Faasera Documentation

SDK Integration Guide (Java / PySpark)

This guide provides instructions to integrate the Faasera SDK within your data pipelines using either Java or PySpark environments.

---

Overview

The Faasera SDK enables you to embed core capabilities like profiling, masking, and validation directly into your ETL or ML workflows using Java or PySpark. This is ideal for high-throughput, low-latency scenarios where data privacy must be enforced in real-time.

---

Supported Environments

Environment	Integration Mode	SDK Format
Java	JAR / Maven	faasera-core.jar
PySpark	pip / .whl	faasera_sdk.whl
Databricks	Notebook / Wheel	%pip install

---

Java SDK Integration

Step 1: Add the SDK

If using Maven:

<dependency>
    <groupId>ai.faasera</groupId>
    <artifactId>faasera-core</artifactId>
    <version>1.0.0</version>
</dependency>

Or download the JAR:

wget https://repo.faasera.ai/releases/faasera-core-1.0.0.jar

Step 2: Initialize SDK

FaaseraMaskingSDK sdk = new FaaseraMaskingSDK();
sdk.initialize("path/to/policy.json");
Dataset<Row> maskedDF = sdk.mask(inputDF);

Step 3: Use in Spark Job

inputDF.createOrReplaceTempView("input_table");
spark.udf().register("mask_udf", sdk.getUDF());
spark.sql("SELECT mask_udf(column1) FROM input_table").show();

---

PySpark SDK Integration

Step 1: Install the Wheel

%pip install faasera_sdk-1.0.0-py3-none-any.whl

Step 2: Import and Initialize

from faasera_sdk import FaaseraProfiler

profiler = FaaseraProfiler(policy_path="policy.json")
profiled_df = profiler.profile(spark_df)

Step 3: Apply Masking

from faasera_sdk import FaaseraMasker

masker = FaaseraMasker(policy_path="policy.json")
masked_df = masker.mask(spark_df)

---

Testing Locally

You can run the SDK with sample CSVs or Parquet files:

java -jar faasera-core.jar --mode=mask --input=./data.csv --output=./masked.csv

Or use PySpark CLI:

pyspark --packages faasera_sdk

---

Tips & Best Practices

• Reuse policy files across stages (profile → mask → validate)
• Use broadcast joins to optimize masking at scale
• Disable hinting in air-gapped environments (set "enabled": false in policy)
• Use deterministic keys for consistent masking across sessions

---

License Activation

For SDK-based deployments, a license token (offline or cloud-linked) is required. Configure the license path as:

faasera.license.path=/etc/faasera/license.jwt

Or set via environment:

export FAASERA_LICENSE_PATH=/etc/faasera/license.jwt

---

Need help? Visit faasera.ai/docs or reach out to support@faasera.ai.