ODC SDK, Python version

Software Development Kit (SDK) to work with ODC platform. https://services.scicrunch.io/odc/docs

For optimal support, I recommend using latest Python versions (3.13.x or 3.14.x at the time of writing). Prior Python versions should work but not guaranteed. See requirements.txt file for all required dependencies.

Note: package name is odc_sdk

Repo overview

• odc_sdk: This is source code for odc_sdk python package you will be installing.
  • configs: Setup .env file to store ODC API token. Functions will read from .env file when necessary.
  • models: Schema used when marshal response json easier to work with data format.
  • client.py: Functions user will interact with the most
  • endpoints.py: Helper functions to create full endpoint
  • request_handler.py: Http requests and middleware
• dist: Store latest pre-built wheel and tar.gz files. pip install git+ command looks into this folder first on install. Skips build step if these files are present.
• docs: Documentation, examples, decisions made during design process.
• tests: Use pytest to run.

Install

Production

python3 -m pip install git+https://github.com/ucsf-ferguson-lab/odc_python_sdk

Note: Unix systems might require --break-system-packages flag

python3 -m pip install git+https://github.com/ucsf-ferguson-lab/odc_python_sdk --break-system-packages

Dev

python3 -m pip install git+https://github.com/ucsf-ferguson-lab/odc_python_sdk.git@dev

Dependencies (not required)

If you need to install the required dependencies. Run

python3 -m pip install -r requirements.txt

Virtual environment (not required)

See https://docs.python.org/3/library/venv.html to setup virtualenv. Follow above steps to install package inside virtualenv.

Uninstall

python3 -m pip uninstall odc_sdk

Quick start

See code examples in /docs. Code below copied from /docs/stats/example.py.

from odc_sdk.endpoints import StatEndpoints
from odc_sdk.client import StatsClient
from odc_sdk.models import stats_responses as stats_models

if __name__ == "__main__":
    stat_endpoints = StatEndpoints.from_base_url()
    stat_client = StatsClient(stat_endpoints)

    # available methods: user_stats, download_stats, lab_stats, dataset_stats
    user_stats = stat_client.user_stats(stats_models)

    print(user_stats)

Supported routes

See https://services.scicrunch.io/odc/docs for full swagger documentation

public stats endpoints (no auth required)

• GET /stats/datasets
• GET /stats/users
• GET /stats/labs
• GET /stats/downloads

Not working

info endpoints (auth with API token required)

• GET /dataset/{dataset_id}/info

• GET /dataset/{dataset_id}/data-dictionary

• GET /user/info

• GET /dataset/{dataset_id}

• POST /dataset/create

• POST /dataset/upload

• POST /dataset/delete

• POST /dataset/{dataset_id}/data-dictionary/upload