antigranular 0.2.0

Antigranular is a community-driven, open-source platform that merges confidential computing and differential privacy. This creates a secure environment for handling and unlocking the full potential of unseen data..

Unlock privacy: Getting along with Antigranular

Antigranular is a community-driven, open-source platform that merges confidential computing and differential privacy. This creates a secure environment for handling and unlocking the full potential of unseen data.

Connect to Antigranular

Antigranular works with just 4 characters %%ag , like magic! Any code written after the magic cell %%ag is run in our remote server which is a restricted environment allowing methods which guarantees differential privacy.

Install the Antigranular package using pip:

!pip install antigranular

Import the Antigranular library:

import antigranular as ag

Use your client credentials and dataset or competition ID to connect to the AG Enclave Server:

ag.login("client id": "<client_secret_id>": competition="<competition_id>")

A succesful login will register the cell magic %%ag.

Loading Private Datasets

Private dataset objects can be loaded in the form of PrivateDataFrames and PrivateSeries using the ag_utils library. ag_utils is a package locally intalled in the remote server. This eliminated the hassle of having to install anything other than antigranular package.

You can learn more about PrivateSeries and PrivateDataFrames on our quick on Private Pandas.

We use load_dataset() method to obtain a collection of private objects in the form of a dictionary. The structure of the response dictionary, dataset path and private object names will be mentioned during the competition.

%%ag
from op_pandas import PrivateDataFrame , PrivateSeries
from ag_utils import load_dataset 
"""
Sample response structure
{
    train_x : priv_train_x,
    train_y : priv_train_y,
    test_x : priv_test_x
}
"""
# Obtaining the dictionary containing private objects
response = load_dataset("<path_to_dataset>")

# Unpacking the response dictionary
train_x = response["train_x"]
train_y = response["train_y"]
test_x = response["test_x"]

Exporting Objects

Since %%ag runs code in a very restricted environment, you need to export the differentially private objects to the local environment in order to do further analysis. The data objects can be exported using the export method in ag_utils.

API info: export(obj, variable_name:str)

The remote object gets exported to the local environment and gets assigned to the stated variable name. It is important to note thatPrivateSeries and PrivateDataFrame objects cannot be exported and will raise an error if tried to be exported in any manner.

%%ag
from ag_utils import export
train_info = train_x.describe(eps=1)
export(train_info , 'variable_name')

Once exported, you can apply any form of data anlysis on the differentially private object.

# Local code block
print(variable_name)
--------------------------------------
                    Age         Salary
    count  99987.000000   99987.000000
    mean      38.435953  120009.334336
    std       12.167379   46255.486093
    min       18.257448   40048.259037
    25%       27.185189   80057.639960
    50%       38.210860  120380.291216
    75%       49.147724  159835.637091
    max       59.282932  199920.664706

Libraries Supported

• pandas: A versatile data manipulation library that offers efficient data structures and tools for data analysis and manipulation.

• op_pandas: A wrapped library specifically designed for differentially private data manipulation within the Pandas framework. It enhances privacy-preserving techniques and enables privacy-aware data processing.

• op_diffprivlib: A differentially private library that provides various privacy-preserving algorithms and mechanisms for machine learning and data analysis tasks.

• op_smartnoise: A library focused on privacy-preserving analysis using the SmartNoise framework. It provides tools for differential privacy and secure computation.

• op_opendp: A library that offers differentially private data analysis and algorithms based on the OpenDP project. It provides privacy-preserving methods and tools for statistical analysis.