Refactor benchmark to use randomized vectors instead of sets

- Replaced std::set with std::vector for sequence handling.
- Added randomization of sequences to avoid ordered input bias.
- Removed unused balance function and related code in bst.cpp and bst.h.
- Fixed bugs in list insertion and search logic.
- Updated plot.py to allow custom y-axis labels and enable log scale for build plots.

Lab2/plot.py

@@ -1,11 +1,11 @@
 import matplotlib.pyplot as plt
 
-def plot(header:str, first:list[int], second:list[int], log=False, labels=("Lista", "Drzewo BST")):
+def plot(header:str, first:list[int], second:list[int], log=False, labels=("Lista", "Drzewo BST"), ylabel= "Czas (ms)"):
     plt.figure(figsize=(10, 6))
     plt.plot(range(1000,25001,1000), first, label=labels[0])
     plt.plot(range(1000,25001,1000), second, label=labels[1])
     plt.xlabel("Rozmiar tablicy")
-    plt.ylabel("Czas (ms)")
+    plt.ylabel(ylabel)
     if log:
         plt.yscale('log', base=2)
     plt.title(header)
@@ -42,7 +42,7 @@ if __name__ == "__main__":
                 listTimes['delete'].append(float(row[3]))
 
 
-        plot("Tworzenie", listTimes['build'], bstTimes['build'])
+        plot("Tworzenie", listTimes['build'], bstTimes['build'], log=True)
         plot("Wyszukiwanie", listTimes['search'], bstTimes['search'], log=True)
         plot("Usuwanie", listTimes['delete'], bstTimes['delete'])
 
@@ -55,4 +55,4 @@ if __name__ == "__main__":
         bstHeights = [int(x[0]) for x in values]
         avlHeights = [int(x[1]) for x in values]
 
-        plot("AVl", bstHeights, avlHeights, log=True, labels=("Drzewo BST", "Drzewo AVL"))
+        plot("AVL", bstHeights, avlHeights, log=True, labels=("Drzewo BST", "Drzewo AVL"), ylabel="Wysokość drzewa")