From 2012e7ea53ed9670f9283f8d6e9328f1c79cb2f6 Mon Sep 17 00:00:00 2001
From: Chinmaya Sahu <chinmayasahu101@gmail.com>
Date: Thu, 7 Nov 2024 16:16:45 +0530
Subject: [PATCH] Made some improvements

---
 package/samplers/hill_climb_search/README.md       |  8 ++++----
 package/samplers/hill_climb_search/__init__.py     |  4 ++--
 package/samplers/hill_climb_search/example.py      | 10 ++++------
 .../hill_climb_search/hill_climb_search.py         | 14 ++++++++++----
 4 files changed, 20 insertions(+), 16 deletions(-)

diff --git a/package/samplers/hill_climb_search/README.md b/package/samplers/hill_climb_search/README.md
index dee21610..4a9ab190 100644
--- a/package/samplers/hill_climb_search/README.md
+++ b/package/samplers/hill_climb_search/README.md
@@ -13,7 +13,7 @@ The **hill climbing algorithm** is an optimization technique that iteratively im
 
 ## Class or Function Names
 
-- HillClimbSearch
+- HillClimbingSampler
 
 ## Example
 
@@ -22,12 +22,12 @@ import optuna
 import optunahub
    
 def objective(trial):
-    x = trial.suggest_discrete_uniform("x", -10, 10)
-    y = trial.suggest_discrete_uniform("y", -10, 10)
+    x = trial.suggest_int("x", -10, 10)
+    y = trial.suggest_int("y", -10, 10)
     return -(x**2 + y**2)
 
 mod = optunahub.load_module("samplers/hill_climb_search")
-sampler = mod.HillClimbSearch()
+sampler = mod.HillClimbingSampler()
 study = optuna.create_study(sampler=sampler)
 study.optimize(objective, n_trials=20)
 ```
diff --git a/package/samplers/hill_climb_search/__init__.py b/package/samplers/hill_climb_search/__init__.py
index 6b6c35bb..cfd376ea 100644
--- a/package/samplers/hill_climb_search/__init__.py
+++ b/package/samplers/hill_climb_search/__init__.py
@@ -1,4 +1,4 @@
-from .hill_climb_search import HillClimbSearch
+from .hill_climb_search import HillClimbingSampler
 
 
-__all__ = ["HillClimbSearch"]
+__all__ = ["HillClimbingSampler"]
diff --git a/package/samplers/hill_climb_search/example.py b/package/samplers/hill_climb_search/example.py
index a2c27703..0d26c2d0 100644
--- a/package/samplers/hill_climb_search/example.py
+++ b/package/samplers/hill_climb_search/example.py
@@ -5,14 +5,12 @@
 if __name__ == "__main__":
 
     def objective(trial: optuna.trial.Trial) -> None:
-        x = trial.suggest_discrete_uniform("x", -10, 10)
-        y = trial.suggest_discrete_uniform("y", -10, 10)
+        x = trial.suggest_int("x", -10, 10)
+        y = trial.suggest_int("y", -10, 10)
         return -(x**2 + y**2)
 
-    module = optunahub.load_module(
-        package="samplers/hill-climb-search", repo_owner="csking101", ref="hill-climb-algorithm"
-    )
-    sampler = module.HillClimbSearch()
+    module = optunahub.load_module(package="samplers/hill-climb-search")
+    sampler = module.HillClimbingSampler()
     study = optuna.create_study(sampler=sampler)
     study.optimize(objective, n_trials=20)
 
diff --git a/package/samplers/hill_climb_search/hill_climb_search.py b/package/samplers/hill_climb_search/hill_climb_search.py
index 0f79d301..24e946ed 100644
--- a/package/samplers/hill_climb_search/hill_climb_search.py
+++ b/package/samplers/hill_climb_search/hill_climb_search.py
@@ -7,7 +7,7 @@
 import optunahub
 
 
-class HillClimbSearch(optunahub.samplers.SimpleBaseSampler):
+class HillClimbingSampler(optunahub.samplers.SimpleBaseSampler):
     """A sampler based on the Hill Climb Local Search Algorithm dealing with discrete values."""
 
     def __init__(
@@ -32,7 +32,7 @@ def _generate_random_point(
         """This function generates a random discrete point in the search space"""
         params = {}
         for param_name, param_distribution in search_space.items():
-            if isinstance(param_distribution, optuna.distributions.FloatDistribution):
+            if isinstance(param_distribution, optuna.distributions.IntDistribution):
                 total_points = int(
                     (param_distribution.high - param_distribution.low) / param_distribution.step
                 )
@@ -70,7 +70,7 @@ def _generate_neighbors(
         """This function generates the neighbors of the current point"""
         neighbors = []
         for param_name, param_distribution in search_space.items():
-            if isinstance(param_distribution, optuna.distributions.FloatDistribution):
+            if isinstance(param_distribution, optuna.distributions.IntDistribution):
                 current_value = current_point[param_name]
                 step = param_distribution.step
 
@@ -127,7 +127,13 @@ def sample_relative(
                 # Store the value of the neighbor, if it improves upon the current point
                 neighbor_value = previous_trial.value
 
-                if neighbor_value < self._current_point_value:
+                criteria = (
+                    neighbor_value < self._current_point_value
+                    if study.direction == "minimize"
+                    else neighbor_value > self._current_point_value
+                )
+
+                if criteria:
                     self._best_neighbor = previous_trial.params
                     self._best_neighbor_value = neighbor_value