Merge pull request #8 from truefoundry/fix-deploy-model

fix: improve deploy-ml-model flow

Author: debajyoti-truefoundry

deploy-ml-model/app.py

@@ -5,7 +5,8 @@
 
 model = joblib.load("iris_classifier.joblib")
 
-app = FastAPI()
+app = FastAPI(docs_url="/", root_path=os.getenv("TFY_SERVICE_ROOT_PATH", "/"))
+
 
 @app.post("/predict")
 def predict(
@@ -18,4 +19,5 @@ def predict(
         petal_width=petal_width,
     )
     prediction = int(model.predict(pd.DataFrame([data]))[0])
-    return {"prediction": prediction}
+    return {"prediction": prediction}
+

deploy-ml-model/deploy.py

@@ -1,23 +1,42 @@
+import argparse
 import logging
-from servicefoundry import Build, PythonBuild, Service, Resources
+from servicefoundry import Build, PythonBuild, Service, Resources, Port
 
 logging.basicConfig(level=logging.INFO)
 
+parser = argparse.ArgumentParser()
+parser.add_argument("--name", required=True, type=str, help="Name of the application.")
+parser.add_argument(
+    "--workspace_fqn",
+    required=True,
+    type=str,
+    help="FQN of the workspace where application will be deployed.",
+)
+parser.add_argument(
+    "--host",
+    required=True,
+    type=str,
+    help="Host where the application will be available for access. Ex:- my-app.my-org.com",
+)
+args = parser.parse_args()
+
 image = Build(
     build_spec=PythonBuild(
         command="uvicorn app:app --port 8000 --host 0.0.0.0",
-        requirements_path="requirements.txt"
+        requirements_path="requirements.txt",
     )
 )
 
 service = Service(
-    name="ml-deploy",
+    name=args.name,
     image=image,
-    ports=[{"port": 8000}],
-    resources=Resources(memory_limit=1500, memory_request=1000),
-    env={
-        "UVICORN_WEB_CONCURRENCY": "1",
-        "ENVIRONMENT": "dev"
-    }
+    ports=[Port(port=8000, host=args.host)],
+    resources=Resources(
+        cpu_request=0.1,
+        cpu_limit=0.1,
+        memory_request=500,
+        memory_limit=500,
+    ),
+    env={"UVICORN_WEB_CONCURRENCY": "1", "ENVIRONMENT": "dev"},
 )
-service.deploy(workspace_fqn="YOUR_WORKSPACE_FQN")
+service.deploy(workspace_fqn=args.workspace_fqn)

deploy-ml-model/requirements.txt

@@ -1,5 +1,5 @@
 fastapi==0.81.0
 uvicorn==0.18.3
 scikit-learn==1.0.2
-joblib
-pandas
+joblib==1.3.2
+pandas==2.1.0