gmicloud 0.1.2__tar.gz → 0.1.4__tar.gz

{gmicloud-0.1.2 → gmicloud-0.1.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: gmicloud
-Version: 0.1.2
+Version: 0.1.4
 Summary: GMI Cloud Python SDK
 Author-email: GMI <gmi@gmitec.net>
 License: MIT
@@ -10,10 +10,13 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.6
 Description-Content-Type: text/markdown
 
-# GMICloud SDK
+# GMICloud SDK (Beta)
 
 ## Overview
 
+Before you start: Our service and GPU resource is currenly invite-only so please contact our team (
+getstarted@gmicloud.ai) to get invited if you don't have one yet.
+
 The GMI Inference Engine SDK provides a Python interface for deploying and managing machine learning models in
 production environments. It allows users to create model artifacts, schedule tasks for serving models, and call
 inference APIs easily.
@@ -62,7 +65,18 @@ client = Client(client_id="<YOUR_CLIENT_ID>", email="<YOUR_EMAIL>", password="<Y
 
 ## Quick Start
 
-### 1. Create a Task from an Artifact Template
+### 1. How to run the code in the example folder
+```bash
+cd path/to/gmicloud-sdk
+# Create a virtual environment
+python -m venv venv
+source venv/bin/activate
+
+pip install -r requirements.txt
+python -m examples.<example_name>
+```
+
+### 2. Create a Task from an Artifact Template
 
 This is the simplest example to deploy an existing artifact template:
 
@@ -91,24 +105,30 @@ response = call_chat_completion(client, task.task_id)
 print(response)
 ```
 
-### 2. Step-by-Step Example: Create Artifact, Task, and Query the Endpoint
+### 3. Step-by-Step Example: Create Artifact, Task, and Query the Endpoint
 
 #### (a) Create an Artifact from a Template
 
 First, you’ll retrieve all templates and create an artifact based on the desired template (e.g., "Llama3.1 8B"):
 
 ```python
-def create_artifact_from_template(client):
+from gmicloud import *
+
+def create_artifact_from_template(client: Client) -> str:
     artifact_manager = client.artifact_manager
 
-    # List all available templates
+    # Get all artifact templates
     templates = artifact_manager.get_artifact_templates()
     for template in templates:
         if template.artifact_template_id == "qwen_2.5_14b_instruct_template_001":
-            return artifact_manager.create_artifact_from_template(
-                artifact_template_id=template.artifact_template_id
+            # Create an artifact from a template
+            artifact_id = artifact_manager.create_artifact_from_template(
+                artifact_template_id=template.artifact_template_id,
             )
-    return None
+
+            return artifact_id
+
+    return ""
 ```
 
 #### (b) Create a Task from the Artifact
@@ -116,43 +136,55 @@ def create_artifact_from_template(client):
 Wait until the artifact becomes "ready" and then deploy it using task scheduling:
 
 ```python
-def create_task_and_start(client, artifact_id):
-    artifact_manager = client.artifact_manager
+from gmicloud import *
+import time
+from datetime import datetime
 
