libica

libica

Python SDK to programmatically call Illumina Connected Analytics (ICA) Genomic (multi-omics) data platform and BioInformatics web services. This SDK supports both ICA v1 and v2 APIs:

• ICAv1 API: https://illumina.gitbook.io/ica-v1/reference/r-api
• ICAv2 API: https://help.ica.illumina.com/reference/r-api
• Install through pip like so:

pip install libica

Overview

• Tested for Python 3.8, 3.9, 3.10, 3.11, 3.12
• See ChangeLog and Milestones
• Test Coverage
• Wiki
• SDK Guide: PyDoc
• User Guide: https://umccr-illumina.github.io/libica/

Getting started with SDK for ICA v2

See pilot.py

# List all data in a project
import os
from contextlib import closing

from libica.openapi.v2 import Configuration, ApiClient, ApiException
from libica.openapi.v2.api.project_data_api import ProjectDataApi
from libica.openapi.v2.model.project_data import ProjectData
from libica.openapi.v2.model.project_data_paged_list import ProjectDataPagedList

if __name__ == '__main__':

    project_id = os.environ['ICAV2_PROJECT_ID']
    file_path = [""]  # empty will list everything under project

    icav2_access_token = os.environ['ICAV2_ACCESS_TOKEN']
    ica_url = "https://ica.illumina.com/ica/rest"

    configuration = Configuration(
        host=ica_url,
        access_token=icav2_access_token,
    )
    # configuration.debug = True  # uncomment to debug API call logging

    api_client = ApiClient(
        configuration=configuration,
        header_name="Content-Type",
        header_value="application/vnd.illumina.v3+json",
    )

    with closing(api_client) as ctx_api_client:
        project_data_api: ProjectDataApi = ProjectDataApi(api_client=ctx_api_client)

        try:
            page_token = ""
            while True:
                project_data_paged_list: ProjectDataPagedList = project_data_api.get_project_data_list(
                    project_id=project_id,
                    file_path=file_path,
                    page_size=str(1000),
                    page_token=page_token,
                )

                for item in project_data_paged_list.items:
                    project_data: ProjectData = item
                    print((
                        project_data.data.id,  # fil.<ID> (or) fol.<ID>
                        project_data.data.details.path,
                        project_data.data.details.data_type,
                    ))

                page_token = project_data_paged_list.next_page_token
                if not project_data_paged_list.next_page_token:
                    break

        except ApiException as e:
            print(e)

You may consider Using App Package (see below).

Getting started with SDK for ICA v1

• Export ICA base URL and JWT Bearer token:

export ICA_BASE_URL=<baseUrl>
export ICA_ACCESS_TOKEN=<tok>

• Generate Bearer JWT token using ICA CLI like so:

ica tokens create --help

• Somewhere in your Python code:

import os
from libica.openapi import libwes
from libica.openapi.libwes import WorkflowList, WorkflowCompact

ica_base_url = os.getenv("ICA_BASE_URL")
ica_access_token = os.getenv("ICA_ACCESS_TOKEN")

configuration = libwes.Configuration(
    host=ica_base_url,
    api_key={
        'Authorization': ica_access_token
    },
    api_key_prefix={
        'Authorization': "Bearer"
    },
)

with libwes.ApiClient(configuration) as api_client:

    wfl_api: libwes.WorkflowsApi = libwes.WorkflowsApi(api_client)
    try:
        page_token = None
        while True:
            wfl_list: WorkflowList = wfl_api.list_workflows(page_size=100, page_token=page_token)
            # print(wfl_list)
            for item in wfl_list.items:
                wfl: WorkflowCompact = item
                print(wfl.id)
                print(wfl.name)
            page_token = wfl_list.next_page_token
            if not wfl_list.next_page_token:
                break
    except libwes.ApiException as e:
        print(e)

Using App Package

NOTE: libica.app package contains reusable modules that are based on use cases around UMCCR Data Portal backend, Workflows automation and orchestration implementations. Hence, it may be a specific domain logic implementation. However, it may still be reusable for your use cases. Starter examples are as follows.

