"FAST: Frequency-Aware Indexing for Spatio-Textual Data Streams" Extensions

---

Usage

FASTkNN index can but utilized by the edu.purdue.cs.fast.FAST class. Following is a simple example of FASTkNN usage. Refer to experiments package for more detailed examples.

import edu.purdue.cs.fast.config.CleanMethod;
import edu.purdue.cs.fast.models.Rectangle;

Config config = new Config();
config.KNN_OMEGA =0.2;

Rectangle spatialBounds = new Rectangle(new Point(0.0, 0.0), new Point(512.0, 512.0));

FAST index = new FAST(
        config,
        spatialBounds,
        512,
        9
);
index.setExternalObjectIndex(512, 9);

index.insertQuery(...);
index.insertObject(...);

Directory Overview

• src/main/java/<package> contains the extension of FAST to support KNN queries.
• src/main/java/<package>/baselines contains the following baselines.
  • ckqst - Reproduction of the paper "Continuous k Nearest Neighbor Queries over Large-Scale Spatial–Textual Data Streams"
  • fast - Original FAST index for continuous MBR queries
  • quadtree - Implementation of a base quadtree
• src/test/java/<package> contains the tests for FAST extension and baselines.
• analysis contains notebooks with different analysis done.
• data contains samples of the datasets.

Running experiments

cd <project_root>

# Build the project
mvn package -Dmaven.test.skip

# Run a experiment (update the required fields within `run_fast.sh`)
./run_fast.sh