feature: git support for hosting models (#878)

* git integration for serving

Author: GaryTu1020

doc/overview.rst

@@ -799,6 +799,23 @@ After that, invoke the ``deploy()`` method on the ``Model``:
 
 This returns a predictor the same way an ``Estimator`` does when ``deploy()`` is called. You can now get inferences just like with any other model deployed on Amazon SageMaker.
 
+Git support is also available when you bring your own model, through which you can use inference scripts stored in your
+Git repositories. The process is similar to using Git support for training jobs. You can simply provide ``git_config``
+when create the ``Model`` object, and let ``entry_point``, ``source_dir`` and ``dependencies`` (if needed) be relative
+paths inside the Git repository:
+
+.. code:: python
+
+    git_config = {'repo': 'https://github.com/username/repo-with-training-scripts.git',
+                  'branch': 'branch1',
+                  'commit': '4893e528afa4a790331e1b5286954f073b0f14a2'}
+
+    sagemaker_model = MXNetModel(model_data='s3://path/to/model.tar.gz',
+                                role='arn:aws:iam::accid:sagemaker-role',
+                                entry_point='inference.py',
+                                source_dir='mxnet',
+                                git_config=git_config)
+
 A full example is available in the `Amazon SageMaker examples repository <https://github.com/awslabs/amazon-sagemaker-examples/tree/master/advanced_functionality/mxnet_mnist_byom>`__.
 
 You can also find this notebook in the **Advanced Functionality** section of the **SageMaker Examples** section in a notebook instance.

src/sagemaker/estimator.py

@@ -955,8 +955,8 @@ def __init__(
         code_location=None,
         image_name=None,
         dependencies=None,
-        git_config=None,
         enable_network_isolation=False,
+        git_config=None,
         **kwargs
     ):
         """Base class initializer. Subclasses which override ``__init__`` should invoke ``super()``
@@ -993,7 +993,7 @@ def __init__(
             source_dir (str): Path (absolute or relative) to a directory with any other training
                 source code dependencies aside from the entry point file (default: None). Structure within this
                 directory are preserved when training on Amazon SageMaker. If 'git_config' is provided,
-                source_dir should be a relative location to a directory in the Git repo.
+                'source_dir' should be a relative location to a directory in the Git repo.
                 Example:
 
                     With the following GitHub repo directory structure:
@@ -1023,6 +1023,8 @@ def __init__(
             dependencies (list[str]): A list of paths to directories (absolute or relative) with
                 any additional libraries that will be exported to the container (default: []).
                 The library folders will be copied to SageMaker in the same folder where the entrypoint is copied.
+                If 'git_config' is provided, 'dependencies' should be a list of relative locations to directories
+                with any additional libraries needed in the Git repo.
                 Example:
 
                     The following call
@@ -1085,12 +1087,12 @@ def _prepare_for_training(self, job_name=None):
         super(Framework, self)._prepare_for_training(job_name=job_name)
 
         if self.git_config:
-            updates = git_utils.git_clone_repo(
+            updated_paths = git_utils.git_clone_repo(
                 self.git_config, self.entry_point, self.source_dir, self.dependencies
             )
-            self.entry_point = updates["entry_point"]
-            self.source_dir = updates["source_dir"]
-            self.dependencies = updates["dependencies"]
+            self.entry_point = updated_paths["entry_point"]
+            self.source_dir = updated_paths["source_dir"]
+            self.dependencies = updated_paths["dependencies"]
 
         # validate source dir will raise a ValueError if there is something wrong with the
         # source directory. We are intentionally not handling it because this is a critical error.

src/sagemaker/git_utils.py

@@ -13,6 +13,7 @@
 from __future__ import absolute_import
 
 import os
+import six
 import subprocess
 import tempfile
 
@@ -39,35 +40,46 @@ def git_clone_repo(git_config, entry_point, source_dir=None, dependencies=None):
                                3. failed to checkout the required commit
         ValueError: If 1. entry point specified does not exist in the repo
                        2. source dir specified does not exist in the repo
+                       3. dependencies specified do not exist in the repo
+                       4. git_config is in bad format
 
     Returns:
         dict: A dict that contains the updated values of entry_point, source_dir and dependencies
     """
+    if entry_point is None:
+        raise ValueError("Please provide an entry point.")
     _validate_git_config(git_config)
     repo_dir = tempfile.mkdtemp()
     subprocess.check_call(["git", "clone", git_config["repo"], repo_dir])
 
     _checkout_branch_and_commit(git_config, repo_dir)
 
