testing: add unit testing (#11)

Co-authored-by: Roberto Pastor Muela <[email protected]>

.github/workflows/ci_cd.yml

@@ -26,6 +26,45 @@ jobs:
       - name: "Run PyAnsys code style checks"
         uses: ansys/actions/code-style@v6
 
+  tests:
+    name: "Tests"
+    runs-on: ${{ matrix.os }}
+    needs: [code-style]
+    strategy:
+      matrix:
+       os: [ubuntu-latest, windows-latest]
+       python-version: ['3.9', '3.12']
+      fail-fast: false
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install -r requirements.txt
+
+      - name: Testing
+        uses: ansys/actions/tests-pytest@v6
+        timeout-minutes: 12
+        with:
+          checkout: false
+          skip-install: true
+          pytest-extra-args: "--cov=ansys.allie.flowkit.python --cov-report=term --cov-report=html:.cov/html --cov-report=xml:.cov/coverage.xml"
+
+      - name: Upload coverage results (HTML)
+        uses: actions/upload-artifact@v4
+        if: (matrix.python-version == env.MAIN_PYTHON_VERSION) && (runner.os == 'Linux')
+        with:
+          name: coverage-html
+          path: .cov/html
+          retention-days: 7
+
   release:
     name: "Release project"
     if: github.event_name == 'push' && contains(github.ref, 'refs/tags')

app/fastapi_utils.py

@@ -77,8 +77,8 @@ def get_parameters_info(params):
         if param.annotation == bytes:
             param_info = ParameterInfo(name=param.name, type="bytes")
             parameters_info.append(param_info)
-        elif hasattr(param.annotation, "schema"):
-            schema = param.annotation.schema()
+        elif hasattr(param.annotation, "model_json_schema"):
+            schema = param.annotation.model_json_schema()
             param_info = extract_fields_from_schema(schema)
             parameters_info.extend(param_info)
         else:
@@ -101,7 +101,7 @@ def get_return_type_info(return_type: Type[BaseModel]):
         A list of ParameterInfo objects representing the return type fields.
 
     """
-    if hasattr(return_type, "schema"):
+    if hasattr(return_type, "model_json_schema"):
         schema = return_type.model_json_schema()
         return extract_fields_from_schema(schema)
     return [ParameterInfo(name="return", type=str(return_type.__name__))]

requirements.txt

@@ -1,6 +1,9 @@
 fastapi==0.112.0
+httpx==0.27.0
 langchain==0.2.12
 pydantic==2.8.2
 pymupdf==1.24.9
+pytest==8.3.2
+pytest-cov==5.0.0
 python_pptx==1.0.1
 PyYAML==6.0.1

setup.py

@@ -0,0 +1,7 @@
+from setuptools import find_packages, setup
+
+setup(
+    name="ansys-allie-flowkit-python",
+    version="0.1.0",
+    packages=find_packages(include=["app", "docker", "configs"]),
+)

tests/__init__.py

@@ -0,0 +1 @@
+"""Tests module."""

tests/conftest.py

@@ -0,0 +1,13 @@
+from unittest.mock import patch
+
+import pytest
+
+# Mock API key for testing
+MOCK_API_KEY = "test_api_key"
+
+
+@pytest.fixture(autouse=True)
+def mock_api_key():
+    """Mock the API key for testing."""
+    with patch("app.config.CONFIG.flowkit_python_api_key", MOCK_API_KEY):
+        yield