-    # Wait until the artifact is ready
+def create_task_and_start(client: Client, artifact_id: str) -> str:
+    artifact_manager = client.artifact_manager
+    # Wait for the artifact to be ready
     while True:
-        artifact = artifact_manager.get_artifact(artifact_id)
-        if artifact.build_status == "SUCCESS":
-            break
-        print("Waiting for artifact to be ready...")
+        try:
+            artifact = artifact_manager.get_artifact(artifact_id)
+            print(f"Artifact status: {artifact.build_status}")
+            # Wait until the artifact is ready
+            if artifact.build_status == BuildStatus.SUCCESS:
+                break
+        except Exception as e:
+            raise e
+        # Wait for 2 seconds
         time.sleep(2)
-
-    # Configure and start the task
-    task_manager = client.task_manager
-    task = task_manager.create_task(Task(
-        config=TaskConfig(
-            ray_task_config=RayTaskConfig(
-                ray_version="2.40.0-py310-gpu",
-                file_path="serve",
-                artifact_id=artifact_id,
-                deployment_name="app",
-                replica_resource=ReplicaResource(
-                    cpu=10,
-                    ram_gb=100,
-                    gpu=1,
+    try:
+        task_manager = client.task_manager
+        # Create a task
+        task = task_manager.create_task(Task(
+            config=TaskConfig(
+                ray_task_config=RayTaskConfig(
+                    ray_version="2.40.0-py310-gpu",
+                    file_path="serve",
+                    artifact_id=artifact_id,
+                    deployment_name="app",
+                    replica_resource=ReplicaResource(
+                        cpu=10,
+                        ram_gb=100,
+                        gpu=1,
+                    ),
+                ),
+                task_scheduling=TaskScheduling(
+                    scheduling_oneoff=OneOffScheduling(
+                        trigger_timestamp=int(datetime.now().timestamp()) + 10,
+                        min_replicas=1,
+                        max_replicas=10,
+                    )
                 ),
             ),
-            task_scheduling=TaskScheduling(
-                scheduling_oneoff=OneOffScheduling(
-                    trigger_timestamp=int(datetime.now().timestamp()) + 10,
-                    min_replicas=1,
-                    max_replicas=10,
-                )
-            ),
-        ),
-    ))
+        ))
+
+        # Start the task
+        task_manager.start_task(task.task_id)
+    except Exception as e:
+        raise e
 
-    task_manager.start_task(task.task_id)
     return task.task_id
 ```
 
@@ -161,14 +193,20 @@ def create_task_and_start(client, artifact_id):
 Once the task is running, use the endpoint for inference:
 
 ```python
+from gmicloud import *
 from examples.completion import call_chat_completion
 
-client = Client()
-artifact_id = create_artifact_from_template(client)
-task_id = create_task_and_start(client, artifact_id)
+# Initialize the Client
+cli = Client()
 
-response = call_chat_completion(client, task_id)
-print(response)
+# Create an artifact from a template
+artifact_id = create_artifact_from_template(cli)
+
+# Create a task and start it
+task_id = create_task_and_start(cli, artifact_id)
+
+# Call chat completion
+print(call_chat_completion(cli, task_id))
 ```
 
 ## API Reference

{gmicloud-0.1.2 → gmicloud-0.1.4}/README.md

@@ -1,7 +1,10 @@
-# GMICloud SDK
+# GMICloud SDK (Beta)
 
 ## Overview
 
+Before you start: Our service and GPU resource is currenly invite-only so please contact our team (
+getstarted@gmicloud.ai) to get invited if you don't have one yet.
+
 The GMI Inference Engine SDK provides a Python interface for deploying and managing machine learning models in
 production environments. It allows users to create model artifacts, schedule tasks for serving models, and call
 inference APIs easily.
@@ -50,7 +53,18 @@ client = Client(client_id="<YOUR_CLIENT_ID>", email="<YOUR_EMAIL>", password="<Y
 
 ## Quick Start
 
-### 1. Create a Task from an Artifact Template
+### 1. How to run the code in the example folder
+```bash
+cd path/to/gmicloud-sdk
+# Create a virtual environment
+python -m venv venv
+source venv/bin/activate
+
+pip install -r requirements.txt
+python -m examples.<example_name>
+```
+
+### 2. Create a Task from an Artifact Template
 
 This is the simplest example to deploy an existing artifact template:
 
@@ -79,24 +93,30 @@ response = call_chat_completion(client, task.task_id)
 print(response)
 ```
 
-### 2. Step-by-Step Example: Create Artifact, Task, and Query the Endpoint
+### 3. Step-by-Step Example: Create Artifact, Task, and Query the Endpoint
 