-    ret = {"entry_point": entry_point, "source_dir": source_dir, "dependencies": dependencies}
+    updated_paths = {
+        "entry_point": entry_point,
+        "source_dir": source_dir,
+        "dependencies": dependencies,
+    }
+
     # check if the cloned repo contains entry point, source directory and dependencies
     if source_dir:
         if not os.path.isdir(os.path.join(repo_dir, source_dir)):
             raise ValueError("Source directory does not exist in the repo.")
         if not os.path.isfile(os.path.join(repo_dir, source_dir, entry_point)):
             raise ValueError("Entry point does not exist in the repo.")
-        ret["source_dir"] = os.path.join(repo_dir, source_dir)
+        updated_paths["source_dir"] = os.path.join(repo_dir, source_dir)
     else:
-        if not os.path.isfile(os.path.join(repo_dir, entry_point)):
+        if os.path.isfile(os.path.join(repo_dir, entry_point)):
+            updated_paths["entry_point"] = os.path.join(repo_dir, entry_point)
+        else:
             raise ValueError("Entry point does not exist in the repo.")
-        ret["entry_point"] = os.path.join(repo_dir, entry_point)
 
-    ret["dependencies"] = []
+    updated_paths["dependencies"] = []
     for path in dependencies:
-        if not os.path.exists(os.path.join(repo_dir, path)):
+        if os.path.exists(os.path.join(repo_dir, path)):
+            updated_paths["dependencies"].append(os.path.join(repo_dir, path))
+        else:
             raise ValueError("Dependency {} does not exist in the repo.".format(path))
-        ret["dependencies"].append(os.path.join(repo_dir, path))
-    return ret
+    return updated_paths
 
 
 def _validate_git_config(git_config):
@@ -84,6 +96,13 @@ def _validate_git_config(git_config):
     """
     if "repo" not in git_config:
         raise ValueError("Please provide a repo for git_config.")
+    allowed_keys = ["repo", "branch", "commit"]
+    for key in allowed_keys:
+        if key in git_config and not isinstance(git_config[key], six.string_types):
+            raise ValueError("'{}' should be a string".format(key))
+    for key in git_config:
+        if key not in allowed_keys:
+            raise ValueError("Unexpected argument(s) provided for git_config!")
 
 
 def _checkout_branch_and_commit(git_config, repo_dir):
@@ -95,8 +114,8 @@ def _checkout_branch_and_commit(git_config, repo_dir):
         repo_dir (str): the directory where the repo is cloned
 
     Raises:
-        ValueError: If 1. entry point specified does not exist in the repo
-                       2. source dir specified does not exist in the repo
+        CalledProcessError: If 1. failed to checkout the required branch
+                               2. failed to checkout the required commit
     """
     if "branch" in git_config:
         subprocess.check_call(args=["git", "checkout", git_config["branch"]], cwd=str(repo_dir))

src/sagemaker/model.py

@@ -17,7 +17,7 @@
 import os
 
 import sagemaker
-from sagemaker import fw_utils, local, session, utils
+from sagemaker import fw_utils, local, session, utils, git_utils
 from sagemaker.fw_utils import UploadedCode
 from sagemaker.transformer import Transformer
 
