Sample Notebook for new CI template code

This is a sample notebook to test new CI template using message exchange.

Description

This is a new template code for language-agnostic worker samples.

Sample 0: Boilarplate code for calling CI services

Find your server IP address first!



In [1]:

    
IP = '192.168.99.100'
BASE = 'http://' + IP + '/v1/'  # You need to change this to your service server



In [2]:

    
import requests
import json

def jprint(data):
    print(json.dumps(data, indent=4))

# Change this to  your Docker container's IP
HEADERS = {'Content-Type': 'application/json'}

# Check API Status
res = requests.get(BASE + 'services')
print("Here is the list of services supported by this submit agent:")
jprint(res.json())

# List of sample file locations
small_network = 'https://s3-us-west-2.amazonaws.com/ci-service-data/small.sif'
med_network = 'https://s3-us-west-2.amazonaws.com/ci-service-data/yeastHighQuality.sif'

# Pre-computed kernel and network (HumanNet)
human_net = 'http://chianti.ucsd.edu/~kono/ci/data/HumanNet.v1.symbol.sif'
human_net_kernel = 'http://chianti.ucsd.edu/~kono/ci/data/human_net.kernel.symbol.txt'









    



Here is the list of services supported by this submit agent:
[
    {
        "portNumber": "5556",
        "description": "Kernel generator service.",
        "serviceName": "kernel",
        "parameters": {
            "network_url": {
                "description": "URL of the source SIF network file.",
                "type": "string",
                "required": true
            }
        }
    },
    {
        "portNumber": "5548",
        "description": "Shell script worker example",
        "serviceName": "shell",
        "parameters": [
            {
                "description": "arg",
                "name": "arg",
                "type": "string",
                "required": true
            }
        ]
    },
    {
        "portNumber": "5557",
        "description": "Heat diffusion sub network finder service.",
        "serviceName": "subnet",
        "parameters": {
            "network_url": {
                "description": "URL of the SIF network file for the kernel.",
                "type": "string",
                "required": true
            },
            "query": {
                "description": "Array of gene names to be used for sub network search.",
                "type": "array",
                "required": true
            },
            "kernel_url": {
                "description": "URL of the pre-computed kernel file.",
                "type": "string",
                "required": true
            }
        }
    },
    {
        "portNumber": "5561",
        "description": "NBS worker sample",
        "serviceName": "nbs",
        "parameters": [
            {
                "description": "Input file location",
                "name": "inputFile",
                "type": "string",
                "required": true
            }
        ]
    },
    {
        "portNumber": "5559",
        "description": "Sample Python worker to display greeting message",
        "serviceName": "hello-python",
        "parameters": [
            {
                "description": "Any string message to be returned.",
                "name": "message",
                "type": "string",
                "required": true
            }
        ]
    },
    {
        "portNumber": "5549",
        "description": "Sample Python worker to process network from NDEx",
        "serviceName": "ndex",
        "parameters": [
            {
                "description": "Unique network ID in NDEx",
                "name": "network_id",
                "type": "string",
                "required": true
            }
        ]
    },
    {
        "portNumber": "5560",
        "description": "Sample Java worker to display greeting message",
        "serviceName": "hello-java",
        "parameters": [
            {
                "description": "Greeting message",
                "name": "message",
                "type": "string",
                "required": true
            }
        ]
    }
]

Sample1: Find Subnetwork using pre-computed kernel file

if you already have pre-computed kernel file, you can use it to find sub networks using:

URL of Kernel file
URL of SIF (network) file

f (kernel, network, query) = subnetwork



In [3]:

    
# Pre-computed sample kernel and network (HumanNet) stored in UCSD web server (about 0.5 GB)
human_net = 'http://chianti.ucsd.edu/~kono/ci/data/HumanNet.v1.symbol.sif'
human_net_kernel = 'http://chianti.ucsd.edu/~kono/ci/data/human_net.kernel.symbol.txt'

# Specify kernel and SIF file locations
base_query = {
    'kernel_url': human_net_kernel,
    'network_url': human_net,
}

gene_list = ['BRCA1', 'MAPK1']
base_query['query'] = gene_list

res = requests.post(BASE + 'services/kernel', data=json.dumps(base_query), headers=HEADERS)
jprint(res.json())
job_id1 = res.json()['job_id']









    



{
    "status": "queued",
    "job_id": "acbbbe04-d4e3-43ba-9f9f-06c7a4aef4d2"
}



In [7]:

    
# Check the status of the job
res = requests.get(BASE + 'queue')
jprint(res.json())









    



[
    {
        "status": "queued",
        "job_id": "acbbbe04-d4e3-43ba-9f9f-06c7a4aef4d2"
    }
]



In [8]:

    
# Get the result (Subnetwork in CX format)
result_url = BASE + 'queue/' + job_id1 + '/result'
print(result_url)
res = requests.get(result_url)
#jprint(res.json())









    



http://192.168.99.100/v1/queue/acbbbe04-d4e3-43ba-9f9f-06c7a4aef4d2/result

Sample 2: Create kernel from SIF file

f (network) = kernel



In [ ]:

    
# Specify locations of the SIF files
sif_file_locations = [small_network, med_network]

kernel_generation_tasks = {}

for sif_location in sif_file_locations:
    body = {
        'network_url': sif_location
    }
    res = requests.post(BASE + 'services/kernel', json=body, headers=HEADERS)
    kernel_generation_tasks[sif_location] = res.json()['job_id']
    jprint(res.json())



In [ ]:

    
# Check the status of the job
res = requests.get(BASE + 'queue')
jprint(res.json())



In [ ]:

    
# Get result
job_id_for_small_network = kernel_generation_tasks[small_network]

kernel_url = BASE + 'queue/' + job_id_for_small_network + '/result'
print(kernel_url)
res = requests.get(kernel_url)
kernel_info = res.json()
jprint(kernel_info)



In [ ]:

    
# Use the result to find subnetwork
base_query = {
    'kernel_url': kernel_info['kernel_file'],
    'network_url': kernel_info['network']
}

gene_list = ["NRAS", "KRAS", "MAPK1"]
base_query['query'] = gene_list

res = requests.post(BASE + 'services/subnet', data=json.dumps(base_query), headers=HEADERS)
find_subnet_job_id = res.json()['job_id']
jprint(res.json())



In [ ]:

    
# Check the result
result_url = BASE + 'queue/' + find_subnet_job_id + '/result'
print(result_url)
res = requests.get(result_url)
#jprint(res.json())

Sample 3: Get actual Kernel file

For current configuration, kernels will be saved in a server in a restricted zone (= docker container separated from internet zone). To access this, you need to open port.

Of course, this may be changed. It depends on deployment target.



In [ ]:

    
import pandas as pd

temp = kernel_info['kernel_file']
parts = temp.split('/')
parts[2] = IP + ':3333' # This is the default opened port...
kernel_url = '/'.join(parts)
print(kernel_url)

df = pd.read_csv(kernel_url, sep="\t")
df.head()