 #### (a) Create an Artifact from a Template
 
 First, you’ll retrieve all templates and create an artifact based on the desired template (e.g., "Llama3.1 8B"):
 
 ```python
-def create_artifact_from_template(client):
+from gmicloud import *
+
+def create_artifact_from_template(client: Client) -> str:
     artifact_manager = client.artifact_manager
 
-    # List all available templates
+    # Get all artifact templates
     templates = artifact_manager.get_artifact_templates()
     for template in templates:
         if template.artifact_template_id == "qwen_2.5_14b_instruct_template_001":
-            return artifact_manager.create_artifact_from_template(
-                artifact_template_id=template.artifact_template_id
+            # Create an artifact from a template
+            artifact_id = artifact_manager.create_artifact_from_template(
+                artifact_template_id=template.artifact_template_id,
             )
-    return None
+
+            return artifact_id
+
+    return ""
 ```
 
 #### (b) Create a Task from the Artifact
@@ -104,43 +124,55 @@ def create_artifact_from_template(client):
 Wait until the artifact becomes "ready" and then deploy it using task scheduling:
 
 ```python
-def create_task_and_start(client, artifact_id):
-    artifact_manager = client.artifact_manager
+from gmicloud import *
+import time
+from datetime import datetime
 
-    # Wait until the artifact is ready
+def create_task_and_start(client: Client, artifact_id: str) -> str:
+    artifact_manager = client.artifact_manager
+    # Wait for the artifact to be ready
     while True:
-        artifact = artifact_manager.get_artifact(artifact_id)
-        if artifact.build_status == "SUCCESS":
-            break
-        print("Waiting for artifact to be ready...")
+        try:
+            artifact = artifact_manager.get_artifact(artifact_id)
+            print(f"Artifact status: {artifact.build_status}")
+            # Wait until the artifact is ready
+            if artifact.build_status == BuildStatus.SUCCESS:
+                break
+        except Exception as e:
+            raise e
+        # Wait for 2 seconds
         time.sleep(2)
-
-    # Configure and start the task
-    task_manager = client.task_manager
-    task = task_manager.create_task(Task(
-        config=TaskConfig(
-            ray_task_config=RayTaskConfig(
-                ray_version="2.40.0-py310-gpu",
-                file_path="serve",
-                artifact_id=artifact_id,
-                deployment_name="app",
-                replica_resource=ReplicaResource(
-                    cpu=10,
-                    ram_gb=100,
-                    gpu=1,
+    try:
+        task_manager = client.task_manager
+        # Create a task
+        task = task_manager.create_task(Task(
+            config=TaskConfig(
+                ray_task_config=RayTaskConfig(
+                    ray_version="2.40.0-py310-gpu",
+                    file_path="serve",
+                    artifact_id=artifact_id,
+                    deployment_name="app",
+                    replica_resource=ReplicaResource(
+                        cpu=10,
+                        ram_gb=100,
+                        gpu=1,
+                    ),
+                ),
+                task_scheduling=TaskScheduling(
+                    scheduling_oneoff=OneOffScheduling(
+                        trigger_timestamp=int(datetime.now().timestamp()) + 10,
+                        min_replicas=1,
+                        max_replicas=10,
+                    )
                 ),
             ),
-            task_scheduling=TaskScheduling(
-                scheduling_oneoff=OneOffScheduling(
-                    trigger_timestamp=int(datetime.now().timestamp()) + 10,
-                    min_replicas=1,
-                    max_replicas=10,
-                )
-            ),
-        ),
-    ))
+        ))
+
+        # Start the task
+        task_manager.start_task(task.task_id)
+    except Exception as e:
+        raise e
 
-    task_manager.start_task(task.task_id)
     return task.task_id
 ```
 
@@ -149,14 +181,20 @@ def create_task_and_start(client, artifact_id):
 Once the task is running, use the endpoint for inference:
 
 ```python
+from gmicloud import *
 from examples.completion import call_chat_completion
 
-client = Client()
-artifact_id = create_artifact_from_template(client)
-task_id = create_task_and_start(client, artifact_id)
+# Initialize the Client
+cli = Client()
 
-response = call_chat_completion(client, task_id)
-print(response)
+# Create an artifact from a template
+artifact_id = create_artifact_from_template(cli)
+
+# Create a task and start it
+task_id = create_task_and_start(cli, artifact_id)
+
+# Call chat completion
+print(call_chat_completion(cli, task_id))
 ```
 
 ## API Reference
@@ -197,4 +235,4 @@ We welcome contributions to enhance the SDK. Please follow these steps:
 1. Fork the repository.
 2. Create a new branch for your feature or bugfix.
 3. Commit changes with clear messages.
-4. Submit a pull request for review.
+4. Submit a pull request for review.