@@ -494,6 +494,7 @@ def __init__(
         code_location=None,
         sagemaker_session=None,
         dependencies=None,
+        git_config=None,
         **kwargs
     ):
         """Initialize a ``FrameworkModel``.
@@ -504,15 +505,54 @@ def __init__(
             role (str): An IAM role name or ARN for SageMaker to access AWS resources on your behalf.
             entry_point (str): Path (absolute or relative) to the Python source file which should be executed
                 as the entry point to model hosting. This should be compatible with either Python 2.7 or Python 3.5.
+                If 'git_config' is provided, 'entry_point' should be a relative location to the Python source file in
+                the Git repo.
+                Example:
+
+                    With the following GitHub repo directory structure:
+
+                    >>> |----- README.md
+                    >>> |----- src
+                    >>>         |----- inference.py
+                    >>>         |----- test.py
+
+                    You can assign entry_point='src/inference.py'.
+            git_config (dict[str, str]): Git configurations used for cloning files, including 'repo', 'branch'
+                and 'commit' (default: None).
+                'branch' and 'commit' are optional. If 'branch' is not specified, 'master' branch will be used. If
+                'commit' is not specified, the latest commit in the required branch will be used.
+                Example:
+
+                    The following config:
+
+                    >>> git_config = {'repo': 'https://github.com/aws/sagemaker-python-sdk.git',
+                    >>>               'branch': 'test-branch-git-config',
+                    >>>               'commit': '329bfcf884482002c05ff7f44f62599ebc9f445a'}
+
+                    results in cloning the repo specified in 'repo', then checkout the 'master' branch, and checkout
+                    the specified commit.
             source_dir (str): Path (absolute or relative) to a directory with any other training
                 source code dependencies aside from the entry point file (default: None). Structure within this
-                directory will be preserved when training on SageMaker.
-                If the directory points to S3, no code will be uploaded and the S3 location will be used instead.
+                directory will be preserved when training on SageMaker. If 'git_config' is provided,
+                'source_dir' should be a relative location to a directory in the Git repo. If the directory points
+                to S3, no code will be uploaded and the S3 location will be used instead.
+                Example:
+
+                    With the following GitHub repo directory structure:
+
+                    >>> |----- README.md
+                    >>> |----- src
+                    >>>         |----- inference.py
+                    >>>         |----- test.py
+
+                    You can assign entry_point='inference.py', source_dir='src'.
             dependencies (list[str]): A list of paths to directories (absolute or relative) with
                 any additional libraries that will be exported to the container (default: []).
                 The library folders will be copied to SageMaker in the same folder where the entrypoint is copied.
-                If the ```source_dir``` points to S3, code will be uploaded and the S3 location will be used
-                instead. Example:
+                If 'git_config' is provided, 'dependencies' should be a list of relative locations to directories
+                with any additional libraries needed in the Git repo. If the ```source_dir``` points to S3, code
+                will be uploaded and the S3 location will be used instead.
+                Example:
 
                     The following call
                     >>> Estimator(entry_point='train.py', dependencies=['my/libs/common', 'virtual-env'])
@@ -554,12 +594,20 @@ def __init__(
         self.entry_point = entry_point
         self.source_dir = source_dir
         self.dependencies = dependencies or []
+        self.git_config = git_config
         self.enable_cloudwatch_metrics = enable_cloudwatch_metrics
         self.container_log_level = container_log_level
         if code_location:
             self.bucket, self.key_prefix = fw_utils.parse_s3_url(code_location)
         else:
             self.bucket, self.key_prefix = None, None
+        if self.git_config:
+            updates = git_utils.git_clone_repo(
+                self.git_config, self.entry_point, self.source_dir, self.dependencies
+            )
+            self.entry_point = updates["entry_point"]
+            self.source_dir = updates["source_dir"]
+            self.dependencies = updates["dependencies"]
         self.uploaded_code = None
         self.repacked_model_data = None
 

tests/integ/test_git.py

@@ -21,11 +21,14 @@
 from tests.integ import lock as lock
 from sagemaker.mxnet.estimator import MXNet
 from sagemaker.pytorch.estimator import PyTorch
+from sagemaker.sklearn.estimator import SKLearn
+from sagemaker.mxnet.model import MXNetModel
+from sagemaker.sklearn.model import SKLearnModel
 from tests.integ import DATA_DIR, PYTHON_VERSION
 
 GIT_REPO = "https://github.com/aws/sagemaker-python-sdk.git"
 BRANCH = "test-branch-git-config"
-COMMIT = "329bfcf884482002c05ff7f44f62599ebc9f445a"
+COMMIT = "ae15c9d7d5b97ea95ea451e4662ee43da3401d73"
 
 # endpoint tests all use the same port, so we use this lock to prevent concurrent execution
 LOCK_PATH = os.path.join(tempfile.gettempdir(), "sagemaker_test_git_lock")
@@ -62,15 +65,16 @@ def test_git_support_with_pytorch(sagemaker_local_session):
 
 
 @pytest.mark.local_mode
-def test_git_support_with_mxnet(sagemaker_local_session, mxnet_full_version):
+def test_git_support_with_mxnet(sagemaker_local_session):
     script_path = "mnist.py"
     data_path = os.path.join(DATA_DIR, "mxnet_mnist")
     git_config = {"repo": GIT_REPO, "branch": BRANCH, "commit": COMMIT}
+    source_dir = "mxnet"
     dependencies = ["foo/bar.py"]
     mx = MXNet(
         entry_point=script_path,
         role="SageMakerRole",
-        source_dir="mxnet",
+        source_dir=source_dir,
         dependencies=dependencies,
         framework_version=MXNet.LATEST_VERSION,
         py_version=PYTHON_VERSION,
@@ -94,10 +98,76 @@ def test_git_support_with_mxnet(sagemaker_local_session, mxnet_full_version):
 
