huajunge/k-sim

Outline

Appendix of TraSS

​ Other Measures

​ Enlarged Element

​ Instantiation

Experiments

​ Setup

​ Experiments

TraSS

We have implemented the proposed framework in HBase, a popular key-value data store. We have published the framework, experimental code and the results in https://github.com/huajunge/k-sim.git. Details of DTW, Hausdorff and Range query，pleas refer to other_measures.pdf

Experiments

1. Setup

• (1) SupplementaryMaterial gives the resource of TraSS and sample data for testing TraSS;
• (2) Putting the trass.jar to the lib path of HBase.
• (3) building this project to get disks.jar (k-sim-traj-1.0-SNAPSHOT-jar-with-dependencies.jar), or using disks.jar in the SupplementaryMaterial.

2. Storing and Querying

• Storing

  spark-submit --class com.just.ksim.experiments.storing.StoringTDriveToHBase --master yarn --deploy-mode cluster --num-executors 30 --executor-memory 8G --executor-cores 2 trass.jar  ./tdrive tdrive_table hdfs:///index_time 8 1 1 16
  

  where, ./tdrive is data path; tdrive_table is the table for storing data; hdfs:///index_time is the path to record the indexing time; 8 is the shards; 1 1 is the data size; 16 is the maximum resolution of XZ* index.

• Threshold Similarity Search

  spark-submit --class com.just.ksim.experiments.query.SimilarityQuery trass.jar  ./tdrive_q tdrive_table 0.015 hdfs:///sim_015 8 16
  

  where, ./tdrive_q is the path query trajectories; tdrive_table is the table stored with tdrive data; 0.015 is the threshold; hdfs:///sim_015 is the path of query time; 8 is the shards; 16 is the maximum resolution of XZ* index.

• Top-k Similarity Search

  spark-submit --class com.just.ksim.experiments.query.KNNQuery trass.jar  ./tdrive_q tdrive_table 150 hdfs:///knn_150 8 0.002 16
  

  where, ./tdrive_q is the path query trajectories; tdrive_table is the table stored with tdrive data; 150 is the k; hdfs:///knn_015 is the path of query time; 8 is the shards; 16 is the maximum resolution of XZ* index.

• Range query

  spark-submit --class com.just.ksim.experiments.query.RangeQuery trass.jar '116.20094,39.97886,116.21094,39.98886;116.20094,39.97886,116.21094,39.98886' tdrive_table hdfs:///range_query_results 8 16
  

k-sim's Issues

代码

请问这是论文TraSS: Efficient Trajectory Similarity Search Based on Key-Value Data Stores的全部代码吗，感觉有很多，如果说只是测试一下论文中的功能是不是其中一部分代码

缺少包

com.just.ksim.index