Moved n_components from cluster estimation into config.yml (#234)

* moved n_components in cluster estimation to config.yaml * moved n_components from cluster estimation to config.yaml * added check verifying max_num_components > 0 * Refined changes to pass tests --------- Co-authored-by: Jane <[email protected]>
UWARG · Jan 18, 2025 · 8633b43 · 8633b43
1 parent 86cc4f2
commit 8633b43
Show file tree

Hide file tree

Showing 5 changed files with 26 additions and 6 deletions.
diff --git a/config.yaml b/config.yaml
@@ -70,6 +70,7 @@ geolocation:
 cluster_estimation:
     min_activation_threshold: 25
     min_new_points_to_run: 5
+    max_num_components: 10
     random_state: 0
 
 communications:

diff --git a/main_2024.py b/main_2024.py
@@ -137,6 +137,7 @@ def main() -> int:
 
         MIN_ACTIVATION_THRESHOLD = config["cluster_estimation"]["min_activation_threshold"]
         MIN_NEW_POINTS_TO_RUN = config["cluster_estimation"]["min_new_points_to_run"]
+        MAX_NUM_COMPONENTS = config["cluster_estimation"]["max_num_components"]
         RANDOM_STATE = config["cluster_estimation"]["random_state"]
 
         COMMUNICATIONS_TIMEOUT = config["communications"]["timeout"]
@@ -327,7 +328,12 @@ def main() -> int:
     result, cluster_estimation_worker_properties = worker_manager.WorkerProperties.create(
         count=1,
         target=cluster_estimation_worker.cluster_estimation_worker,
-        work_arguments=(MIN_ACTIVATION_THRESHOLD, MIN_NEW_POINTS_TO_RUN, RANDOM_STATE),
+        work_arguments=(
+            MIN_ACTIVATION_THRESHOLD,
+            MIN_NEW_POINTS_TO_RUN,
+            MAX_NUM_COMPONENTS,
+            RANDOM_STATE,
+        ),
         input_queues=[geolocation_to_cluster_estimation_queue],
         output_queues=[cluster_estimation_to_communications_queue],
         controller=controller,

diff --git a/modules/cluster_estimation/cluster_estimation.py b/modules/cluster_estimation/cluster_estimation.py
@@ -28,6 +28,9 @@ class ClusterEstimation:
     min_new_points_to_run: int
         Minimum number of new data points that must be collected before running model.
 
+    max_num_components: int
+        Max number of real landing pads.
+
     random_state: int
         Seed for randomizer, to get consistent results.
 
@@ -62,9 +65,6 @@ class ClusterEstimation:
     __MEAN_PRECISION_PRIOR = 1e-6
     __MAX_MODEL_ITERATIONS = 1000
 
-    # Real-world scenario Hyperparameters
-    __MAX_NUM_COMPONENTS = 10  # assumed maximum number of real landing pads
-
     # Hyperparameters to clean up model outputs
     __WEIGHT_DROP_THRESHOLD = 0.1
     __MAX_COVARIANCE_THRESHOLD = 10
@@ -74,6 +74,7 @@ def create(
         cls,
         min_activation_threshold: int,
         min_new_points_to_run: int,
+        max_num_components: int,
         random_state: int,
         local_logger: logger.Logger,
     ) -> "tuple[bool, ClusterEstimation | None]":
@@ -88,10 +89,15 @@ def create(
         if min_activation_threshold < 1:
             return False, None
 
+        # This must be greater than 0
+        if max_num_components < 0:
+            return False, None
+
         return True, ClusterEstimation(
             cls.__create_key,
             min_activation_threshold,
             min_new_points_to_run,
+            max_num_components,
             random_state,
             local_logger,
         )
@@ -101,6 +107,7 @@ def __init__(
         class_private_create_key: object,
         min_activation_threshold: int,
         min_new_points_to_run: int,
+        max_num_components: int,
         random_state: int,
         local_logger: logger.Logger,
     ) -> None:
@@ -112,7 +119,7 @@ def __init__(
         # Initializes VGMM
         self.__vgmm = sklearn.mixture.BayesianGaussianMixture(
             covariance_type=self.__COVAR_TYPE,
-            n_components=self.__MAX_NUM_COMPONENTS,
+            n_components=max_num_components,
             init_params=self.__MODEL_INIT_PARAM,
             weight_concentration_prior=self.__WEIGHT_CONCENTRATION_PRIOR,
             mean_precision_prior=self.__MEAN_PRECISION_PRIOR,

diff --git a/modules/cluster_estimation/cluster_estimation_worker.py b/modules/cluster_estimation/cluster_estimation_worker.py
@@ -14,6 +14,7 @@
 def cluster_estimation_worker(
     min_activation_threshold: int,
     min_new_points_to_run: int,
+    max_num_components: int,
     random_state: int,
     input_queue: queue_proxy_wrapper.QueueProxyWrapper,
     output_queue: queue_proxy_wrapper.QueueProxyWrapper,
@@ -30,6 +31,9 @@ def cluster_estimation_worker(
     min_new_points_to_run: int
         Minimum number of new data points that must be collected before running model.
 
+    max_num_components: int
+        Max number of real landing pads.
+
     random_state: int
         Seed for randomizer, to get consistent results.
 
@@ -56,6 +60,7 @@ def cluster_estimation_worker(
     result, estimator = cluster_estimation.ClusterEstimation.create(
         min_activation_threshold,
         min_new_points_to_run,
+        max_num_components,
         random_state,
         local_logger,
     )

diff --git a/tests/unit/test_cluster_detection.py b/tests/unit/test_cluster_detection.py
@@ -13,10 +13,10 @@
 
 MIN_TOTAL_POINTS_THRESHOLD = 100
 MIN_NEW_POINTS_TO_RUN = 10
+MAX_NUM_COMPONENTS = 10
 RNG_SEED = 0
 CENTRE_BOX_SIZE = 500
 
-
 # Test functions use test fixture signature names and access class privates
 # No enable
 # pylint: disable=protected-access,redefined-outer-name
@@ -34,6 +34,7 @@ def cluster_model() -> cluster_estimation.ClusterEstimation:  # type: ignore
     result, model = cluster_estimation.ClusterEstimation.create(
         MIN_TOTAL_POINTS_THRESHOLD,
         MIN_NEW_POINTS_TO_RUN,
+        MAX_NUM_COMPONENTS,
         RNG_SEED,
         test_logger,
     )