gen3 2.2.1

The Gen3 SDK makes it easy to utilize functionality in Gen3 data commons.

Gen3 SDK for Python

The Gen3 SDK for Python provides classes for handling the authentication flow using a refresh token and getting an access token from the commons. The access token is then refreshed as necessary while the refresh token remains valid. The submission client contains various functions for submitting, exporting, and deleting data from a Gen3 data commons.

Docs for this SDK are available at http://gen3sdk-python.rtfd.io/

Auth

This contains an auth wrapper for supporting JWT based authentication with requests. The access token is generated from the refresh token and is regenerated on expiration.

IndexClient

This is the client for interacting with the Indexd service for GUID brokering and resolution.

SubmissionClient

This is the client for interacting with the Gen3 submission service including GraphQL queries.

Indexing Tools

Download Manifest

How to download a manifest object-manifest.csv of all file objects in indexd for a given commons:

import sys
import logging
import asyncio

from gen3.index import Gen3Index
from gen3.tools import indexing
from gen3.tools.indexing.verify_manifest import manifest_row_parsers

logging.basicConfig(filename="output.log", level=logging.DEBUG)
logging.getLogger().addHandler(logging.StreamHandler(sys.stdout))

COMMONS = "https://{{insert-commons-here}}/"

def main():
    loop = asyncio.new_event_loop()
    asyncio.set_event_loop(loop)

    loop.run_until_complete(
        indexing.async_download_object_manifest(
            COMMONS,
            output_filename="object-manifest.csv",
            num_processes=8,
            max_concurrent_requests=24,
        )
    )


if __name__ == "__main__":
    main()

The output file will contain columns guid, urls, authz, acl, md5, file_size with info populated from indexd.

Verify Manifest

How to verify the file objects in indexd against a "source of truth" manifest.

Bonus: How to override default parsing of manifest to match a different structure.

In the example below we assume a manifest named alternate-manifest.csv already exists with info of what's expected in indexd. The headers in the alternate-manifest.csv are guid, urls, authz, acl, md5, size.

NOTE: The alternate manifest headers differ rfom the default headers described above (file_size doesn't exist and should be taken from size)

import sys
import logging

from gen3.index import Gen3Index
from gen3.tools import indexing
from gen3.tools.indexing.verify_manifest import manifest_row_parsers

logging.basicConfig(filename="output.log", level=logging.INFO)
logging.getLogger().addHandler(logging.StreamHandler(sys.stdout))

COMMONS = "https://{{insert-commons-here}}/"


def main():
    def _get_file_size(row):
        try:
            return int(row.get("size"))
        except Exception:
            logging.warning(f"could not convert this to an int: {row.get('size')}")
            return row.get("size")

    # override default parsers
    manifest_row_parsers["file_size"] = _get_file_size

    indexing.verify_object_manifest(
        COMMONS, manifest_file="alternate-manifest.csv", num_processes=20
    )


if __name__ == "__main__":
    main()

A more complex example is below. In this example:

• The input file is a tab-separated value file (instead of default CSV)
  • Note the manifest_file_delimiter argument
• The arrays in the file are represented with Python-like list syntax
  • ex: ['DEV', 'test'] for the acl column
• We are using more Python processes (20) to speed up the verify process
  • NOTE: You need to be careful about this, as indexd itself needs to support scaling to this number of concurrent requests coming in

import sys
import logging

from gen3.index import Gen3Index
from gen3.tools import indexing
from gen3.tools.indexing.verify_manifest import manifest_row_parsers

logging.basicConfig(filename="output.log", level=logging.INFO)
logging.getLogger().addHandler(logging.StreamHandler(sys.stdout))

COMMONS = "https://{{insert-commons-here}}/"


def main():
    def _get_file_size(row):
        try:
            return int(row.get("size"))
        except Exception:
            logging.warning(f"could not convert this to an int: {row.get('size')}")
            return row.get("size")

    def _get_acl_from_row(row):
        return [row.get("acl").strip().strip("[").strip("]").strip("'")]

    def _get_authz_from_row(row):
        return [row.get("authz").strip().strip("[").strip("]").strip("'")]

    def _get_urls_from_row(row):
        return [row.get("url").strip()]

    # override default parsers
    manifest_row_parsers["file_size"] = _get_file_size
    manifest_row_parsers["acl"] = _get_acl_from_row
    manifest_row_parsers["authz"] = _get_authz_from_row
    manifest_row_parsers["urls"] = _get_urls_from_row

    indexing.verify_object_manifest(
        COMMONS,
        manifest_file="output-manifest.csv",
        manifest_file_delimiter="\t",
        num_processes=20,
    )


if __name__ == "__main__":
    main()

Changelog

0.1.0

Initial release Functionality for IndexClient, and Submission client