tests/test_endpoints_splitter.py

@@ -0,0 +1,116 @@
+import base64
+
+from app.app import app
+from app.endpoints.splitter import validate_request
+from app.models.splitter import SplitterRequest
+from fastapi import HTTPException
+from fastapi.testclient import TestClient
+import pytest
+
+from tests.conftest import MOCK_API_KEY
+
+# Create a test client
+client = TestClient(app)
+
+
+def encode_file_to_base64(file_path):
+    """Encode a file to base64 string."""
+    with open(file_path, "rb") as file:
+        return base64.b64encode(file.read()).decode("utf-8")
+
+
+@pytest.mark.asyncio
+async def test_split_ppt():
+    """Test splitting text in a PowerPoint document into chunks."""
+    ppt_content_base64 = encode_file_to_base64("./tests/test_files/test_presentation.pptx")
+    request_payload = {
+        "document_content": ppt_content_base64,
+        "chunk_size": 100,
+        "chunk_overlap": 10,
+    }
+    response = client.post("/splitter/ppt", json=request_payload, headers={"api-key": MOCK_API_KEY})
+    if response.status_code != 200:
+        print(f"Response status code: {response.status_code}")
+        print(f"Response content: {response.json()}")
+    assert response.status_code == 200
+    assert "chunks" in response.json()
+
+
+@pytest.mark.asyncio
+async def test_split_py():
+    """Test splitting Python code into chunks."""
+    python_code = """
+    def hello_world():
+        print("Hello, world!")
+    """
+    python_code_base64 = base64.b64encode(python_code.encode()).decode("utf-8")
+    request_payload = {"document_content": python_code_base64, "chunk_size": 50, "chunk_overlap": 5}
+    response = client.post("/splitter/py", json=request_payload, headers={"api-key": MOCK_API_KEY})
+    assert response.status_code == 200
+    assert "chunks" in response.json()
+
+
+@pytest.mark.asyncio
+async def test_split_pdf():
+    """Test splitting text in a PDF document into chunks."""
+    pdf_content_base64 = encode_file_to_base64("./tests/test_files/test_document.pdf")
+    request_payload = {
+        "document_content": pdf_content_base64,
+        "chunk_size": 200,
+        "chunk_overlap": 20,
+    }
+    response = client.post("/splitter/pdf", json=request_payload, headers={"api-key": MOCK_API_KEY})
+    assert response.status_code == 200
+    assert "chunks" in response.json()
+
+
+# Define test cases for validate_request()
+validate_request_test_cases = [
+    # Test case 1: valid request
+    (
+        SplitterRequest(
+            document_content="dGVzdA==", chunk_size=100, chunk_overlap=10  # base64 for "test"
+        ),
+        MOCK_API_KEY,
+        None,
+    ),
+    # Test case: invalid API key
+    (
+        SplitterRequest(document_content="dGVzdA==", chunk_size=100, chunk_overlap=10),
+        "invalid_api_key",
+        HTTPException(status_code=401, detail="Invalid API key"),
+    ),
+    # Test case 2: missing document content
+    (
+        SplitterRequest(document_content="", chunk_size=100, chunk_overlap=10),
+        MOCK_API_KEY,
+        HTTPException(status_code=400, detail="No document content provided"),
+    ),
+    # Test case 4: invalid chunk size
+    (
+        SplitterRequest(document_content="dGVzdA==", chunk_size=0, chunk_overlap=10),
+        MOCK_API_KEY,
+        HTTPException(status_code=400, detail="No chunk size provided"),
+    ),
+    # Test case 5: invalid chunk overlap
+    (
+        SplitterRequest(document_content="dGVzdA==", chunk_size=100, chunk_overlap=-1),
+        MOCK_API_KEY,
+        HTTPException(status_code=400, detail="Chunk overlap must be greater than or equal to 0"),
+    ),
+]
+
+
+@pytest.mark.parametrize("api_request, api_key, expected_exception", validate_request_test_cases)
+def test_validate_request(api_request, api_key, expected_exception):
+    """Test the validate_request function with various scenarios."""
+    if expected_exception:
+        with pytest.raises(HTTPException) as exc_info:
+            validate_request(api_request, api_key)
+        assert exc_info.value.status_code == expected_exception.status_code
+        assert exc_info.value.detail == expected_exception.detail
+    else:
+        try:
+            validate_request(api_request, api_key)
+        except HTTPException:
+            pytest.fail("validate_request() raised HTTPException unexpectedly!")

tests/test_list_functions.py

@@ -0,0 +1,58 @@
+from app.app import app
+from fastapi.testclient import TestClient
+import pytest
+
+# Initialize the test client
+client = TestClient(app)
+
+
+@pytest.mark.asyncio
+async def test_list_functions():
+    """Test listing available functions."""
+    # Test splitter results
+    response = client.get("/", headers={"api-key": "test_api_key"})
+    assert response.status_code == 200
+    response_data = response.json()
+
+    expected_response_start = [
+        {
+            "name": "split_ppt",
+            "path": "/splitter/ppt",
+            "inputs": [
+                {"name": "document_content", "type": "string(binary)"},
+                {"name": "chunk_size", "type": "integer"},
+                {"name": "chunk_overlap", "type": "integer"},
+            ],
+            "outputs": [{"name": "chunks", "type": "array<string>"}],
+            "definitions": {},
+        },
+        {
+            "name": "split_py",
+            "path": "/splitter/py",
+            "inputs": [
+                {"name": "document_content", "type": "string(binary)"},
+                {"name": "chunk_size", "type": "integer"},
+                {"name": "chunk_overlap", "type": "integer"},
+            ],
+            "outputs": [{"name": "chunks", "type": "array<string>"}],
+            "definitions": {},
+        },
+        {
+            "name": "split_pdf",
+            "path": "/splitter/pdf",
+            "inputs": [
+                {"name": "document_content", "type": "string(binary)"},
+                {"name": "chunk_size", "type": "integer"},
+                {"name": "chunk_overlap", "type": "integer"},
+            ],
+            "outputs": [{"name": "chunks", "type": "array<string>"}],
+            "definitions": {},
+        },
+    ]
+
+    assert response_data[:3] == expected_response_start
+
+    # Test invalid API key
+    response = client.get("/", headers={"api-key": "invalid_api_key"})
+    assert response.status_code == 401
+    assert response.json() == {"detail": "Invalid API key"}