tests in `tests/string-set-tests.sml` were ported from [kpol's Trie data structure in C3](https://github.com/kpol/trie), although this is not true for the files in the `src/` directory.
+- The tests in `tests/string-set-tests.sml` were ported from [kpol's Trie data structure in C3](https://github.com/kpol/trie).
+- The words.txt dataset in `bench/words.txt` is from [this repository](https://github.com/dwyl/english-words).

bench/conv-words.sml

@@ -1,3 +1,5 @@
+(* generate a words.sml file with a vector of strings, 
+ * from a line-delimited words.txt file *)
 val inIo = TextIO.openIn "words.txt"
 val outIO = TextIO.openOut "words.sml"
 
@@ -11,11 +13,14 @@ fun writeLines (outIO, lst) =
     [] => ()
   | word :: tl =>
       let
-        val word = String.substring (word, 0, String.size word - 2)
+        (* remove \r and \n from the word *)
-        val isLast = tl = []
+        val word = Substring.full word
         val word =
-          if isLast then "\"" ^ word ^ "\""
+          Substring.dropr (fn chr => chr = #"\n" orelse chr = #"\r") word
-          else "\"" ^ word ^ "\",\n"
+        val word = Substring.string word
+
+        val isLast = tl = []
+        val word = if isLast then "\"" ^ word ^ "\"" else "\"" ^ word ^ "\",\n"
         val _ = TextIO.output (outIO, word)
       in
         writeLines (outIO, tl)