Add unit tests for calculations and utils modules

Author: cobanov

setup.py

@@ -1,5 +1,4 @@
-from setuptools import setup, find_packages
-
+from setuptools import find_packages, setup
 
 setup(
     name="tasnif",

tasnif/__init__.py

@@ -9,11 +9,7 @@
 __url__ = "https://github.com/cobanov/tasnif"
 
 
+from tasnif.calculations import calculate_kmeans, calculate_pca, get_embeddings
 from tasnif.tasnif import Tasnif
-from tasnif.calculations import get_embeddings, calculate_pca, calculate_kmeans
-from tasnif.utils import (
-    read_images_from_directory,
-    read_with_pil,
-    create_dir,
-    create_image_grid,
-)
+from tasnif.utils import (create_dir, create_image_grid,
+                          read_images_from_directory, read_with_pil)

tasnif/tasnif.py

@@ -5,13 +5,9 @@
 
 from tqdm import tqdm
 
-from .calculations import get_embeddings, calculate_pca, calculate_kmeans
-from .utils import (
-    read_images_from_directory,
-    read_with_pil,
-    create_dir,
-    create_image_grid,
-)
+from .calculations import calculate_kmeans, calculate_pca, get_embeddings
+from .utils import (create_dir, create_image_grid, read_images_from_directory,
+                    read_with_pil)
 
 warnings.filterwarnings("ignore")
 

tests/test_calculations.py

@@ -1,26 +1,24 @@
 import numpy as np
-from tasnif.calculations import get_embeddings, calculate_pca, calculate_kmeans
-from tasnif.utils import read_with_pil, read_images_from_directory
+
+from tasnif.calculations import calculate_kmeans, calculate_pca, get_embeddings
+from tasnif.utils import read_images_from_directory, read_with_pil
 
 
 def test_get_embeddings():
-    # Assuming "test_images" contains at least one image for embedding generation
     image_paths = read_images_from_directory("tests/test_images")
     images = read_with_pil(image_paths)
     embeddings = get_embeddings(images=images, use_gpu=False)
     assert embeddings is not None, "Embeddings were not generated."
 
 
 def test_calculate_pca():
-    # Generate some random data for PCA
-    embeddings = np.random.rand(20, 2048)  # Simulate 10 embeddings of dimension 2048
+    embeddings = np.random.rand(20, 2048)
     pca_embeddings = calculate_pca(embeddings, pca_dim=16)
     assert pca_embeddings.shape[1] == 16, "PCA did not reduce to the correct dimension."
 
 
 def test_calculate_kmeans():
-    # Use PCA embeddings from the previous test or generate new random data
-    pca_embeddings = np.random.rand(10, 16)  # Simulate 10 samples in 16 dimensions
+    pca_embeddings = np.random.rand(10, 16)
     num_classes = 2
     centroid, labels, counts = calculate_kmeans(pca_embeddings, num_classes)
     assert (

tests/test_utils.py

@@ -1,15 +1,12 @@
-import os
 from tasnif.utils import read_images_from_directory, read_with_pil
 
 
 def test_read_images_from_directory():
-    # Assuming you have a directory named "test_images" with some image files
     images = read_images_from_directory("tests/test_images")
     assert len(images) > 0, "No images found in the directory."
 
 
 def test_read_with_pil():
-    # Use the same directory as above or make sure it contains at least one image
     image_paths = read_images_from_directory("tests/test_images")
     images = read_with_pil(image_paths)
     assert len(images) == len(image_paths), "Not all images were read successfully."