zgiai 1.0.0

A lightweight utility library for modern Python development

ZGI SDK

A lightweight Python SDK with OpenAI-compatible API interface.

Installation

pip install zgiai

Quick Start

OpenAI-Compatible API Client

from zgi import ZGI

# Initialize client
client = ZGI(api_key="your-api-key")

# Or use environment variable
# export ZGI_API_KEY="your-api-key"
client = ZGI()

# Chat completions
response = client.chat.completions.create(
    model="gpt-4",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"}
    ]
)
print(response['choices'][0]['message']['content'])

# Upload a file
with open("training_data.jsonl", "rb") as f:
    file = client.files.create(
        file=f,
        purpose="fine-tune"
    )
print(f"Uploaded file: {file['id']}")

# Create fine-tuning job
job = client.fine_tuning.jobs.create(
    training_file=file['id'],
    model="gpt-3.5-turbo",
    hyperparameters={
        "n_epochs": 3
    }
)
print(f"Fine-tuning job created: {job['id']}")

# Check job status
job_status = client.fine_tuning.jobs.retrieve(job['id'])
print(f"Status: {job_status['status']}")

# List models
models = client.models.list()
for model in models['data']:
    print(model['id'])

Utility Functions

from zgi import is_empty, deep_clone, debounce, sleep, random_string, format_date

# Check if a value is empty
is_empty('')  # True
is_empty([])  # True
is_empty({})  # True

# Deep clone an object
original = {'a': 1, 'b': {'c': 2}}
cloned = deep_clone(original)

# Debounce a function
@debounce(wait=0.3)
def my_function():
    print('Debounced!')

# Sleep for a duration
sleep(1)  # Sleep for 1 second

# Generate a random string
random_str = random_string(10)  # Random 10-character string

# Format a date
from datetime import datetime
formatted = format_date(datetime.now())  # "2024-01-01"

API Reference

ZGI Client

ZGI(api_key, base_url, organization, timeout)

Initialize the ZGI client.

Parameters:

• api_key (str): Your ZGI API key (or set ZGI_API_KEY env var)
• base_url (str, optional): Base URL for API (default: https://api.zgi.ai/v1)
• organization (str, optional): Organization ID
• timeout (int, optional): Request timeout in seconds (default: 60)

Chat API

client.chat.completions.create(**kwargs)

Create a chat completion.

Parameters:

• model (str): Model ID (e.g., "gpt-4", "gpt-3.5-turbo")
• messages (list): List of message dicts with 'role' and 'content'
• temperature (float): Sampling temperature (0-2, default: 1.0)
• max_tokens (int, optional): Maximum tokens to generate
• stream (bool): Whether to stream responses (default: False)

Files API

client.files.create(file, purpose)

Upload a file.

Parameters:

• file (BinaryIO): File object to upload
• purpose (str): Purpose ("fine-tune", "assistants", etc.)

client.files.list(purpose=None)

List all files.

client.files.retrieve(file_id)

Get file information.

client.files.delete(file_id)

Delete a file.

Fine-tuning API

client.fine_tuning.jobs.create(training_file, model, **kwargs)

Create a fine-tuning job.

Parameters:

• training_file (str): File ID of training data
• model (str): Base model to fine-tune
• validation_file (str, optional): File ID of validation data
• hyperparameters (dict, optional): Training hyperparameters
• suffix (str, optional): Custom model name suffix

client.fine_tuning.jobs.list(limit=20, after=None)

List fine-tuning jobs.

client.fine_tuning.jobs.retrieve(job_id)

Get job details.

client.fine_tuning.jobs.cancel(job_id)

Cancel a job.

Models API

client.models.list()

List all available models.

client.models.retrieve(model_id)

Get model information.

Utility Functions

is_empty(value: Any) -> bool

Check if a value is empty (None, empty string, empty list, or empty dict).

deep_clone(obj: T) -> T

Create a deep copy of an object.

debounce(wait: float = 0.3)

Decorator to debounce a function with an optional wait time (default: 0.3 seconds).

sleep(seconds: float) -> None

Sleep for a specified duration in seconds.

random_string(length: int = 10) -> str

Generate a random alphanumeric string with specified length (default: 10).

format_date(date: datetime = None, fmt: str = "%Y-%m-%d") -> str

Format a date to a string (default: current date, format: YYYY-MM-DD).

Environment Variables

• ZGI_API_KEY: Your ZGI API key
• ZGI_BASE_URL: Custom API base URL (optional)
• ZGI_ORGANIZATION: Organization ID (optional)

Requirements

• Python >= 3.7
• requests >= 2.20.0

License

MIT