Skip to content

IKNL/vtg

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

92 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation


vantage6

A privacy preserving federated learning solution


vtg

R package that supports running algorithms on the Vantage6 federated infrastructure. Contains a Client and functions that facilitate running dockerized algorithms.

The user can also specify the data format for the input and output of an algorithm. The client and wrapper handle the (de)serialization for this. This feature makes it possible to run R-based algorithms from the Python client as well.

Note: If you want cross-language algorithms be sure that, while creating an algorithm, the type of your results can be converted to json.

Installation

Run the following in the R console to install the package and its dependencies:

devtools::install_github('IKNL/vtg')

Usage

# Function to create a client
setup.client <- function() {
    # Define parameters
    username <- "username@example.com"
    password <- "password"
    host <- "https://harukas.vantage6.ai:443"
    api_path <- ""
    
    # Create the client
    client <- vtg::Client$new(host, api_path)
    client$authenticate(username, password)

    return(client)
}

# Create a client
client <- setup.client()

# Get a list of available collaborations
print(client$getCollaborations())

# Could output something like this:
#   id     name
# 1  1 ZEPPELIN
# 2  2 PIPELINE

# Select a collaboration
client$setCollaborationId(1)

# Data format can be set to either 'json' or 'rds' (default)
client$data_format <- 'json'

# User can specify to use the master function instead of RPC calls
client$use.master.container <- TRUE

Then the client can be used to create new (sub-)tasks and retrieve the results.

Run an algorithm

For example if we want to run the master method (dcoxph) from the Cox Proportional Hazards (coxph) algorithm.

Run the following in the R console to install the package and its dependencies:

# This also installs the package vtg
devtools::install_github('mellesies/vtg.coxph', subdir="src")

Run the following for the coxph model:

# Define explanatory variables, time column and censor column
expl_vars <- c("Age","Race2","Race3","Mar2","Mar3","Mar4","Mar5","Mar9",
               "Hist8520","hist8522","hist8480","hist8501","hist8201",
               "hist8211","grade","ts","nne","npn","er2","er4")
time_col <- "Time"
censor_col <- "Censor"

# The master function of vtg.coxph is called 'dcoxph'
result <- vtg.coxph::dcoxph(client, expl_vars, time_col, censor_col)

print(result)

vantage6

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •  

Languages