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud/_internal/_client/_artifact_client.py

@@ -5,7 +5,6 @@ from ._iam_client import IAMClient
 from ._decorator import handle_refresh_token
 from .._models import *
 from .._config import ARTIFACT_SERVICE_BASE_URL
-from .._constants import ACCESS_TOKEN_HEADER, CLIENT_ID_HEADER
 
 
 class ArtifactClient:
@@ -33,11 +32,7 @@ class ArtifactClient:
         :return: The Artifact object.
         :rtype: Artifact
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.get(f"/get_artifact", custom_headers, {"artifact_id": artifact_id})
+        result = self.client.get(f"/get_artifact", self.iam_client.get_custom_headers(), {"artifact_id": artifact_id})
 
         return Artifact.model_validate(result)
 
@@ -49,11 +44,7 @@ class ArtifactClient:
         :return: A list of Artifact objects.
         :rtype: List[Artifact]
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.get("/get_all_artifacts", custom_headers)
+        result = self.client.get("/get_all_artifacts", self.iam_client.get_custom_headers())
         if not result:
             return []
         return [Artifact.model_validate(item) for item in result]
@@ -67,11 +58,7 @@ class ArtifactClient:
         :return: The response object containing the created artifact details.
         :rtype: CreateArtifactResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.post("/create_artifact", custom_headers, request.model_dump())
+        result = self.client.post("/create_artifact", self.iam_client.get_custom_headers(), request.model_dump())
 
         return CreateArtifactResponse.model_validate(result)
 
@@ -84,11 +71,7 @@ class ArtifactClient:
         :return: The response object containing the created artifact details.
         :rtype: CreateArtifactFromTemplateResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.post("/create_artifact_from_template", custom_headers,
+        result = self.client.post("/create_artifact_from_template", self.iam_client.get_custom_headers(),
                                   {"artifact_template_id": artifact_template_id})
 
         return CreateArtifactFromTemplateResponse.model_validate(result)
@@ -102,11 +85,8 @@ class ArtifactClient:
         :return: The response object containing the rebuilt artifact details.
         :rtype: RebuildArtifactResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.post("/rebuild_artifact", custom_headers, {"artifact_id": artifact_id})
+        result = self.client.post("/rebuild_artifact", self.iam_client.get_custom_headers(),
+                                  {"artifact_id": artifact_id})
 
         return CreateArtifactResponse.model_validate(result)
 
@@ -119,11 +99,8 @@ class ArtifactClient:
         :return: The response object containing the deleted artifact details.
         :rtype: DeleteArtifactResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.delete("/delete_artifact", custom_headers, {"artifact_id": artifact_id})
+        result = self.client.delete("/delete_artifact", self.iam_client.get_custom_headers(),
+                                    {"artifact_id": artifact_id})
 
         return DeleteArtifactResponse.model_validate(result)
 
@@ -136,11 +113,7 @@ class ArtifactClient:
         :return: The response object containing the pre-signed URL and upload details.
         :rtype: GetBigFileUploadUrlResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.post("/get_bigfile_upload_url", custom_headers, request.model_dump())
+        result = self.client.post("/get_bigfile_upload_url", self.iam_client.get_custom_headers(), request.model_dump())
 
         return GetBigFileUploadUrlResponse.model_validate(result)
 
