superb-ai-curate

Overview

superb-ai-curate is the Python client for interacting with Superb Curate.

Installation

You don't need this source code unless you want to modify the package. If you just want to use the package, just run:

$ pip install --upgrade superb-ai-curate

Requirements

Python 3.7+

Documentation

You can also find the documentation for superb-ai-curate on the website.

• Introduction
• Installation
• Tutorial

Usage

An Access Key is required to use the python client. This can be generated from the Settings > Access menu on the Superb AI Curate website. For more details on access key issuance and management, you can check the Access Key Management documentation. The Team Name refers to the organization name that your personal account belongs to.

import spb_curate
from spb_curate import curate

spb_curate.access_key = "..."
spb_curate.team_name = "..."

dataset = curate.fetch_dataset(id="...")

images = [
    curate.Image(
        key="<unique image key>",
        source=curate.ImageSourceLocal(asset="/path/to/image"),
        metadata={"weather": "clear", "timeofday": "daytime"},
    ),
    curate.Image(
        key="<unique image key>",
        source=curate.ImageSourceLocal(asset="/path/to/image"),
        metadata={"weather": "clear", "timeofday": "daytime"},
    ),
]

job: curate.Job = dataset.add_images(images=images)
job.wait_until_complete()

Configuring per-request

For use with multiple credentials, the requests can be configured at the function level.

from spb_curate import curate

dataset = curate.fetch_dataset(access_key="...", team_name="...", id="...")

Logging

If required, the client can be configured to produce basic logging output. There are two levels that are logged to, INFO and DEBUG. For production use, INFO is the recommended logging level, however DEBUG can be used for more verbosity.

There are several methods for setting the log level.

1. Environment variable

$ export SPB_LOG_LEVEL = "INFO"

2. Superb AI Curate Python client global setting

import spb_curate

spb_curate.log_level = "INFO"

3. Python logging library

import logging

logging.basicConfig()
logging.getLogger("superb-ai").setLevel(logging.INFO)

Development

The development environment relies on Poetry for package management, testing and building.

$ poetry install -E dev
$ poetry run pytest --cov=spb_curate tests