     with lock.lock(LOCK_PATH):
         try:
-            predictor = mx.deploy(initial_instance_count=1, instance_type="local")
+            serving_script_path = "mnist_hosting_with_custom_handlers.py"
+            client = sagemaker_local_session.sagemaker_client
+            desc = client.describe_training_job(TrainingJobName=mx.latest_training_job.name)
+            model_data = desc["ModelArtifacts"]["S3ModelArtifacts"]
+            model = MXNetModel(
+                model_data,
+                "SageMakerRole",
+                entry_point=serving_script_path,
+                source_dir=source_dir,
+                dependencies=dependencies,
+                py_version=PYTHON_VERSION,
+                sagemaker_session=sagemaker_local_session,
+                framework_version=MXNet.LATEST_VERSION,
+                git_config=git_config,
+            )
+            predictor = model.deploy(initial_instance_count=1, instance_type="local")
 
             data = numpy.zeros(shape=(1, 1, 28, 28))
             result = predictor.predict(data)
             assert result is not None
         finally:
             predictor.delete_endpoint()
+
+
+@pytest.mark.skipif(PYTHON_VERSION != "py3", reason="Scikit-learn image supports only python 3.")
+@pytest.mark.local_mode
+def test_git_support_with_sklearn(sagemaker_local_session, sklearn_full_version):
+    script_path = "mnist.py"
+    data_path = os.path.join(DATA_DIR, "sklearn_mnist")
+    git_config = {
+        "repo": "https://github.com/GaryTu1020/python-sdk-testing.git",
+        "branch": "branch1",
+        "commit": "aafa4e96237dd78a015d5df22bfcfef46845c3c5",
+    }
+    source_dir = "sklearn"
+    sklearn = SKLearn(
+        entry_point=script_path,
+        role="SageMakerRole",
+        source_dir=source_dir,
+        py_version=PYTHON_VERSION,
+        train_instance_count=1,
+        train_instance_type="local",
+        sagemaker_session=sagemaker_local_session,
+        framework_version=sklearn_full_version,
+        hyperparameters={"epochs": 1},
+        git_config=git_config,
+    )
+    train_input = "file://" + os.path.join(data_path, "train")
+    test_input = "file://" + os.path.join(data_path, "test")
+    sklearn.fit({"train": train_input, "test": test_input})
+
+    assert os.path.isdir(sklearn.source_dir)
+
+    with lock.lock(LOCK_PATH):
+        try:
+            client = sagemaker_local_session.sagemaker_client
+            desc = client.describe_training_job(TrainingJobName=sklearn.latest_training_job.name)
+            model_data = desc["ModelArtifacts"]["S3ModelArtifacts"]
+            model = SKLearnModel(
+                model_data,
+                "SageMakerRole",
+                entry_point=script_path,
+                source_dir=source_dir,
+                sagemaker_session=sagemaker_local_session,
+                git_config=git_config,
+            )
+            predictor = model.deploy(1, "local")
+
+            data = numpy.zeros((100, 784), dtype="float32")
+            result = predictor.predict(data)
+            assert result is not None
+        finally:
+            predictor.delete_endpoint()

tests/unit/test_estimator.py

@@ -50,7 +50,7 @@
 OUTPUT_PATH = "s3://bucket/prefix"
 GIT_REPO = "https://github.com/aws/sagemaker-python-sdk.git"
 BRANCH = "test-branch-git-config"
-COMMIT = "329bfcf884482002c05ff7f44f62599ebc9f445a"
+COMMIT = "ae15c9d7d5b97ea95ea451e4662ee43da3401d73"
 
 DESCRIBE_TRAINING_JOB_RESULT = {"ModelArtifacts": {"S3ModelArtifacts": MODEL_DATA}}
 INSTANCE_TYPE = "c4.4xlarge"
@@ -898,12 +898,12 @@ def test_git_support_bad_repo_url_format(sagemaker_session):
 
 
 @patch(
-    "subprocess.check_call",
+    "sagemaker.git_utils.git_clone_repo",
     side_effect=subprocess.CalledProcessError(
-        returncode=1, cmd="git clone https://github.com/aws/no-such-repo.git"
+        returncode=1, cmd="git clone https://github.com/aws/no-such-repo.git /tmp/repo_dir"
     ),
 )
-def test_git_support_git_clone_fail(check_call, sagemaker_session):
+def test_git_support_git_clone_fail(sagemaker_session):
     git_config = {"repo": "https://github.com/aws/no-such-repo.git", "branch": BRANCH}
     fw = DummyFramework(
         entry_point="entry_point",