Refactored 'jpmml_evaluator.make_evaluator()' utility function

jpmml_evaluator/__init__.py

@@ -1,8 +1,10 @@
+import os
 import pickle
 
 from abc import abstractmethod, abstractclassmethod, ABC
 from importlib.resources import files
 from pandas import DataFrame
+from pathlib import Path
 
 import numpy
 
@@ -228,6 +230,8 @@ def setLocatable(self, locatable = False):
 		return self
 
 	def loadFile(self, path):
+		if isinstance(path, Path):
+			path = str(path)
 		javaFile = self.backend.newObject("java.io.File", path)
 		self.javaModelEvaluatorBuilder.load(javaFile)
 		return self
@@ -270,13 +274,14 @@ def make_backend(alias):
 		aliases = ["jpype", "pyjnius", "py4j"]
 		raise ValueError("Java backend alias {0} not in {1}".format(alias, aliases))
 
-def make_evaluator(path, backend = "jpype", lax = False, locatable = False, reporting = False, transpile = False):
+def make_evaluator(obj, backend = "jpype", lax = False, locatable = False, reporting = False, transpile = False):
 	""" Builds an Evaluator based on a PMML file.
 
 	Parameters:
 	----------
-	path: string
-		The path to the PMML file in local filesystem.
+	obj: string or bytes
+		The object to load. Either a path to a PMML file in local filesystem,
+		or a PMML string or byte array.
 
 	backend: JavaBackend or string
 		The Java backend or its alias
@@ -303,12 +308,26 @@ def make_evaluator(path, backend = "jpype", lax = False, locatable = False, repo
 		raise TypeError()
 
 	evaluatorBuilder = LoadingModelEvaluatorBuilder(backend, lax) \
-		.setLocatable(locatable) \
-		.loadFile(path)
+		.setLocatable(locatable)
+
+	if isinstance(obj, Path):
+		evaluatorBuilder = evaluatorBuilder.loadFile(obj)
+	elif isinstance(obj, str):
+		if len(obj) < 1024 and os.path.isfile(obj):
+			evaluatorBuilder = evaluatorBuilder.loadFile(obj)
+		else:
+			evaluatorBuilder = evaluatorBuilder.loadString(obj)
+	elif isinstance(obj, bytes):
+		evaluatorBuilder = evaluatorBuilder.loadBytes(obj)
+	else:
+		raise TypeError()
+
 	if reporting:
-		evaluatorBuilder.setReportingValueFactoryFactory()
+		evaluatorBuilder = evaluatorBuilder.setReportingValueFactoryFactory()
+
 	if transpile:
-		evaluatorBuilder.transpile(transpile if isinstance(transpile, str) else None)
+		evaluatorBuilder = evaluatorBuilder.transpile(transpile if isinstance(transpile, str) else None)
+
 	return evaluatorBuilder.build()
 
 def _package_data_jars(package_data_dir):

jpmml_evaluator/tests/__init__.py

@@ -1,11 +1,12 @@
 import os
 
+from pathlib import Path
 from unittest import TestCase
 
 import numpy
 import pandas
 
-from jpmml_evaluator import make_backend, make_evaluator, Evaluator, JavaError, LoadingModelEvaluatorBuilder
+from jpmml_evaluator import make_backend, make_evaluator, Evaluator, JavaError
 
 def _resource(name):
 	return os.path.join(os.path.dirname(__file__), "resources", name)
@@ -27,9 +28,15 @@ def workflow(self, backend):
 
 		self.assertIsInstance(resource, str)
 
-		evaluator = LoadingModelEvaluatorBuilder(backend) \
-			.loadFile(resource) \
-			.build()
+		evaluator = make_evaluator(resource, backend = backend)
+
+		self.assertIsInstance(evaluator, Evaluator)
+
+		resource_path = Path(resource)
+
+		self.assertIsInstance(resource_path, Path)
+
+		evaluator = make_evaluator(resource_path, backend = backend)
 
 		self.assertIsInstance(evaluator, Evaluator)
 
@@ -38,9 +45,7 @@ def workflow(self, backend):
 
 		self.assertIsInstance(resource_bytes, bytes)
 
-		evaluator = LoadingModelEvaluatorBuilder(backend) \
-			.loadBytes(resource_bytes) \
-			.build()
+		evaluator = make_evaluator(resource_bytes, backend = backend)
 
 		self.assertIsInstance(evaluator, Evaluator)
 
@@ -49,9 +54,7 @@ def workflow(self, backend):
 
 		self.assertIsInstance(resource_string, str)
 
-		evaluator = LoadingModelEvaluatorBuilder(backend) \
-			.loadString(resource_string) \
-			.build()
+		evaluator = make_evaluator(resource_string, backend = backend)
 
 		self.assertIsInstance(evaluator, Evaluator)