Introduction

Idea

In this project, I plan to Implement pre-sorting method in z-order pattern or Hilbert order pattern to improve the performance of GiST index building period.

Project proposal

My proposal for GSoC 2021 can be found at ​https://docs.google.com/document/d/1_mY_F2hPDk3vmXH5PPp2z9BuQWt-ZMORk6KxtdVQ3HY/edit?usp=sharing.

Link to Github repository: ​https://github.com/HanwGeek/postgis

Timeline

17th May - 7th June

Community bonding period:

• I have Introduced myself over the channel and shared my proposal over the mailing list for suggestions.
• Communicating with mentors and learned about community, working, etc. It is a great experience talking with experts in the domain.
• Forked the repository of PostGIS ​https://github.com/HanwGeek/postgis
• Updated wiki User page and added my personal information ​https://wiki.osgeo.org/wiki/User:HanwGeek
• Currently understanding the codebase.

Coding Week 1 (7th June - 13th June)

Coding Phase :

• Create sort support function bindings for geometry data type
• Implement a abbrev_convert with the current Hilbert hash function
• Test random point data set with sort support function
• Communicate with mentors and the Postgres/PostGIS community for more information and help

Plans for next week:

• Push current version of codes to my branch
• Prepare more complex test data

Coding Week 2 (14th June - 20th June)

Coding Phase :

• Fix the signature bug of sort support function
• Test random data of 10,000 point with pre-sort support function
• Communicate with mentors and the Postgres/PostGIS community for more information and help

Plans for next week:

• Ask for the suggestion about the Hilbert function signature from PostGIS community
• Refactor the code structure
• Create the signature of z-order hash function

Coding Week 3 (21th June - 27th June)

Coding Phase :

• Refactor the code structure, moving hash function to inline module
• Create a pull request and receive suggestions from community and mentors
• Finish morton hash function

Plans for next week:

• Create larger random data or use real-world data for testing
• Evaluate the stability and efficiency of hash functions
• Searching for a more efficient hash function

Coding Week 4 (28th June - 4th July)

Coding Phase :

• Finish hash function
• Create FlameGraph for cpu time analysis
• Prepare io access test

Plans for next week:

• Check the performance of hash function in detail
• Prepare for evaluation of performance and logic

Coding Week 5 (5th July - 11th July)

Coding Phase :

• Check the original paper of GiST in details
• Check the implement of GiST in Postgresql
• Prepare for evaluation of IO access

Plans for next week:

• Do the IO(buffer) access test
• Prepare for GSoC evaluation
• Confirm the proper hash function

Coding Week 6 (12th July - 18th July)

Coding Phase :

• Prepare for GSoC evaluation
• Test the IO access with EXPLAIN

Plans for next week:

• Confirm the proper hash function
• Do more research on hash function
• Go deep into GiST in postgres

Coding Week 7 (19th July - 25th July)

Coding Phase :

• Check the implementation of GiST in Postgres
• Test the Buffer hit and Execution Time of index query performance with random data

Plans for next week:

• Do more detailed tests to figure out the reason of performance test results
• Modify the hash functions

Coding Week 8 (26th July - 1st August)

Coding Phase :

• Complete the tests of Buffer hit and Execution Time of index query performance with random data
• Write a [document](​https://docs.google.com/document/d/1m4oxBAsKCyjAnYmkCmQ0X_ltiid5tliFwF3rtdzlKsc/edit?usp=sharing) with the test result

Plans for next week:

• Connect with mentors and community for the decision of pre-sort methods
• Check the page status of different pre-sort methods.

Coding Week 9 (2th August - 8st August)

Coding Phase :

• Apply pageinspect to debug the pre-sorting methods
• Do more traversal tests
• Propose a issue about gist_page_items function

Plans for next week:

• Fix the bugs in the pages of pre-sorting methods
• Do more detailed tests
• Organize existing code and test cases

Plans for next week:

• Connect with mentors and community for the decision of pre-sort methods
• Check the page status of different pre-sort methods.

Student's Biography

My name is Han WANG. I am a first year graduate student majoring in GIS at Peking University, and will get my Master's degree in 2023. And this is my github(​https://github.com/HanwGeek) and my linkedin(​https://www.linkedin.com/in/hanwgeek/). I am interested in all cool things. And it is very exciting to join the open source community! My research interest includes massive spatial temporal data management and analysis. Currently, I am working on a machine learning project based on big trajectory data, which is stored in PostgreSQL database and managed by PostGIS.

Mentors

*Regina Obe ​https://wiki.osgeo.org/wiki/User:Robe *Giuseppe Broccolo

[[Category: Google Summer of Code]]