gmicloud 0.1.0__tar.gz

gmicloud-0.1.0/PKG-INFO

@@ -0,0 +1,208 @@
+Metadata-Version: 2.2
+Name: gmicloud
+Version: 0.1.0
+Summary: GMI Cloud Python SDK
+Home-page: https://github.com/GMISWE/python-sdk
+Author: GMI
+Author-email: GMI <gmi@gmitec.net>
+License: MIT
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.6
+Description-Content-Type: text/markdown
+Dynamic: author
+Dynamic: home-page
+Dynamic: requires-python
+
+# GMICloud
+
+## Overview
+
+This project is an open-source SDK for interacting with the GMICLOUD platform. It provides functionalities to manage
+artifacts, tasks, and usage data efficiently.
+
+## Features
+
+- Create and manage artifacts
+- Create and manage tasks
+- Fetch usage data
+- Integration with OpenAI for chat completions
+
+## Installation
+
+To install the SDK, use pip:
+
+```bash
+pip install gmicloud-sdk
+```
+
+## Usage
+
+### Initialize the Client
+
+```python
+from gmicloud import Client
+
+client = Client(username="your_username", password="your_password")
+```
+
+### Create an Artifact with a File
+
+```python
+artifact_id = client.artifact_manager.create_artifact_with_file(
+    artifact_name="Llama3.1 8B",
+    artifact_file_path="./files/Llama-3.1-8B-Instruct.zip",
+    description="This is a test artifact",
+    tags=['example', 'test']
+)
+```
+
+### Create an Artifact from a Template
+
+```python
+artifact_id = client.artifact_manager.create_artifact_from_template(
+    artifact_template_id="template_id"
+)
+```
+
+### Create and Start a Task
+
+```python
+task_id = client.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=2,
+                ram_gb=128,
+                gpu=2,
+            ),
+        ),
+        task_scheduling=TaskScheduling(
+            scheduling_oneoff=OneOffScheduling(
+                trigger_timestamp=int(datetime.now().timestamp()) + 60,
+                min_replicas=1,
+                max_replicas=10,
+            )
+        ),
+    ),
+))
+
+client.task_manager.start_task(task_id)
+```
+
+### Call Chat Completion
+
+```python
+from openai import OpenAI
+
+task = client.task_manager.get_task(task_id)
+open_ai = OpenAI(
+    api_key="YOUR_API_KEY",
+    base_url=task.info.endpoint
+)
+
+completion = open_ai.chat.completions.create(
+    model="meta-llama/Llama-3.1-8B-Instruct",
+    messages=[
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user", "content": "Translate the sentences to Chinese"},
+    ],
+    max_tokens=200,
+    temperature=0.7
+)
+
+print(completion.choices[0].message.content)
+```
+
+## Configuration
+
+### One-off Task Configuration
+
+`examples/config/one-off_task.json`:
+
+```json
+{
+  "config": {
+    "ray_task_config": {
+      "ray_version": "2.40.0",
+      "file_path": "serve",
+      "deployment_name": "string",
+      "replica_resource": {
+        "cpu": 2,
+        "ram_gb": 12,
+        "gpu": 1
+      }
+    },
+    "task_scheduling": {
+      "scheduling_oneoff": {
+        "min_replicas": 1,
+        "max_replicas": 1
+      }
+    }
+  }
+}
+```
+
+### Daily Task Configuration
+
+`examples/config/daily_task.json`:
+
+```json
+{
+  "config": {
+    "ray_task_config": {
+      "ray_version": "2.40.0-py310-gpu",
+      "file_path": "serve",
+      "deployment_name": "string",
+      "replica_resource": {
+        "cpu": 6,
+        "ram_gb": 64,
+        "gpu": 2
+      }
+    },
+    "task_scheduling": {
+      "scheduling_daily": {
+        "triggers": [
+          {
+            "timezone": "UTC",
+            "Hour": 0,
+            "minute": 10,
+            "second": 0,
+            "min_replicas": 1,
+            "max_replicas": 2
+          },
+          {
+            "timezone": "UTC",
+            "Hour": 0,
+            "minute": 10,
+            "second": 30,
+            "min_replicas": 1,
+            "max_replicas": 4
+          }
+        ]
+      }
+    }
+  }
+}
+```
+
+## Running Tests
+
+To run the unit tests, use the following command:
+
+```bash
+pytest
+```
+
+## Contributing
+
+Contributions are welcome! Please open an issue or submit a pull request.
+
+## License
+
+This project is licensed under the MIT License. See the `LICENSE` file for details.

