Remove wait time when stopping and starting torchserve in tests

Make sanity test run with pytest 90%

Added missing test for snapshotting

Use pytest tests in torchserve sanity checks

Author: mreso

test/pytest/sanity/conftest.py

@@ -0,0 +1,27 @@
+import json
+import sys
+from pathlib import Path
+
+import pytest
+
+REPO_ROOT = Path(__file__).parents[3]
+
+
+MAR_CONFIG = REPO_ROOT.joinpath("ts_scripts", "mar_config.json")
+
+
+@pytest.fixture(name="gen_models", scope="module")
+def load_gen_models() -> dict:
+    with open(MAR_CONFIG) as f:
+        models = json.load(f)
+    models = {m["model_name"]: m for m in models}
+    return models
+
+
+@pytest.fixture(scope="module")
+def ts_scripts_path():
+    sys.path.append(REPO_ROOT.as_posix())
+
+    yield
+
+    sys.path.pop()

test/pytest/sanity/test_config_snapshot.py

@@ -0,0 +1,54 @@
+import json
+from pathlib import Path
+
+import pytest
+import test_utils
+
+REPO_ROOT = Path(__file__).parents[3]
+SANITY_MODELS_CONFIG = REPO_ROOT.joinpath("ts_scripts", "configs", "sanity_models.json")
+
+
+def load_resnet18() -> dict:
+    with open(SANITY_MODELS_CONFIG) as f:
+        models = json.load(f)
+    return list(filter(lambda x: x["name"] == "resnet-18", models))[0]
+
+
+@pytest.fixture(name="resnet18")
+def generate_resnet18(model_store, gen_models, ts_scripts_path):
+    model = load_resnet18()
+
+    from ts_scripts.marsgen import generate_model
+
+    generate_model(gen_models[model["name"]], model_store)
+
+    yield model
+
+
+@pytest.fixture(scope="module")
+def torchserve_with_snapshot(model_store):
+    test_utils.torchserve_cleanup()
+
+    test_utils.start_torchserve(
+        model_store=model_store, no_config_snapshots=False, gen_mar=False
+    )
+
+    yield
+
+    test_utils.torchserve_cleanup()
+
+
+def test_config_snapshotting(
+    resnet18, model_store, torchserve_with_snapshot, ts_scripts_path
+):
+    from ts_scripts.sanity_utils import run_rest_test
+
+    run_rest_test(resnet18, unregister_model=False)
+
+    test_utils.stop_torchserve()
+
+    test_utils.start_torchserve(
+        model_store=model_store, no_config_snapshots=False, gen_mar=False
+    )
+
+    run_rest_test(resnet18, register_model=False)

test/pytest/sanity/test_model_registering.py

@@ -0,0 +1,55 @@
+import json
+from pathlib import Path
+
+import pytest
+
+REPO_ROOT = Path(__file__).parents[3]
+SANITY_MODELS_CONFIG = REPO_ROOT.joinpath("ts_scripts", "configs", "sanity_models.json")
+
+
+@pytest.fixture(scope="module")
+def grpc_client_stubs(ts_scripts_path):
+    from ts_scripts.shell_utils import rm_file
+    from ts_scripts.tsutils import generate_grpc_client_stubs
+
+    generate_grpc_client_stubs()
+
+    yield
+
+    rm_file(REPO_ROOT.joinpath("ts_scripts", "*_pb2*.py").as_posix(), True)
+
+
+def load_models() -> dict:
+    with open(SANITY_MODELS_CONFIG) as f:
+        models = json.load(f)
+    return models
+
+
+@pytest.fixture(name="model", params=load_models(), scope="module")
+def models_to_validate(request, model_store, gen_models, ts_scripts_path):
+    model = request.param
+
+    if model["name"] in gen_models:
+        from ts_scripts.marsgen import generate_model
+
+        generate_model(gen_models[model["name"]], model_store)
+
+    yield model
+
+
+def test_models_with_grpc(model, torchserve, ts_scripts_path, grpc_client_stubs):
+    from ts_scripts.sanity_utils import run_grpc_test
+
+    run_grpc_test(model)
+
+
+def test_models_with_rest(model, torchserve, ts_scripts_path):
+    from ts_scripts.sanity_utils import run_rest_test
+
+    run_rest_test(model)
+
+
+def test_gpu_setup(ts_scripts_path):
+    from ts_scripts.sanity_utils import test_gpu_setup
+
+    test_gpu_setup()

test/pytest/test_utils.py

@@ -135,6 +135,7 @@ def model_archiver_command_builder(
     extra_files=None,
     force=False,
     config_file=None,
+    archive_format=None,
 ):
     # Initialize a list to store the command-line arguments
     cmd_parts = ["torch-model-archiver"]
@@ -164,9 +165,11 @@ def model_archiver_command_builder(
     if force:
         cmd_parts.append("--force")
 
