Auto detect Python wheel packages and infer build command (#603)

.github/workflows/push.yml

@@ -41,7 +41,7 @@ jobs:
       - name: Pull external libraries
         run: |
           make vendor
-          pip install wheel
+          pip3 install wheel
 
       - name: Run tests
         run: make test

bundle/artifacts/autodetect.go

@@ -0,0 +1,32 @@
+package artifacts
+
+import (
+	"context"
+
+	"github.com/databricks/cli/bundle"
+	"github.com/databricks/cli/bundle/artifacts/whl"
+	"github.com/databricks/cli/libs/log"
+)
+
+func DetectPackages() bundle.Mutator {
+	return &autodetect{}
+}
+
+type autodetect struct {
+}
+
+func (m *autodetect) Name() string {
+	return "artifacts.DetectPackages"
+}
+
+func (m *autodetect) Apply(ctx context.Context, b *bundle.Bundle) error {
+	// If artifacts section explicitly defined, do not try to auto detect packages
+	if b.Config.Artifacts != nil {
+		log.Debugf(ctx, "artifacts block is defined, skipping auto-detecting")
+		return nil
+	}
+
+	return bundle.Apply(ctx, b, bundle.Seq(
+		whl.DetectPackage(),
+	))
+}

bundle/artifacts/build.go

@@ -33,12 +33,13 @@ func (m *build) Apply(ctx context.Context, b *bundle.Bundle) error {
 		return fmt.Errorf("artifact doesn't exist: %s", m.name)
 	}
 
-	if len(artifact.Files) == 0 && artifact.BuildCommand == "" {
-		return fmt.Errorf("artifact %s misconfigured: 'files' or 'build' property is required", m.name)
+	// Skip building if build command is not specified or infered
+	if artifact.BuildCommand == "" {
+		// If no build command was specified or infered and there is no
+		// artifact output files specified, artifact is misconfigured
+		if len(artifact.Files) == 0 {
+			return fmt.Errorf("misconfigured artifact: please specify 'build' or 'files' property")
 		}
-
-	// If artifact file is explicitly defined, skip building the artifact
-	if len(artifact.Files) != 0 {
 		return nil
 	}
 

bundle/artifacts/infer.go

@@ -0,0 +1,60 @@
+package artifacts
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/databricks/cli/bundle"
+	"github.com/databricks/cli/bundle/artifacts/whl"
+	"github.com/databricks/cli/bundle/config"
+)
+
+var inferMutators map[config.ArtifactType]mutatorFactory = map[config.ArtifactType]mutatorFactory{
+	config.ArtifactPythonWheel: whl.InferBuildCommand,
+}
+
+func getInferMutator(t config.ArtifactType, name string) bundle.Mutator {
+	mutatorFactory, ok := inferMutators[t]
+	if !ok {
+		return nil
+	}
+
+	return mutatorFactory(name)
+}
+
+func InferMissingProperties() bundle.Mutator {
+	return &all{
+		name: "infer",
+		fn:   inferArtifactByName,
+	}
+}
+
+func inferArtifactByName(name string) (bundle.Mutator, error) {
+	return &infer{name}, nil
+}
+
+type infer struct {
+	name string
+}
+
+func (m *infer) Name() string {
+	return fmt.Sprintf("artifacts.Infer(%s)", m.name)
+}
+
+func (m *infer) Apply(ctx context.Context, b *bundle.Bundle) error {
+	artifact, ok := b.Config.Artifacts[m.name]
+	if !ok {
+		return fmt.Errorf("artifact doesn't exist: %s", m.name)
+	}
+
+	if artifact.BuildCommand != "" {
+		return nil
+	}
+
+	inferMutator := getInferMutator(artifact.Type, m.name)
+	if inferMutator != nil {
+		return bundle.Apply(ctx, b, inferMutator)
+	}
+
+	return nil
+}