App for ICA v2

See pilotapp.py

Example: ProjectDataOps app to list project files, download a file, etc...

from contextlib import closing

from libica.app import AppHelper
from libica.app.dataops import ProjectDataOps
from libica.openapi.v2 import ApiClient
from libica.openapi.v2.model.project_data import ProjectData

# --- Use AppHelper to build SDK API client

app_helper = AppHelper(debug=False)
project_id = app_helper.get_icav2_cli_session_project_id()

cli_session_api_client: ApiClient = app_helper \
    .build_icav2_configuration_with_cli_session() \
    .get_icav2_api_client()

# --- Construct ProjectDataOps from dataops module

project_dataops = ProjectDataOps(project_id=project_id, api_client=cli_session_api_client)

# --- List all files under given project's folder

# If you do not cd into the folder, it will list all files under the project
project_dataops.cd("/test_folder/")

for item in project_dataops.list_files():
    project_data: ProjectData = item
    print((
        project_data.data.id,  # fil.<ID> (or) fol.<ID>
        project_data.data.details.path,
        project_data.data.details.data_type,
        project_data.data.details.name,
        project_data.data.details.status,
        project_data.data.details.file_size_in_bytes,
        project_data.data.details.time_created,
    ))

# --- Download csv file from given project and file path

file_path = "/test_folder/SampleSheet.csv"
print(f"Downloading {file_path} from project_id {project_id}")
project_dataops.cd(file_path=file_path)
ntf = project_dataops.download_file()
with closing(ntf) as cf:
    with open(cf.name, 'r') as f:
        for line in f.readlines():
            print(line)

For more, see PyDoc:

• https://umccr-illumina.github.io/libica/libica/app/dataops.html

App for ICA v1

Example: Configuration Object Builder

from libica.app import configuration
from libica.openapi import libgds

gds_config = configuration(
  lib=libgds,  # pass in library of interest e.g. libwes, libtes, etc 
  secret_name=["FROM_AWS_SECRET_MANAGER_THAT_STORE_ICA_ACCESS_TOKEN"],
  base_url="https://use1.platform.illumina.com",  # overwrite if not https://aps2.platform.illumina.com
  debug=False,  # True if you like to debug API calls, False by default anyway, just for demo
)

with libgds.ApiClient(gds_config) as api_client:
    ...

Example: Listing Files from GDS

from typing import List

from libica.app import gds
from libica.openapi import libgds

vol, path = gds.parse_path("gds://development/some/folder/path/")
files: List[libgds.FileResponse] = gds.get_file_list(volume_name=vol, path=path)

for file in files:
  print(f"{file.name}, {file.volume_name}, {file.path}, {file.presigned_url}")

Development

• Setup virtual environment and activate it

• Install dev dependencies

make install

• To bring up mock API μ-services

make up

• To run tests suites

make unit
make autounit

• To run full suite, smoke test

make test

AutoGen Workflow

• SDK is autogenerated from OpenAPI (Swagger) definitions
• There are few tools required for this autogen workflow to work.
  1. openapi-generator-cli -- used to generate code and doc
  2. redocly/cli -- validate definitions
• These tools are Node.js app, hence, required build tools node, npm, npx and yarn as follows.

node -v
v20.15.1

npm i -g yarn
yarn -v
4.3.1

yarn install

npx openapi-generator-cli help
npx redocly lint --help

• API mock server prism is set up through docker compose as follows.

docker --version
Docker version 27.0.3, build 7d4bcd8

make up
make ps
make test_ica_mock
make test_icav2_mock
docker compose logs

• All the autogen config and arrangement refer to syncapi.sh and syncapi2.sh which is called through Makefile targets.

Generator Version

• See generator version compatibility
• Select generator version as follows:

npx openapi-generator-cli version-manager list

Notice

• MIT License and DISCLAIMER
• The Advanced Genomics Collaboration (TAGC)