Add Python Dependency Deck

.github/workflows/pythonbuild.yml

@@ -149,6 +149,41 @@ jobs:
           fail_ci_if_error: false
           files: coverage.xml
 
+  test-hypothesis:
+    needs:
+      - detect-python-versions
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest]
+        python-version: ${{fromJson(needs.detect-python-versions.outputs.python-versions)}}
+    steps:
+      - uses: actions/checkout@v4
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
+      - name: Cache pip
+        uses: actions/cache@v3
+        with:
+          # This path is specific to Ubuntu
+          path: ~/.cache/pip
+          # Look to see if there is a cache hit for the corresponding requirements files
+          key: ${{ format('{0}-pip-{1}', runner.os, hashFiles('dev-requirements.in', 'requirements.in')) }}
+      - name: Install dependencies
+        run: make setup && pip freeze
+      - name: Test with coverage
+        env:
+          FLYTEKIT_HYPOTHESIS_PROFILE: ci
+        run: |
+          make unit_test_hypothesis
+      - name: Codecov
+        uses: codecov/[email protected]
+        with:
+          fail_ci_if_error: false
+          files: coverage.xml
+
   test-serialization:
     needs:
       - detect-python-versions

Makefile

@@ -62,10 +62,13 @@ unit_test_extras_codecov:
 unit_test:
 	# Skip all extra tests and run them with the necessary env var set so that a working (albeit slower)
 	# library is used to serialize/deserialize protobufs is used.
-	$(PYTEST_AND_OPTS) -m "not (serial or sandbox_test)" tests/flytekit/unit/ --ignore=tests/flytekit/unit/extras/ --ignore=tests/flytekit/unit/models --ignore=tests/flytekit/unit/extend ${CODECOV_OPTS}
+	$(PYTEST_AND_OPTS) -m "not (serial or sandbox_test or hypothesis)" tests/flytekit/unit/ --ignore=tests/flytekit/unit/extras/ --ignore=tests/flytekit/unit/models --ignore=tests/flytekit/unit/extend ${CODECOV_OPTS}
 	# Run serial tests without any parallelism
 	$(PYTEST) -m "serial" tests/flytekit/unit/ --ignore=tests/flytekit/unit/extras/ --ignore=tests/flytekit/unit/models --ignore=tests/flytekit/unit/extend ${CODECOV_OPTS}
 
+.PHONY: unit_test_hypothesis
+unit_test_hypothesis:
+	$(PYTEST_AND_OPTS) -m "hypothesis" tests/flytekit/unit/experimental ${CODECOV_OPTS}
 
 .PHONY: unit_test_extras
 unit_test_extras:

flytekit/core/python_function_task.py

@@ -349,15 +349,21 @@ def dynamic_execute(self, task_function: Callable, **kwargs) -> Any:
         raise ValueError(f"Invalid execution provided, execution state: {ctx.execution_state}")
 
     def _write_decks(self, native_inputs, native_outputs_as_map, ctx, new_user_params):
-        # These errors are raised if the source code can not be retrieved
-        with suppress(OSError, TypeError):
-            source_code = inspect.getsource(self._task_function)
-
+        if self._disable_deck is False:
             from flytekit.deck import Deck
-            from flytekit.deck.renderer import SourceCodeRenderer
+            from flytekit.deck.renderer import PythonDependencyRenderer
+
+            # These errors are raised if the source code can not be retrieved
+            with suppress(OSError, TypeError):
+                source_code = inspect.getsource(self._task_function)
+                from flytekit.deck.renderer import SourceCodeRenderer
+
+                source_code_deck = Deck("Source Code")
+                renderer = SourceCodeRenderer()
+                source_code_deck.append(renderer.to_html(source_code))
 
-            source_code_deck = Deck("Source Code")
-            renderer = SourceCodeRenderer()
-            source_code_deck.append(renderer.to_html(source_code))
+            python_dependencies_deck = Deck("Dependencies")
+            renderer = PythonDependencyRenderer()
+            python_dependencies_deck.append(renderer.to_html())
 
         return super()._write_decks(native_inputs, native_outputs_as_map, ctx, new_user_params)

