From 88ec026921aabcdc14c971284a7a0564870fe927 Mon Sep 17 00:00:00 2001
From: Humza Shahid <humzasaur@gmail.com>
Date: Tue, 3 Sep 2024 22:09:31 +0100
Subject: [PATCH] progress with coding insert function

---
 src/string-set.sml | 76 ++++++++++++++++++++++++++++++++++------------
 1 file changed, 57 insertions(+), 19 deletions(-)

diff --git a/src/string-set.sml b/src/string-set.sml
index d1c35d1..a8e88a6 100644
--- a/src/string-set.sml
+++ b/src/string-set.sml
@@ -58,10 +58,9 @@ struct
       else SEARCH_KEY_CONTAINS_TRIE_KEY
     else 
       (* implicit: keyPos > String.size searchKey *) 
-      if keyPos <= String.size trieKey then
-        TRIE_KEY_CONTAINS_SEARCH_KEY
-      else
-        NO_SEARCH_MATCH
+      if keyPos <= String.size trieKey 
+      then TRIE_KEY_CONTAINS_SEARCH_KEY
+      else NO_SEARCH_MATCH
 
   fun isFoundNode node =
     case node of
@@ -123,14 +122,14 @@ struct
 
   datatype insert_string_match =
     NO_INSERT_MATCH
-    (* may need to split string if difference found but prefix matched *)
+  (* may need to split string if difference found but prefix matched *)
   | DIFFERENCE_FOUND_AT of int
-    (* no need to do anything if insert key matched, 
-     * as this is a set where only strings are stored *)
+  (* no need to do anything if insert key matched, 
+   * as this is a set where only strings are stored *)
   | FULL_INSERT_MATCH
-    (* if insert key contains trie key, may need to recurse down node *)
+  (* if insert key contains trie key, may need to recurse down node *)
   | INSERT_KEY_CONTAINS_TRIE_KEY
-    (* if trie key contains insert key, need to split node *)
+  (* if trie key contains insert key, need to split node *)
   | TRIE_KEY_CONTAINS_INSERT_KEY
 
   fun insertKeyMatch (insertKey, trieKey, keyPos) =
@@ -157,20 +156,59 @@ struct
       else INSERT_KEY_CONTAINS_TRIE_KEY
     else 
       (* implicit: keyPos > String.size insertKey *) 
-      if keyPos <= String.size trieKey then
-        TRIE_KEY_CONTAINS_INSERT_KEY
+      if keyPos <= String.size trieKey 
+      then TRIE_KEY_CONTAINS_INSERT_KEY
+      else NO_INSERT_MATCH
+
+  datatype insert_bin_search_result =
+    INSERT_NEW_CHILD of int
+  | FOUND_INSERT_POS of int
+  | APPEND_NEW_CHILD
+
+  fun linearSearch (findChr, keyPos, idx, children) =
+    if idx = Vector.length children then
+      APPEND_NEW_CHILD
+    else
+      let
+        val curStr = Vector.sub (children, idx)
+        val curChr = String.sub (curStr, keyPos)
+      in
+        if curChr > findChr then INSERT_NEW_CHILD idx
+        else linearSearch (findChr, keyPos, idx + 1, children)
+      end
+
+  fun helpInsertBinSearch (findChr, keyPos, children, low, high) =
+    let
+      val mid = low + ((high - low) div 2)
+    in
+      if high >= low then
+        let
+          val midStr = Vector.sub (children, mid)
+          val midChr = String.sub (midStr, keyPos)
+        in
+          if midChr = findChr then
+            FOUND_INSERT_POS mid
+          else if midChr < findChr then
+            helpInsertBinSearch (findChr, keyPos, children, mid + 1, high)
+          else
+            helpInsertBinSearch (findChr, keyPos, children, low, mid - 1)
+        end
       else
-        NO_INSERT_MATCH
+        linearSearch (findChr, keyPos, mid, children)
+    end
+
+  fun insertBinSearch (findChr, keyPos, children) =
+    helpInsertBinSearch
+      (findChr, keyPos, children, 0, Vector.length children - 1)
 
   (* 
    * todo: 
-   * Code another function for binary search; this time for insert.
-   * The difference between findBinSearch and insertBinSearch
-   * should be that, while findBinSearch returns NONE if the value is not found,
-   * insertBinSearch should return the appropriate index to insert at
-   * if the value is not found.
-   * The main insert function can then create new key/children vectors,
-   * containing the insertKey and FOUND at the appropriate index.
+   * Complete code for CHILDREN and FOUND_WITH_CHILDREN cases in insert function.
+   * The start will be similar to the exists function, 
+   * with binary search possibly followed by a string matching comparison.
+   * No string matching comparison is needed if the current char is not found
+   * in the binary search stage, however; 
+   * instead, we will insert a new element into the array when that happens.
    *)
 
   fun insert (insKey, keyPos, trie) =