-    # Append the export-path argument to the list
-    cmd_parts.append(f"--export-path {MODEL_STORE}")
+    if archive_format:
+        cmd += " --archive-format {0}".foramt(archive_format)
 
+    cmd += " --export-path {0}".format(MODEL_STORE)
+    
     # Convert the list into a string to represent the complete command
     cmd = " ".join(cmd_parts)
 

ts_scripts/marsgen.py

@@ -43,6 +43,64 @@ def gen_mar(model_store=None):
                 print(f"## Symlink {src}, {dst} successfully.")
 
 
+def generate_model(model, model_store_dir):
+    serialized_file_path = None
+    if model.get("serialized_file_remote", None):
+        if model.get("gen_scripted_file_path", None):
+            subprocess.run(["python", model["gen_scripted_file_path"]])
+        else:
+            serialized_model_file_url = (
+                f"https://download.pytorch.org/models/{model['serialized_file_remote']}"
+            )
+            urllib.request.urlretrieve(
+                serialized_model_file_url,
+                f'{model_store_dir}/{model["serialized_file_remote"]}',
+            )
+        serialized_file_path = os.path.join(
+            model_store_dir, model["serialized_file_remote"]
+        )
+    elif model.get("serialized_file_local", None):
+        serialized_file_path = model["serialized_file_local"]
+
+    handler = model.get("handler", None)
+
+    extra_files = model.get("extra_files", None)
+
+    runtime = model.get("runtime", None)
+
+    archive_format = model.get("archive_format", "zip-store")
+
+    requirements_file = model.get("requirements_file", None)
+
+    export_path = model.get("export_path", model_store_dir)
+
+    cmd = model_archiver_command_builder(
+        model["model_name"],
+        model["version"],
+        model["model_file"],
+        serialized_file_path,
+        handler,
+        extra_files,
+        runtime,
+        archive_format,
+        requirements_file,
+        export_path,
+    )
+    print(f"## In directory: {os.getcwd()} | Executing command: {cmd}\n")
+    try:
+        subprocess.check_call(cmd, shell=True)
+        marfile = "{}.mar".format(model["model_name"])
+        print("## {} is generated.\n".format(marfile))
+        mar_set.add(marfile)
+    except subprocess.CalledProcessError as exc:
+        print("## {} creation failed !, error: {}\n".format(model["model_name"], exc))
+
+    if model.get("serialized_file_remote", None) and os.path.exists(
+        serialized_file_path
+    ):
+        os.remove(serialized_file_path)
+
+
 def generate_mars(mar_config=MAR_CONFIG_FILE_PATH, model_store_dir=MODEL_STORE_DIR):
     """
     By default generate_mars reads ts_scripts/mar_config.json and outputs mar files in dir model_store_gen
@@ -67,72 +125,7 @@ def generate_mars(mar_config=MAR_CONFIG_FILE_PATH, model_store_dir=MODEL_STORE_D
         models = json.loads(f.read())
 