flytekit/deck/renderer.py

@@ -86,3 +86,77 @@
         css = formatter.get_style_defs(".highlight").replace("#fff0f0", "#ffffff")
         html = highlight(source_code, PythonLexer(), formatter)
         return f"<style>{css}</style>{html}"
+
+
+class PythonDependencyRenderer:
+    """
+    PythonDependencyDeck is a deck that contains information about packages installed via pip.
+    """
+
+    def __init__(self, title: str = "Dependencies"):
+        self._title = title
+
+    def to_html(self) -> str:
+        import json
+        import subprocess
+        import sys
+
+        from flytekit.loggers import logger
+
+        try:
+            installed_packages = json.loads(
+                subprocess.check_output([sys.executable, "-m", "pip", "list", "--format", "json"])
+            )
+            requirements_txt = (
+                subprocess.check_output([sys.executable, "-m", "pip", "freeze"])
+                .decode("utf-8")
+                .replace("\\n", "\n")
+                .rstrip()
+            )
+        except subprocess.CalledProcessError as e:
+            logger.error(f"Error occurred while fetching installed packages: {e}")
+            return "Error occurred while fetching installed packages."
+
+        table = (
+            "<table>\n<tr>\n<th style='text-align:left;'>Name</th>\n<th style='text-align:left;'>Version</th>\n</tr>\n"
+        )
+
+        for entry in installed_packages:
+            table += f"<tr>\n<td>{entry['name']}</td>\n<td>{entry['version']}</td>\n</tr>\n"
+
+        table += "</table>"
+
+        html = f"""
+        <!DOCTYPE html>
+        <html lang="en">
+        <head>
+        <meta charset="UTF-8">
+        <meta name="viewport" content="width=device-width, initial-scale=1.0">
+        <title>Flyte Dependencies</title>
+        <script>
+        async function copyTable() {{
+          var requirements_txt = document.getElementById('requirements_txt');
+
+          try {{
+            await navigator.clipboard.writeText(requirements_txt.innerText);
+          }} catch (err) {{
+            console.log('Error accessing the clipboard: ' + err);
+          }}
+        }}
+        </script>
+        </head>
+        <body>
+
+        <button onclick="copyTable()">
+          <span>Copy table as requirements.txt</span>
+        </button>
+        <h3>Python Dependencies</h3>
+
+        {table}
+
+        <div id="requirements_txt" style="display:none">{requirements_txt}</div>
+
+        </body>
+        </html>
+        """
+        return html

pyproject.toml

@@ -99,6 +99,7 @@ log_cli_level = 20
 markers = [
     "sandbox_test: fake integration tests", # unit tests that are really integration tests that run on a sandbox environment
     "serial: tests to avoid using with pytest-xdist",
+    "hypothesis: tests that use they hypothesis library",
 ]
 
 [tool.coverage.report]

tests/flytekit/unit/conftest.py

@@ -1,4 +1,7 @@
+import os
+
 import pytest
+from hypothesis import settings
 
 from flytekit.image_spec.image_spec import ImageSpecBuilder
 
@@ -11,3 +14,9 @@ def build_image(self, img):
 @pytest.fixture()
 def mock_image_spec_builder():
     return MockImageSpecBuilder()
+
+
+settings.register_profile("ci", max_examples=5, deadline=100_000)
+settings.register_profile("dev", max_examples=10, deadline=10_000)
+
+settings.load_profile(os.getenv("FLYTEKIT_HYPOTHESIS_PROFILE", "dev"))

tests/flytekit/unit/deck/test_deck.py

@@ -3,12 +3,13 @@
 
 import pytest
 from markdown_it import MarkdownIt
-from mock import mock
+from mock import mock, patch
 
 import flytekit
 from flytekit import Deck, FlyteContextManager, task
 from flytekit.deck import MarkdownRenderer, SourceCodeRenderer, TopFrameRenderer
 from flytekit.deck.deck import _output_deck
+from flytekit.deck.renderer import PythonDependencyRenderer
 