@@ -153,11 +126,7 @@ class ArtifactClient:
         :return: The response object containing the deletion status.
         :rtype: DeleteBigfileResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.delete("/delete_bigfile", custom_headers, request.dict())
+        result = self.client.delete("/delete_bigfile", self.iam_client.get_custom_headers(), request.dict())
 
         return DeleteBigfileResponse.model_validate(result)
 
@@ -169,9 +138,5 @@ class ArtifactClient:
         :return: A list of ArtifactTemplate objects.
         :rtype: List[ArtifactTemplate]
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.get("/get_artifact_templates", custom_headers)
+        result = self.client.get("/get_artifact_templates", self.iam_client.get_custom_headers())
         return GetArtifactTemplatesResponse.model_validate(result)

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud/_internal/_client/_http_client.py

@@ -2,7 +2,6 @@ import requests
 from .._exceptions import APIError
 from .._exceptions import UnauthorizedError
 from .._constants import *
-from .._config import *
 
 
 class HTTPClient:
@@ -62,6 +61,7 @@ class HTTPClient:
                 raise APIError(f"HTTP Request failed: {error_message}")
             # Raise for HTTP errors
             response.raise_for_status()
+            print(response.text)
 
         except requests.exceptions.RequestException as e:
             raise APIError(f"HTTP Request failed: {str(e)}")

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud/_internal/_client/_iam_client.py

@@ -3,7 +3,7 @@ import jwt
 from ._http_client import HTTPClient
 from .._config import IAM_SERVICE_BASE_URL
 from .._models import *
-from .._constants import CLIENT_ID_HEADER
+from .._constants import CLIENT_ID_HEADER, AUTHORIZATION_HEADER
 
 
 class IAMClient:
@@ -34,12 +34,35 @@ class IAMClient:
         custom_headers = {
             CLIENT_ID_HEADER: self._client_id
         }
-        req = LoginRequest(email=self._email, password=self._password)
-        result = self.client.post("/me/sessions", custom_headers, req.model_dump())
+        req = AuthTokenRequest(email=self._email, password=self._password)
+        auth_tokens_result = self.client.post("/me/auth-tokens", custom_headers, req.model_dump())
+        auth_tokens_resp = AuthTokenResponse.model_validate(auth_tokens_result)
 
-        resp = LoginResponse.model_validate(result)
-        self._access_token = resp.accessToken
-        self._refresh_token = resp.refreshToken
+        create_session_result = None
+        if auth_tokens_resp.is2FARequired:
+            max_attempts = 3
+            for attempt in range(max_attempts):
+                code = input(f"Attempt {attempt + 1}/{max_attempts}: Please enter the 2FA code: ")
+
+                create_session_req = CreateSessionRequest(
+                    type="native", authToken=auth_tokens_resp.authToken, otpCode=code
+                )
+                try:
+                    create_session_result = self.client.post(
+                        "/me/sessions", custom_headers, create_session_req.model_dump()
+                    )
+                    break
+                except Exception as e:
+                    print("Invalid 2FA code, please try again.")
+                    if attempt == max_attempts - 1:
+                        raise Exception("Failed to create session after 3 incorrect 2FA attempts.") from e
+        else:
+            create_session_req = CreateSessionRequest(type="native", authToken=auth_tokens_resp.authToken, otpCode=None)
+            create_session_result = self.client.post("/me/sessions", custom_headers, create_session_req.model_dump())
+
+        create_session_resp = CreateSessionResponse.model_validate(create_session_result)
+        self._access_token = create_session_resp.accessToken
+        self._refresh_token = create_session_resp.refreshToken
         self._user_id = self.parse_user_id()
 
     def refresh_token(self):
@@ -49,9 +72,9 @@ class IAMClient:
         custom_headers = {
             CLIENT_ID_HEADER: self._client_id
         }
-        result = self.client.patch("/me/sessions", custom_headers, {"refreshToken": self.refresh_token})
+        result = self.client.patch("/me/sessions", custom_headers, {"refreshToken": self._refresh_token})
 
-        resp = LoginResponse.model_validate(result)
+        resp = CreateSessionResponse.model_validate(result)
         self._access_token = resp.accessToken
         self._refresh_token = resp.refreshToken
 