bundle/artifacts/whl/autodetect.go

@@ -0,0 +1,74 @@
+package whl
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"path/filepath"
+	"regexp"
+	"time"
+
+	"github.com/databricks/cli/bundle"
+	"github.com/databricks/cli/bundle/config"
+	"github.com/databricks/cli/libs/cmdio"
+)
+
+type detectPkg struct {
+}
+
+func DetectPackage() bundle.Mutator {
+	return &detectPkg{}
+}
+
+func (m *detectPkg) Name() string {
+	return "artifacts.whl.AutoDetect"
+}
+
+func (m *detectPkg) Apply(ctx context.Context, b *bundle.Bundle) error {
+	cmdio.LogString(ctx, "artifacts.whl.AutoDetect: Detecting Python wheel project...")
+
+	// checking if there is setup.py in the bundle root
+	setupPy := filepath.Join(b.Config.Path, "setup.py")
+	_, err := os.Stat(setupPy)
+	if err != nil {
+		cmdio.LogString(ctx, "artifacts.whl.AutoDetect: No Python wheel project found at bundle root folder")
+		return nil
+	}
+
+	cmdio.LogString(ctx, fmt.Sprintf("artifacts.whl.AutoDetect: Found Python wheel project at %s", b.Config.Path))
+	module := extractModuleName(setupPy)
+
+	if b.Config.Artifacts == nil {
+		b.Config.Artifacts = make(map[string]*config.Artifact)
+	}
+
+	pkgPath, err := filepath.Abs(b.Config.Path)
+	if err != nil {
+		return err
+	}
+	b.Config.Artifacts[module] = &config.Artifact{
+		Path: pkgPath,
+		Type: config.ArtifactPythonWheel,
+	}
+
+	return nil
+}
+
+func extractModuleName(setupPy string) string {
+	bytes, err := os.ReadFile(setupPy)
+	if err != nil {
+		return randomName()
+	}
+
+	content := string(bytes)
+	r := regexp.MustCompile(`name=['"](.*)['"]`)
+	matches := r.FindStringSubmatch(content)
+	if len(matches) == 0 {
+		return randomName()
+	}
+	return matches[1]
+}
+
+func randomName() string {
+	return fmt.Sprintf("artifact%d", time.Now().Unix())
+}

bundle/artifacts/whl/autodetect_test.go

@@ -0,0 +1,22 @@
+package whl
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestExtractModuleName(t *testing.T) {
+	moduleName := extractModuleName("./testdata/setup.py")
+	assert.Equal(t, "my_test_code", moduleName)
+}
+
+func TestExtractModuleNameMinimal(t *testing.T) {
+	moduleName := extractModuleName("./testdata/setup_minimal.py")
+	assert.Equal(t, "my_test_code", moduleName)
+}
+
+func TestExtractModuleNameIncorrect(t *testing.T) {
+	moduleName := extractModuleName("./testdata/setup_incorrect.py")
+	assert.Contains(t, moduleName, "artifact")
+}

bundle/artifacts/whl/build.go

@@ -32,12 +32,6 @@ func (m *build) Apply(ctx context.Context, b *bundle.Bundle) error {
 		return fmt.Errorf("artifact doesn't exist: %s", m.name)
 	}
 
-	// TODO: If not set, BuildCommand should be infer prior to this
-	// via a mutator so that it can be observable.
-	if artifact.BuildCommand == "" {
-		return fmt.Errorf("artifacts.whl.Build(%s): missing build property for the artifact", m.name)
-	}
-
 	cmdio.LogString(ctx, fmt.Sprintf("artifacts.whl.Build(%s): Building...", m.name))
 
 	dir := artifact.Path

bundle/artifacts/whl/infer.go

@@ -0,0 +1,34 @@
+package whl
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/databricks/cli/bundle"
+	"github.com/databricks/cli/python"
+)
+
+type infer struct {
+	name string
+}
+
+func (m *infer) Apply(ctx context.Context, b *bundle.Bundle) error {
+	artifact := b.Config.Artifacts[m.name]
+	py, err := python.DetectExecutable(ctx)
+	if err != nil {
+		return err
+	}
+	artifact.BuildCommand = fmt.Sprintf("%s setup.py bdist_wheel", py)
+
+	return nil
+}
+
+func (m *infer) Name() string {
+	return fmt.Sprintf("artifacts.whl.Infer(%s)", m.name)
+}
+
+func InferBuildCommand(name string) bundle.Mutator {
+	return &infer{
+		name: name,
+	}
+}

bundle/artifacts/whl/testdata/setup.py

@@ -0,0 +1,15 @@
+from setuptools import setup, find_packages
+
+import my_test_code
+
+setup(
+    name="my_test_code",
+    version=my_test_code.__version__,
+    author=my_test_code.__author__,
+    url="https://databricks.com",
+    author_email="john.doe@databricks.com",
+    description="my test wheel",
+    packages=find_packages(include=["my_test_code"]),
+    entry_points={"group_1": "run=my_test_code.__main__:main"},
+    install_requires=["setuptools"],
+)

bundle/artifacts/whl/testdata/setup_incorrect.py

@@ -0,0 +1,14 @@
+from setuptools import setup, find_packages
+
+import my_test_code
+
+setup(
+    version=my_test_code.__version__,
+    author=my_test_code.__author__,
+    url="https://databricks.com",
+    author_email="john.doe@databricks.com",
+    description="my test wheel",
+    packages=find_packages(include=["my_test_code"]),
+    entry_points={"group_1": "run=my_test_code.__main__:main"},
+    install_requires=["setuptools"],
+)

bundle/artifacts/whl/testdata/setup_minimal.py

@@ -0,0 +1,3 @@
+from setuptools import setup
+
+setup(name="my_test_code")

