Add classification tests

Author: kipcole9

test/classification_test.exs

@@ -0,0 +1,60 @@
+defmodule Image.ClassificationTest do
+  use ExUnit.Case, async: false
+
+  @moduletag :ml
+
+  @images Path.join(__DIR__, "support/images")
+
+  # Start the classifier serving once for the whole suite. Loading
+  # ConvNeXt-tiny-224 takes several seconds, so we do it here rather
+  # than per-test.
+  setup_all do
+    spec = Image.Classification.classifier()
+    start_supervised!(spec)
+    :ok
+  end
+
+  describe "classify/2" do
+    test "returns a predictions map with label and score keys" do
+      image = Image.open!(Path.join(@images, "puppy.webp"))
+      assert %{predictions: [%{label: label, score: score} | _rest]} =
+               Image.Classification.classify(image)
+
+      assert is_binary(label)
+      assert is_float(score)
+      assert score >= 0.0 and score <= 1.0
+    end
+  end
+
+  describe "labels/2" do
+    test "classifies a Cavalier King Charles Spaniel as a Blenheim spaniel" do
+      image = Image.open!(Path.join(@images, "puppy.webp"))
+      labels = Image.Classification.labels(image)
+      assert "Blenheim spaniel" in labels
+    end
+
+    test "classifies a Lamborghini as a sports car" do
+      image = Image.open!(Path.join(@images, "lamborghini-forsennato-concept.jpg"))
+      labels = Image.Classification.labels(image)
+      assert "sports car" in labels
+    end
+
+    test "classifies a tabby cat as a cat" do
+      image = Image.open!(Path.join(@images, "cat.png"))
+      labels = Image.Classification.labels(image)
+      assert Enum.any?(labels, &String.contains?(&1, "cat"))
+    end
+
+    test "returns an empty list when min_score is 1.0" do
+      image = Image.open!(Path.join(@images, "puppy.webp"))
+      assert Image.Classification.labels(image, min_score: 1.0) == []
+    end
+
+    test "returns more labels with a lower min_score threshold" do
+      image = Image.open!(Path.join(@images, "puppy.webp"))
+      high = Image.Classification.labels(image, min_score: 0.8)
+      low = Image.Classification.labels(image, min_score: 0.1)
+      assert length(low) >= length(high)
+    end
+  end
+end