Quick Tutorial

This quick tutorial has two parts:

Example is a complete fragment of code that will generate a hierarchy from an input network.
The steps section below provide more information on how to use this library.

Example

Open a terminal and run the block of code below either in a Python terminal or as a Python script (requires Python 3.4+ with cdapsutil and ndex2 packages installed)

import json
import cdapsutil
import ndex2


# Create CommunityDetection object
cd = cdapsutil.CommunityDetection(runner=cdapsutil.ServiceRunner())

# Create NDEx2 python client
client = ndex2.client.Ndex2()

# Download BioGRID: Protein-Protein Interactions (SARS-CoV) from NDEx
# http://ndexbio.org/viewer/networks/669f30a3-cee6-11ea-aaef-0ac135e8bacf
client_resp = client.get_network_as_cx_stream('669f30a3-cee6-11ea-aaef-0ac135e8bacf')

# Convert downloaded network to NiceCXNetwork object
net_cx = ndex2.create_nice_cx_from_raw_cx(json.loads(client_resp.content))

# Run HiDeF on CDAPS REST service
hier_net = cd.run_community_detection(net_cx, algorithm='hidef')

# Print information about hierarchy
print('Hierarchy name: ' + str(hier_net.get_name()))
print('# nodes: ' + str(len(hier_net.get_nodes())))
print('# edges: ' + str(len(hier_net.get_edges())))

# Display 1st 500 characters of hierarchy network CX
print(json.dumps(hier_net.to_cx())[0:500])

The code blocks above use the NDEx2 Python client to download BioGRID: Protein-Protein Interactions (SARS-CoV) network from NDEx as a NiceCXNetwork. The Community Detection algorithm HiDeF is then run on the network using the CDAPS REST Service. The result is a NiceCXNetwork stored in hier_net object.

Step 1 - Choose what to run

import cdapsutil

sr = cdapsutil.ServiceRunner()
algos = sr.get_algorithms()['algorithms']
for key in algos.keys():
    if 'EDGELIST' not in algos[key]['inputDataFormat']:
        continue
    if 'COMMUNITYDETECT' not in algos[key]['outputDataFormat']:
        continue
    print('Algorithm name: ' + str(key))
    print('\tDocker image: ' + str(algos[key]['dockerImage']))

Step 2 - Choose where to run

Alternate Runner objects can be set in the CommunityDetection constructor object to denote where to run the Community Detection Algorithm.

Remotely Via CDAPS REST Service

If runner is omitted this is the default behavior. When used the Community Detection Algorithm is run remotely on CDAPS REST Service

cd = cdapsutil.CommunityDetection(runner=cdapsutil.ServiceRunner())

Locally Via Docker

The DockerRunner runs the Community Detection Algorithm using a locally installed Docker.

cd = cdapsutil.CommunityDetection(runner=cdapsutil.DockerRunner())

Use already generated output

The ExternalResultsRunner assumes the algorithm passed into run_community_detection is a path to a file containing the results from an externally run Community Detection Algorithm. This is useful for the case where the algorithm was run on a cluster via Singularity.

cd = cdapsutil.CommunityDetection(runner=cdapsutil.ExternalResultsRunner())