@@ -90,3 +113,12 @@ class IAMClient:
         Gets the current client ID.
         """
         return self._client_id
+
+    def get_custom_headers(self) -> dict:
+        """
+        Gets the custom headers for the IAM client.
+        """
+        return {
+            AUTHORIZATION_HEADER: f'Bearer {self._access_token}',
+            CLIENT_ID_HEADER: self._client_id
+        }

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud/_internal/_client/_task_client.py

@@ -3,7 +3,6 @@ from ._decorator import handle_refresh_token
 from ._iam_client import IAMClient
 from .._config import TASK_SERVICE_BASE_URL
 from .._models import *
-from .._constants import ACCESS_TOKEN_HEADER, CLIENT_ID_HEADER
 
 
 class TaskClient:
@@ -30,11 +29,7 @@ class TaskClient:
         :return: An instance of Task containing the details of the retrieved task.
         :rtype: Task
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.get("/get_task", custom_headers, {"task_id": task_id})
+        result = self.client.get("/get_task", self.iam_client.get_custom_headers(), {"task_id": task_id})
 
         return Task.model_validate(result)
 
@@ -46,11 +41,7 @@ class TaskClient:
         :return: An instance of GetAllTasksResponse containing the retrieved tasks.
         :rtype: GetAllTasksResponse
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.get("/get_tasks", custom_headers)
+        result = self.client.get("/get_tasks", self.iam_client.get_custom_headers())
         if not result:
             return GetAllTasksResponse(tasks=[])
 
@@ -63,12 +54,7 @@ class TaskClient:
 
         :param task: The Task object containing the details of the task to be created.
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-
-        result = self.client.post("/create_task", custom_headers, task.model_dump())
+        result = self.client.post("/create_task", self.iam_client.get_custom_headers(), task.model_dump())
 
         return CreateTaskResponse.model_validate(result)
 
@@ -79,11 +65,7 @@ class TaskClient:
 
         :param task: The Task object containing the updated task details.
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        self.client.put("/update_schedule", custom_headers, task.model_dump())
+        self.client.put("/update_schedule", self.iam_client.get_custom_headers(), task.model_dump())
 
     @handle_refresh_token
     def start_task(self, task_id: str):
@@ -92,11 +74,7 @@ class TaskClient:
 
         :param task_id: The ID of the task to be started.
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        self.client.post("/start_task", custom_headers, {"task_id": task_id})
+        self.client.post("/start_task", self.iam_client.get_custom_headers(), {"task_id": task_id})
 
     @handle_refresh_token
     def stop_task(self, task_id: str):
@@ -105,11 +83,7 @@ class TaskClient:
 
         :param task_id: The ID of the task to be stopped.
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        self.client.post("/stop_task", custom_headers, {"task_id": task_id})
+        self.client.post("/stop_task", self.iam_client.get_custom_headers(), {"task_id": task_id})
 
     @handle_refresh_token
     def get_usage_data(self, start_timestamp: str, end_timestamp: str) -> GetUsageDataResponse:
@@ -119,11 +93,7 @@ class TaskClient:
         :param start_timestamp: The start timestamp of the usage data.
         :param end_timestamp: The end timestamp of the usage data.
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        result = self.client.get("/get_usage_data", custom_headers,
+        result = self.client.get("/get_usage_data", self.iam_client.get_custom_headers(),
                                  {"start_timestamp": start_timestamp, "end_timestamp": end_timestamp})
 
         return result
@@ -135,8 +105,4 @@ class TaskClient:
 
         :param task_id: The ID of the task to be archived.
         """
-        custom_headers = {
-            ACCESS_TOKEN_HEADER: self.iam_client.get_access_token(),
-            CLIENT_ID_HEADER: self.iam_client.get_client_id()
-        }
-        self.client.post("/archive_task", custom_headers, {"task_id": task_id})
+        self.client.post("/archive_task", self.iam_client.get_custom_headers(), {"task_id": task_id})

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud/_internal/_constants.py

@@ -1,4 +1,5 @@
 ACCEPT_HEADER = 'Accept'
+AUTHORIZATION_HEADER = 'Authorization'
 CONTENT_TYPE_HEADER = 'Content-Type'
 CLIENT_ID_HEADER = 'CE-ClientId'
 ACCESS_TOKEN_HEADER = 'CE-AccessToken'

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud/_internal/_models.py

@@ -307,10 +307,35 @@ class CreateTaskResponse(BaseModel):
     upload_link: str  # URL to upload the task data.
 