gmicloud-0.1.0/README.md

@@ -0,0 +1,191 @@
+# GMICloud
+
+## Overview
+
+This project is an open-source SDK for interacting with the GMICLOUD platform. It provides functionalities to manage
+artifacts, tasks, and usage data efficiently.
+
+## Features
+
+- Create and manage artifacts
+- Create and manage tasks
+- Fetch usage data
+- Integration with OpenAI for chat completions
+
+## Installation
+
+To install the SDK, use pip:
+
+```bash
+pip install gmicloud-sdk
+```
+
+## Usage
+
+### Initialize the Client
+
+```python
+from gmicloud import Client
+
+client = Client(username="your_username", password="your_password")
+```
+
+### Create an Artifact with a File
+
+```python
+artifact_id = client.artifact_manager.create_artifact_with_file(
+    artifact_name="Llama3.1 8B",
+    artifact_file_path="./files/Llama-3.1-8B-Instruct.zip",
+    description="This is a test artifact",
+    tags=['example', 'test']
+)
+```
+
+### Create an Artifact from a Template
+
+```python
+artifact_id = client.artifact_manager.create_artifact_from_template(
+    artifact_template_id="template_id"
+)
+```
+
+### Create and Start a Task
+
+```python
+task_id = client.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=2,
+                ram_gb=128,
+                gpu=2,
+            ),
+        ),
+        task_scheduling=TaskScheduling(
+            scheduling_oneoff=OneOffScheduling(
+                trigger_timestamp=int(datetime.now().timestamp()) + 60,
+                min_replicas=1,
+                max_replicas=10,
+            )
+        ),
+    ),
+))
+
+client.task_manager.start_task(task_id)
+```
+
+### Call Chat Completion
+
+```python
+from openai import OpenAI
+
+task = client.task_manager.get_task(task_id)
+open_ai = OpenAI(
+    api_key="YOUR_API_KEY",
+    base_url=task.info.endpoint
+)
+
+completion = open_ai.chat.completions.create(
+    model="meta-llama/Llama-3.1-8B-Instruct",
+    messages=[
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user", "content": "Translate the sentences to Chinese"},
+    ],
+    max_tokens=200,
+    temperature=0.7
+)
+
+print(completion.choices[0].message.content)
+```
+
+## Configuration
+
+### One-off Task Configuration
+
+`examples/config/one-off_task.json`:
+
+```json
+{
+  "config": {
+    "ray_task_config": {
+      "ray_version": "2.40.0",
+      "file_path": "serve",
+      "deployment_name": "string",
+      "replica_resource": {
+        "cpu": 2,
+        "ram_gb": 12,
+        "gpu": 1
+      }
+    },
+    "task_scheduling": {
+      "scheduling_oneoff": {
+        "min_replicas": 1,
+        "max_replicas": 1
+      }
+    }
+  }
+}
+```
+
+### Daily Task Configuration
+
+`examples/config/daily_task.json`:
+
+```json
+{
+  "config": {
+    "ray_task_config": {
+      "ray_version": "2.40.0-py310-gpu",
+      "file_path": "serve",
+      "deployment_name": "string",
+      "replica_resource": {
+        "cpu": 6,
+        "ram_gb": 64,
+        "gpu": 2
+      }
+    },
+    "task_scheduling": {
+      "scheduling_daily": {
+        "triggers": [
+          {
+            "timezone": "UTC",
+            "Hour": 0,
+            "minute": 10,
+            "second": 0,
+            "min_replicas": 1,
+            "max_replicas": 2
+          },
+          {
+            "timezone": "UTC",
+            "Hour": 0,
+            "minute": 10,
+            "second": 30,
+            "min_replicas": 1,
+            "max_replicas": 4
+          }
+        ]
+      }
+    }
+  }
+}
+```
+
+## Running Tests
+
+To run the unit tests, use the following command:
+
+```bash
+pytest
+```
+
+## Contributing
+
+Contributions are welcome! Please open an issue or submit a pull request.
+
+## License
+
+This project is licensed under the MIT License. See the `LICENSE` file for details.