 
 @pytest.mark.skipif("pandas" not in sys.modules, reason="Pandas is not installed.")
@@ -50,9 +51,9 @@ def test_timeline_deck():
 @pytest.mark.parametrize(
     "disable_deck,expected_decks",
     [
-        (None, 2),  # time line deck + source code deck
-        (False, 4),  # time line deck + source code deck + input and output decks
-        (True, 2),  # time line deck + source code deck
+        (None, 1),  # time line deck
+        (False, 5),  # time line deck + source code deck + python dependency deck + input and output decks
+        (True, 1),  # time line deck
     ],
 )
 def test_deck_for_task(disable_deck, expected_decks):
@@ -75,11 +76,21 @@ def t1(a: int) -> str:
 @pytest.mark.parametrize(
     "enable_deck,disable_deck, expected_decks, expect_error",
     [
-        (None, None, 3, False),  # default deck and time line deck + source code deck
-        (None, False, 5, False),  # default deck and time line deck + source code deck + input and output decks
-        (None, True, 3, False),  # default deck and time line deck + source code deck
-        (True, None, 5, False),  # default deck and time line deck + source code deck + input and output decks
-        (False, None, 3, False),  # default deck and time line deck + source code deck
+        (None, None, 2, False),  # default deck and time line deck
+        (
+            None,
+            False,
+            6,
+            False,
+        ),  # default deck and time line deck + source code deck + python dependency deck + input and output decks
+        (None, True, 2, False),  # default deck and time line deck
+        (
+            True,
+            None,
+            6,
+            False,
+        ),  # default deck and time line deck + source code deck + python dependency deck + input and output decks
+        (False, None, 2, False),  # default deck and time line deck
         (True, True, -1, True),  # Set both disable_deck and enable_deck to True and confirm that it fails
         (False, False, -1, True),  # Set both disable_deck and enable_deck to False and confirm that it fails
     ],
@@ -176,3 +187,19 @@ def test_source_code_renderer():
     # Assert that the color #ffffff is used instead of #fff0f0
     assert "#ffffff" in result
     assert "#fff0f0" not in result
+
+
+def test_python_dependency_renderer():
+    with patch("subprocess.check_output") as mock_check_output:
+        mock_check_output.return_value = '[{"name": "numpy", "version": "1.21.0"}]'.encode()
+        renderer = PythonDependencyRenderer()
+        result = renderer.to_html()
+        assert "numpy" in result
+        assert "1.21.0" in result
+
+        # Assert that the result includes parts of the python dependency
+        assert "Name" in result
+        assert "Version" in result
+
+        # Assert that the button of copy
+        assert 'button onclick="copyTable()"' in result

tests/flytekit/unit/experimental/test_eager_workflows.py

@@ -6,7 +6,7 @@
 
 import hypothesis.strategies as st
 import pytest
-from hypothesis import given, settings
+from hypothesis import given
 
 from flytekit import dynamic, task, workflow
 from flytekit.exceptions.user import FlyteValidationException
@@ -15,7 +15,6 @@
 from flytekit.types.file import FlyteFile
 from flytekit.types.structured import StructuredDataset
 
-DEADLINE = 2000
 INTEGER_ST = st.integers(min_value=-10_000_000, max_value=10_000_000)
 
 
@@ -48,7 +47,7 @@ def dynamic_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_simple_eager_workflow(x_input: int):
     """Testing simple eager workflow with just tasks."""
 
@@ -62,7 +61,7 @@ async def eager_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_conditional_eager_workflow(x_input: int):
     """Test eager workflow with conditional logic."""
 
@@ -80,7 +79,7 @@ async def eager_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_try_except_eager_workflow(x_input: int):
     """Test eager workflow with try/except logic."""
 
@@ -99,7 +98,7 @@ async def eager_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST, n_input=st.integers(min_value=1, max_value=20))
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_gather_eager_workflow(x_input: int, n_input: int):
     """Test eager workflow with asyncio gather."""
 
@@ -113,7 +112,7 @@ async def eager_wf(x: int, n: int) -> typing.List[int]:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_eager_workflow_with_dynamic_exception(x_input: int):
     """Test eager workflow with dynamic workflow is not supported."""
 
@@ -131,7 +130,7 @@ async def nested_eager_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_nested_eager_workflow(x_input: int):
     """Testing running nested eager workflows."""
 
@@ -145,7 +144,7 @@ async def eager_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_eager_workflow_within_workflow(x_input: int):
     """Testing running eager workflow within a static workflow."""
 
@@ -168,7 +167,7 @@ def subworkflow(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_workflow_within_eager_workflow(x_input: int):
     """Testing running a static workflow within an eager workflow."""
 
@@ -182,7 +181,7 @@ async def eager_wf(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
+@pytest.mark.hypothesis
 def test_local_task_eager_workflow_exception(x_input: int):
     """Testing simple eager workflow with a local function task doesn't work."""
 
@@ -199,8 +198,8 @@ async def eager_wf_with_local(x: int) -> int:
 
 
 @given(x_input=INTEGER_ST)
-@settings(deadline=DEADLINE, max_examples=5)
 @pytest.mark.filterwarnings("ignore:coroutine 'AsyncEntity.__call__' was never awaited")
+@pytest.mark.hypothesis
 def test_local_workflow_within_eager_workflow_exception(x_input: int):
     """Cannot call a locally-defined workflow within an eager workflow"""
 
@@ -243,6 +242,7 @@ def create_directory() -> FlyteDirectory:
 
 
 @pytest.mark.skipif("pandas" not in sys.modules, reason="Pandas is not installed.")
+@pytest.mark.hypothesis
 def test_eager_workflow_with_offloaded_types():
     """Test eager workflow that eager workflows work with offloaded types."""
     import pandas as pd