bundle/phases/build.go

@@ -11,6 +11,8 @@ func Build() bundle.Mutator {
 	return newPhase(
 		"build",
 		[]bundle.Mutator{
+			artifacts.DetectPackages(),
+			artifacts.InferMissingProperties(),
 			artifacts.BuildAll(),
 			interpolation.Interpolate(
 				interpolation.IncludeLookupsInPath("artifacts"),

bundle/tests/bundle/python_wheel/bundle.yml

@@ -5,7 +5,7 @@ artifacts:
     my_test_code:
       type: whl
       path: "./my_test_code"
-      build: "python setup.py bdist_wheel"
+      build: "python3 setup.py bdist_wheel"
 
 resources:
   jobs:

bundle/tests/bundle/python_wheel_no_artifact/.gitignore

@@ -0,0 +1,3 @@
+build/
+*.egg-info
+.databricks

bundle/tests/bundle/python_wheel_no_artifact/bundle.yml

@@ -0,0 +1,13 @@
+bundle:
+  name: python-wheel
+
+resources:
+  jobs:
+    test_job:
+      name: "[${bundle.environment}] My Wheel Job"
+      tasks:
+        - task_key: TestTask
+          existing_cluster_id: "0717-aaaaa-bbbbbb"
+          python_wheel_task:
+            package_name: "my_test_code"
+            entry_point: "run"

bundle/tests/bundle/python_wheel_no_artifact/my_test_code/__init__.py

@@ -0,0 +1,2 @@
+__version__ = "0.0.1"
+__author__ = "Databricks"

bundle/tests/bundle/python_wheel_no_artifact/my_test_code/__main__.py

@@ -0,0 +1,16 @@
+"""
+The entry point of the Python Wheel
+"""
+
+import sys
+
+
+def main():
+    # This method will print the provided arguments
+    print('Hello from my func')
+    print('Got arguments:')
+    print(sys.argv)
+
+
+if __name__ == '__main__':
+    main()

bundle/tests/bundle/python_wheel_no_artifact/setup.py

@@ -0,0 +1,15 @@
+from setuptools import setup, find_packages
+
+import my_test_code
+
+setup(
+    name="my_test_code",
+    version=my_test_code.__version__,
+    author=my_test_code.__author__,
+    url="https://databricks.com",
+    author_email="john.doe@databricks.com",
+    description="my test wheel",
+    packages=find_packages(include=["my_test_code"]),
+    entry_points={"group_1": "run=my_test_code.__main__:main"},
+    install_requires=["setuptools"],
+)

bundle/tests/bundle/wheel_test.go

@@ -22,3 +22,16 @@ func TestBundlePythonWheelBuild(t *testing.T) {
 	require.NoError(t, err)
 	require.Equal(t, 1, len(matches))
 }
+
+func TestBundlePythonWheelBuildAutoDetect(t *testing.T) {
+	b, err := bundle.Load("./python_wheel_no_artifact")
+	require.NoError(t, err)
+
+	m := phases.Build()
+	err = m.Apply(context.Background(), b)
+	require.NoError(t, err)
+
+	matches, err := filepath.Glob("python_wheel/my_test_code/dist/my_test_code-*.whl")
+	require.NoError(t, err)
+	require.Equal(t, 1, len(matches))
+}

python/runner.go

@@ -15,7 +15,7 @@ func PyInline(ctx context.Context, inlinePy string) (string, error) {
 }
 
 func Py(ctx context.Context, script string, args ...string) (string, error) {
-	py, err := detectExecutable(ctx)
+	py, err := DetectExecutable(ctx)
 	if err != nil {
 		return "", err
 	}
@@ -70,7 +70,7 @@ func detectVirtualEnv() (string, error) {
 
 var pyExec string
 
-func detectExecutable(ctx context.Context) (string, error) {
+func DetectExecutable(ctx context.Context) (string, error) {
 	if pyExec != "" {
 		return pyExec, nil
 	}

python/runner_test.go

@@ -25,14 +25,14 @@ func TestExecAndPassError(t *testing.T) {
 
 func TestDetectPython(t *testing.T) {
 	pyExec = ""
-	py, err := detectExecutable(context.Background())
+	py, err := DetectExecutable(context.Background())
 	assert.NoError(t, err)
 	assert.Contains(t, py, "python3")
 }
 
 func TestDetectPythonCache(t *testing.T) {
 	pyExec = "abc"
-	py, err := detectExecutable(context.Background())
+	py, err := DetectExecutable(context.Background())
 	assert.NoError(t, err)
 	assert.Equal(t, "abc", py)
 	pyExec = ""
@@ -82,7 +82,7 @@ func TestPyInline(t *testing.T) {
 }
 
 func TestPyInlineStderr(t *testing.T) {
-	detectExecutable(context.Background())
+	DetectExecutable(context.Background())
 	inline := "import sys; sys.stderr.write('___msg___'); sys.exit(1)"
 	_, err := PyInline(context.Background(), inline)
 	assert.EqualError(t, err, "___msg___")