gmicloud-0.1.0/examples/example.py

@@ -0,0 +1,145 @@
+import os
+import time
+from datetime import datetime
+
+from openai import OpenAI
+
+from gmicloud import *
+
+
+def create_artifact_with_file(client: Client) -> str:
+    artifact_manager = client.artifact_manager
+
+    # Create an artifact with a file
+    artifact_id = artifact_manager.create_artifact_with_file(
+        artifact_name="Llama3.1 8B",
+        artifact_file_path="./files/Llama-3.1-8B-Instruct.zip",
+        description="This is a test artifact",
+        tags=['example', 'test']
+    )
+
+    return artifact_id
+
+
+def create_artifact_from_template(client: Client) -> str:
+    artifact_manager = client.artifact_manager
+
+    # Get all artifact templates
+    templates = artifact_manager.get_artifact_templates()
+    print(templates)
+    for template in templates:
+        if template.artifact_name == "Llama3.1 8B":
+            # Create an artifact from a template
+            artifact_id = artifact_manager.create_artifact_from_template(
+                artifact_template_id=template.artifact_template_id,
+            )
+
+            return artifact_id
+
+    return ""
+
+
+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:
+        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)
+    try:
+        task_manager = client.task_manager
+        # Create a task
+        task = task_manager.create_task(Task(
+            config=TaskConfig(
+                ray_task_config=RayTaskConfig(
+                    ray_version="latest-py311-gpu",
+                    file_path="serve",
+                    artifact_id=artifact_id,
+                    deployment_name="app",
+                    replica_resource=ReplicaResource(
+                        cpu=24,
+                        ram_gb=128,
+                        gpu=2,
+                    ),
+                ),
+                task_scheduling=TaskScheduling(
+                    scheduling_oneoff=OneOffScheduling(
+                        trigger_timestamp=int(datetime.now().timestamp()) + 60,
+                        min_replicas=1,
+                        max_replicas=10,
+                    )
+                ),
+            ),
+        ))
+
+        # Start the task
+        task_manager.start_task(task.task_id)
+    except Exception as e:
+        raise e
+
+    return task.task_id
+
+
+def call_chat_completion(client: Client, task_id: str):
+    task_manager = client.task_manager
+    # Wait for the task to be ready
+    while True:
+        try:
+            task = task_manager.get_task(task_id)
+            print(f"task status: {task.task_status}")
+            # Wait until the task is ready
+            if task.task_status == "ready":
+                break
+        except Exception as e:
+            print(e)
+            return
+        # Wait for 2 seconds
+        time.sleep(2)
+
+    if not task.info.endpoint or not task.info.endpoint.strip():
+        raise Exception("Task endpoint is not ready yet")
+
+    open_ai = OpenAI(
+        api_key=os.getenv("OPENAI_API_KEY", "YOUR_DEFAULT_API_KEY"),
+        base_url=os.getenv("OPENAI_API_BASE", task.info.endpoint)
+    )
+    # Make a chat completion request using the new OpenAI client.
+    completion = open_ai.chat.completions.create(
+        model="meta-llama/Llama-3.1-8B-Instruct",
+        messages=[
+            {"role": "system", "content": "You are a helpful assistant."},
+            {"role": "user",
+             "content": f"Translate the sentences to Chinese"},
+        ],
+        max_tokens=200,
+        temperature=0.7
+    )
+
+    print(completion.choices[0].message.content)
+
+
+if __name__ == '__main__':
+    # Initialize the Client
+    cli = Client()
+
+    # print(cli.artifact_manager.get_all_artifacts())
+
+    # Create an artifact with a file
+    # artifact_id = create_artifact_with_file(cli)
+
+    # Create an artifact from a template
+    #artifact_id = create_artifact_from_template(cli)
+    artifact_id = "cba6db2f-315a-4765-9e94-1e692f7fdb39"
+
+    # Create a task and start it
+    task_id = create_task_and_start(cli, artifact_id)
+
+    # Call chat completion
+    call_chat_completion(cli, task_id)

gmicloud-0.1.0/gmicloud/__init__.py

@@ -0,0 +1,41 @@
+from ._internal._models import (
+    Artifact,
+    ArtifactData,
+    ArtifactMetadata,
+    Task,
+    TaskOwner,
+    TaskConfig,
+    TaskInfo,
+    RayTaskConfig,
+    TaskScheduling,
+    ReplicaResource,
+    OneOffScheduling,
+    DailyScheduling,
+    DailyTrigger,
+    ArtifactTemplate
+)
+from ._internal._enums import (
+    BuildStatus,
+    TaskEndpointStatus
+)
+from .client import Client
+
+__all__ = [
+    "Client",
+    "Artifact",
+    "ArtifactData",
+    "ArtifactMetadata",
+    "Task",
+    "TaskOwner",
+    "TaskConfig",
+    "TaskInfo",
+    "RayTaskConfig",
+    "TaskScheduling",
+    "ReplicaResource",
+    "OneOffScheduling",
+    "DailyScheduling",
+    "DailyTrigger",
+    "ArtifactTemplate",
+    "BuildStatus",
+    "TaskEndpointStatus"
+]