         for model in models:
-            serialized_file_path = None
-            if model.get("serialized_file_remote") and model["serialized_file_remote"]:
-                if (
-                    model.get("gen_scripted_file_path")
-                    and model["gen_scripted_file_path"]
-                ):
-                    subprocess.run(["python", model["gen_scripted_file_path"]])
-                else:
-                    serialized_model_file_url = (
-                        "https://download.pytorch.org/models/{}".format(
-                            model["serialized_file_remote"]
-                        )
-                    )
-                    urllib.request.urlretrieve(
-                        serialized_model_file_url,
-                        f'{model_store_dir}/{model["serialized_file_remote"]}',
-                    )
-                serialized_file_path = os.path.join(
-                    model_store_dir, model["serialized_file_remote"]
-                )
-            elif model.get("serialized_file_local") and model["serialized_file_local"]:
-                serialized_file_path = model["serialized_file_local"]
-
-            handler = model.get("handler", None)
-
-            extra_files = model.get("extra_files", None)
-
-            runtime = model.get("runtime", None)
-
-            archive_format = model.get("archive_format", "zip-store")
-
-            requirements_file = model.get("requirements_file", None)
-
-            export_path = model.get("export_path", model_store_dir)
-
-            cmd = model_archiver_command_builder(
-                model["model_name"],
-                model["version"],
-                model["model_file"],
-                serialized_file_path,
-                handler,
-                extra_files,
-                runtime,
-                archive_format,
-                requirements_file,
-                export_path,
-            )
-            print(f"## In directory: {os.getcwd()} | Executing command: {cmd}\n")
-            try:
-                subprocess.check_call(cmd, shell=True)
-                marfile = "{}.mar".format(model["model_name"])
-                print("## {} is generated.\n".format(marfile))
-                mar_set.add(marfile)
-            except subprocess.CalledProcessError as exc:
-                print(
-                    "## {} creation failed !, error: {}\n".format(
-                        model["model_name"], exc
-                    )
-                )
-
-            if (
-                model.get("serialized_file_remote")
-                and model["serialized_file_remote"]
-                and os.path.exists(serialized_file_path)
-            ):
-                os.remove(serialized_file_path)
+            generate_model(model, model_store_dir)
     os.chdir(cwd)
 
 

ts_scripts/sanity_utils.py

@@ -10,7 +10,6 @@
 from ts_scripts import marsgen as mg
 from ts_scripts import tsutils as ts
 from ts_scripts import utils
-from ts_scripts.tsutils import generate_grpc_client_stubs
 
 REPO_ROOT = os.path.join(os.path.dirname(os.path.abspath(__file__)), "..")
 sys.path.append(REPO_ROOT)
@@ -163,51 +162,18 @@ def run_rest_test(model, register_model=True, unregister_model=True):
 
 
 def test_sanity():
-    generate_grpc_client_stubs()
-
-    print("## Started sanity tests")
-
-    models_to_validate = load_model_to_validate()
-
-    test_gpu_setup()
-
-    ts_log_file = os.path.join("logs", "ts_console.log")
-
-    os.makedirs("model_store", exist_ok=True)
-    os.makedirs("logs", exist_ok=True)
-
-    mg.mar_set = set(os.listdir("model_store"))
-    started = ts.start_torchserve(log_file=ts_log_file, gen_mar=False)
-    if not started:
-        sys.exit(1)
-
-    resnet18_model = models_to_validate["resnet-18"]
-
-    models_to_validate = {
-        k: v for k, v in models_to_validate.items() if k != "resnet-18"
-    }
-
-    for _, model in models_to_validate.items():
-        run_grpc_test(model)
-        run_rest_test(model)
-
-    run_rest_test(resnet18_model, unregister_model=False)
-
-    stopped = ts.stop_torchserve()
-    if not stopped:
-        sys.exit(1)
-
-    # Restarting torchserve
-    # This should restart with the generated snapshot and resnet-18 model should be automatically registered
-    started = ts.start_torchserve(log_file=ts_log_file, gen_mar=False)
-    if not started:
-        sys.exit(1)
-
-    run_rest_test(resnet18_model, register_model=False)
-
-    stopped = ts.stop_torchserve()
-    if not stopped:
-        sys.exit(1)
+    # Execute python tests
+    print("## Started TorchServe sanity pytests")
+    test_dir = os.path.join("test", "pytest", "sanity")
+    coverage_dir = os.path.join("ts")
+    report_output_dir = os.path.join(test_dir, "coverage.xml")
+
+    ts_test_cmd = f"python -m pytest --cov-report xml:{report_output_dir} --cov={coverage_dir} {test_dir}"
+    print(f"## In directory: {os.getcwd()} | Executing command: {ts_test_cmd}")
+    ts_test_error_code = os.system(ts_test_cmd)
+
+    if ts_test_error_code != 0:
+        sys.exit("## TorchServe sanity test failed !")
 
 
 def test_workflow_sanity():