Add tests for notebooks

.github/workflows/notebooks.yml

@@ -0,0 +1,45 @@
+# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
+
+# Notebooks (CI) GitHub Actions tests
+
+name: Run Notebooks
+
+permissions:
+  contents: read
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+  schedule:
+    - cron: "0 8 * * *"  # Daily at 08:00 UTC
+
+jobs:
+  notebooks:
+    name: Tests
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v3
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: '3.12'
+
+      - name: Install dependencies  # Transformer dependency for Florence-2
+        run: |
+          python -m pip install --upgrade pip
+          pip install ultralytics pytest transformers==4.49.0
+
+      - name: Run notebook tests
+        run: |
+          echo "CI event: ${{ github.event_name }}"
+          if [ "${{ github.event_name }}" = "schedule" ]; then
+            echo "Running full test suite including @slow tests"
+            pytest tests/test_notebooks.py -p no:warnings
+          else
+            echo "Running fast tests only (excluding @slow)"
+            pytest -m "not slow" tests/test_notebooks.py -p no:warnings
+          fi

tests/test_notebooks.py

@@ -0,0 +1,73 @@
+# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
+
+# Tests Ultralytics Notebooks: https://github.com/ultralytics/notebooks/
+
+import cv2
+import pytest
+from ultralytics import YOLO, SAM, ASSETS
+from ultralytics.utils.downloads import safe_download
+from ultralytics.utils import ROOT
+from ultralytics.solutions import ObjectCounter, TrackZone, Heatmap, AIGym
+
+TMP = (ROOT / "../tests/tmp").resolve()  # temp directory for test files
+
+DEMO_VIDEO = "https://github.com/ultralytics/notebooks/releases/download/v0.0.0/solutions-ci-demo.mp4"
+WORKOUT_VIDEO = "https://github.com/ultralytics/assets/releases/download/v0.0.0/Legextension.demo.video.mp4"
+
+@pytest.mark.parametrize("solution_class, solution_kwargs, video_source", [
+    (ObjectCounter, {"region": [(20, 400), (1080, 400), (1080, 360), (20, 360)]}, DEMO_VIDEO),
+    (TrackZone, {"region": [(20, 400), (1080, 400), (1080, 360), (20, 360)]}, DEMO_VIDEO),
+    (Heatmap, {"colormap": cv2.COLORMAP_PARULA}, DEMO_VIDEO),
+    (AIGym, {"kpts": [12, 14, 16]}, WORKOUT_VIDEO)
+])
+def test_solution_inference(solution_class, solution_kwargs, video_source):
+    """Parameterized test for multiple Ultralytics solution demos."""
+    safe_download(video_source)
+    source = "Legextension.demo.video.mp4" if solution_class=="AIGym" else "solutions-ci-demo.mp4"
+    cap = cv2.VideoCapture(source)
+    solution = solution_class(**solution_kwargs)
+
+    while cap.isOpened():
+        success, frame = cap.read()
+        if not success:
+            break
+        _ = solution(frame)  # Run inference
+    cap.release()
+
+
+def test_export_val_results():
+    """Test: notebooks/how-to-export-the-validation-results-into-dataframe-csv-sql-and-other-formats.ipynb"""
+    model = YOLO("yolo11n.pt")
+    metrics = model.val(data="coco8.yaml", verbose=True)
+    metrics.to_df()
+    metrics.to_csv()
+    metrics.to_html()
+    metrics.to_sql()
+    metrics.to_json()
+    metrics.confusion_matrix.to_df()
+    metrics.confusion_matrix.to_html()
+    metrics.confusion_matrix.to_sql()
+    metrics.confusion_matrix.to_xml()
+    metrics.confusion_matrix.to_json()
+    with open("validation_results.html", "w") as f:
+        f.write(metrics.to_xml())
+    with open("confusion_matrix.csv", "w") as f:
+        f.write(metrics.confusion_matrix.to_csv())
+
+
+@pytest.mark.slow
+def test_sam_inference():
+    """Test: notebooks/inference-with-meta-sam-and-sam2-using-ultralytics-python-package.ipynb"""
+    from ultralytics.data.annotator import auto_annotate
+
+    model = SAM("sam2.1_b.pt")
+    _ = model(ASSETS/"bus.jpg")  # image.
+    _ = model(ASSETS/"bus.jpg", points=[34, 714])  # point prompt.
+    _ = model(ASSETS/"bus.jpg", points=[[34, 714], [283, 634]]) # multiple points prompt.
+    _ = model(ASSETS / "bus.jpg", bboxes=[3.83287, 229.3560, 796.2098, 728.4313])  # bbox prompt.
+
+    for img in ["bus.jpg", "zidane.jpg"]:
+        safe_download(f"https://ultralytics.com/assets/{img}", dir="assets")
+    auto_annotate(data="/content/assets",  # return the annotation in the Ultralytics YOLO segmentation format.
+                  det_model="yolo11x.pt",
+                  sam_model="sam_b.pt")