 
-class LoginResponse(BaseModel):
+class AuthTokenRequest(BaseModel):
+    """
+    Request object for user login.
+    """
+    email: str  # User email.
+    password: str  # User password.
+
+
+class AuthTokenResponse(BaseModel):
     """
     Response object for user login.
     """
+    authToken: str  # Access token for the user session.
+    is2FARequired: bool  # Indicates if 2FA is required for the user.
+
+
+class CreateSessionRequest(BaseModel):
+    """
+    Request object for creating a user session.
+    """
+    type: str  # Type of the session (e.g., native).
+    authToken: str  # Access token for the user session.
+    otpCode: Optional[str]  # 2FA code for the user session.
+
+
+class CreateSessionResponse(BaseModel):
+    """
+    Response object for creating a user session.
+    """
     accessToken: str  # Access token for the user session.
     refreshToken: str  # Refresh token for the user session.
 

{gmicloud-0.1.2 → gmicloud-0.1.4}/gmicloud.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: gmicloud
-Version: 0.1.2
+Version: 0.1.4
 Summary: GMI Cloud Python SDK
 Author-email: GMI <gmi@gmitec.net>
 License: MIT
@@ -10,10 +10,13 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.6
 Description-Content-Type: text/markdown
 
-# GMICloud SDK
+# GMICloud SDK (Beta)
 
 ## Overview
 
+Before you start: Our service and GPU resource is currenly invite-only so please contact our team (
+getstarted@gmicloud.ai) to get invited if you don't have one yet.
+
 The GMI Inference Engine SDK provides a Python interface for deploying and managing machine learning models in
 production environments. It allows users to create model artifacts, schedule tasks for serving models, and call
 inference APIs easily.
@@ -62,7 +65,18 @@ client = Client(client_id="<YOUR_CLIENT_ID>", email="<YOUR_EMAIL>", password="<Y
 
 ## Quick Start
 
-### 1. Create a Task from an Artifact Template
+### 1. How to run the code in the example folder
+```bash
+cd path/to/gmicloud-sdk
+# Create a virtual environment
+python -m venv venv
+source venv/bin/activate
+
+pip install -r requirements.txt
+python -m examples.<example_name>
+```
+
+### 2. Create a Task from an Artifact Template
 
 This is the simplest example to deploy an existing artifact template:
 
@@ -91,24 +105,30 @@ response = call_chat_completion(client, task.task_id)
 print(response)
 ```
 
-### 2. Step-by-Step Example: Create Artifact, Task, and Query the Endpoint
+### 3. Step-by-Step Example: Create Artifact, Task, and Query the Endpoint
 
 #### (a) Create an Artifact from a Template
 
 First, you’ll retrieve all templates and create an artifact based on the desired template (e.g., "Llama3.1 8B"):
 
 ```python
-def create_artifact_from_template(client):
+from gmicloud import *
+
+def create_artifact_from_template(client: Client) -> str:
     artifact_manager = client.artifact_manager
 
-    # List all available templates
+    # Get all artifact templates
     templates = artifact_manager.get_artifact_templates()
     for template in templates:
         if template.artifact_template_id == "qwen_2.5_14b_instruct_template_001":
-            return artifact_manager.create_artifact_from_template(
-                artifact_template_id=template.artifact_template_id
+            # Create an artifact from a template
+            artifact_id = artifact_manager.create_artifact_from_template(
+                artifact_template_id=template.artifact_template_id,
             )
-    return None
+
+            return artifact_id
+
+    return ""
 ```
 
 #### (b) Create a Task from the Artifact
@@ -116,43 +136,55 @@ def create_artifact_from_template(client):
 Wait until the artifact becomes "ready" and then deploy it using task scheduling:
 
