oslo.vision • training notebooks • sdk • blog

OsloVision Python Client

This Python package provides a client for interacting with the Oslo API, a platform for creating and managing datasets for machine learning projects. The client allows you to easily upload images, create annotations, and download dataset exports.

Features

• Simple interface for interacting with the Oslo API
• Support for adding images to projects
• Creation of annotations for images
• Downloading of dataset exports
• Automatic handling of authentication

Installation

You can install the Oslo API Python client using pip:

pip install oslovision

Usage

Here's a quick example of how to use the Oslo API client:

from oslovision import OsloVision

# Initialize the client
api = OsloVision("your_api_token_here")

# Test the API connection
print(api.test_api())

# Add an image to a project
with open("image.jpg", "rb") as img_file:
    image_data = api.add_image("your_project_identifier", img_file)
print(f"Added image: {image_data['id']}")

# Create an annotation
annotation = api.create_annotation(
    "your_project_identifier",
    image_data['id'],
    "cat",
    x0=10,
    y0=20,
    width_px=100,
    height_px=150
)
print(f"Created annotation: {annotation['id']}")

# Download an export
download_path = api.download_export("your_project_identifier", 1)
print(f"Export downloaded to: {download_path}")

API Reference

OsloVision(base_url: str, token: str)

Initialize the Oslo API client.

• base_url: The base URL of the Oslo API (e.g., "https://app.oslo.vision/api/v1")
• token: Your Oslo API authentication token

Methods

test_api() -> Dict[str, str]

Test if the API is up and running and the token is valid.

add_image(project_identifier: str, image: Union[str, IOBase], split: str = "train", status: str = "pending") -> Dict

Add an image to a project.

• project_identifier: The ID of the project to add the image to
• image: Either a file object or a URL string of the image
• split: The dataset split for the image (default: "train")
• status: The status of the image (default: "pending")

Returns a dictionary with the added image's data.

create_annotation(project_identifier: str, image_identifier: str, label: str, x0: float, y0: float, width_px: float, height_px: float) -> Dict

Create a new annotation for an image.

• project_identifier: The ID of the project
• image_identifier: The ID of the image to annotate
• label: The label for the annotation
• x0, y0: The top-left coordinates of the bounding box
• width_px, height_px: The width and height of the bounding box in pixels

Returns a dictionary with the created annotation's data.

download_export(project_identifier: str, version: int, output_dir: str) -> str

Download a dataset export and extract from the zip file.

• project_identifier: The ID of the project
• version: The version number of the export
• output_dir: The directory to save the downloaded files (default: current directory)

Returns the path to the downloaded export.

Contributing

Contributions to the Oslo API Python client are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.