 ```python
-def create_task_and_start(client, artifact_id):
-    artifact_manager = client.artifact_manager
+from gmicloud import *
+import time
+from datetime import datetime
 
-    # Wait until the artifact is ready
+def create_task_and_start(client: Client, artifact_id: str) -> str:
+    artifact_manager = client.artifact_manager
+    # Wait for the artifact to be ready
     while True:
-        artifact = artifact_manager.get_artifact(artifact_id)
-        if artifact.build_status == "SUCCESS":
-            break
-        print("Waiting for artifact to be ready...")
+        try:
+            artifact = artifact_manager.get_artifact(artifact_id)
+            print(f"Artifact status: {artifact.build_status}")
+            # Wait until the artifact is ready
+            if artifact.build_status == BuildStatus.SUCCESS:
+                break
+        except Exception as e:
+            raise e
+        # Wait for 2 seconds
         time.sleep(2)
-
-    # Configure and start the task
-    task_manager = client.task_manager
-    task = task_manager.create_task(Task(
-        config=TaskConfig(
-            ray_task_config=RayTaskConfig(
-                ray_version="2.40.0-py310-gpu",
-                file_path="serve",
-                artifact_id=artifact_id,
-                deployment_name="app",
-                replica_resource=ReplicaResource(
-                    cpu=10,
-                    ram_gb=100,
-                    gpu=1,
+    try:
+        task_manager = client.task_manager
+        # Create a task
+        task = task_manager.create_task(Task(
+            config=TaskConfig(
+                ray_task_config=RayTaskConfig(
+                    ray_version="2.40.0-py310-gpu",
+                    file_path="serve",
+                    artifact_id=artifact_id,
+                    deployment_name="app",
+                    replica_resource=ReplicaResource(
+                        cpu=10,
+                        ram_gb=100,
+                        gpu=1,
+                    ),
+                ),
+                task_scheduling=TaskScheduling(
+                    scheduling_oneoff=OneOffScheduling(
+                        trigger_timestamp=int(datetime.now().timestamp()) + 10,
+                        min_replicas=1,
+                        max_replicas=10,
+                    )
                 ),
             ),
-            task_scheduling=TaskScheduling(
-                scheduling_oneoff=OneOffScheduling(
-                    trigger_timestamp=int(datetime.now().timestamp()) + 10,
-                    min_replicas=1,
-                    max_replicas=10,
-                )
-            ),
-        ),
-    ))
+        ))
+
+        # Start the task
+        task_manager.start_task(task.task_id)
+    except Exception as e:
+        raise e
 
-    task_manager.start_task(task.task_id)
     return task.task_id
 ```
 
@@ -161,14 +193,20 @@ def create_task_and_start(client, artifact_id):
 Once the task is running, use the endpoint for inference:
 
 ```python
+from gmicloud import *
 from examples.completion import call_chat_completion
 
-client = Client()
-artifact_id = create_artifact_from_template(client)
-task_id = create_task_and_start(client, artifact_id)
+# Initialize the Client
+cli = Client()
 
-response = call_chat_completion(client, task_id)
-print(response)
+# Create an artifact from a template
+artifact_id = create_artifact_from_template(cli)
+
+# Create a task and start it
+task_id = create_task_and_start(cli, artifact_id)
+
+# Call chat completion
+print(call_chat_completion(cli, task_id))
 ```
 
 ## API Reference

{gmicloud-0.1.2 → gmicloud-0.1.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "gmicloud"
-version = "0.1.2"
+version = "0.1.4"
 description = "GMI Cloud Python SDK"
 authors = [{ name = "GMI", email = "gmi@gmitec.net" }]
 license = { text = "MIT" }