srvanrell / ranking-table-tennis Goto Github PK

View Code? Open in Web Editor NEW

0.0 3.0 1.0 1.65 MB

Ranking system based on rating points for table tennis players.

Home Page: http://srvanrell.github.io/ranking-table-tennis/

License: GNU General Public License v2.0

Python 99.79% Shell 0.13% Makefile 0.08%

rating-points ranking table-tennis ranking-system

ranking-table-tennis's Introduction

ranking-table-tennis

Ranking system for table tennis players.

The system keeps a record of championship points and rating points.

Championship points aim at tracking the tournament achievements of players. Their computation is based on the best rounds reached by players in each tournament.

Rating points aim at tracking the relative skill level of players. Their computation is based on the outcomes of one-versus-one matches.

Installation

This system has been developed for a Linux environment.

Poetry installation (recommended)

poetry install

System-wide installation

sudo pip install ranking-table-tennis

Single-user installation

pip install --user ranking-table-tennis

Uninstallation

[sudo] pip uninstall ranking-table-tennis

Update

System-wide update

sudo pip install -U ranking-table-tennis

Single-user update (recommended)

pip install --user -U ranking-table-tennis

Usage

The commands must be run in a bash terminal.

Fill a sheet with the tournament matches. It must be saved in the Tournaments spreadsheet (xlsx).

Players and Initial Ranking sheets must be in the same spreadsheet (it is used as a database).
Run rtt preprocess.

The scripts will read the Tournament spreadsheet and will ask for missing information of new players (city, affiliation, initial rating points, and category). This information will be saved in the Players and Initial Ranking sheets.
Run rtt compute.

It will ask for the tournament that you want to process. 0 will compute all from the beggining. The outcome will be saved in the Ranking spreadsheet.
Run rtt publish.

It will ask for the index of the tournament that you want to publish. The outcome will be saved in a new spreadsheet.

Development

Install locally from source (source directory will immediately affect the installed package without needing to re-install):

pip install --user --editable .

sudo apt install graphviz (still required?)

Update version at pyproject.toml and then create a source distribution

poetry build

Upload to PyPI

poetry publish [--dry-run] [--build]

Upload to TestPyPI

poetry publish -r testpypi [--dry-run] [--build]

ranking-table-tennis's People

Contributors

Watchers

Forkers

steap

ranking-table-tennis's Issues

Initial ranking should log when the player was added

Addition date or TID should be included into initial ranking

Improve config system

Configs are simple yaml files. That could be replaced for a more elegant strategy like OmegaConf.

Publish should order by rating and pid to be consistent

Multiple run of publish commands should give the same results.

Raw rankings have incorrect metadata

Raw rankings have initial tid metadata. Revise.

Consider using logging instead of prints

Single command rtt

Create a single command line rtt that call the other scripts as:

rtt preprocess
rtt compute-ranking
rtt publish

Change publish defaults to offline

Due that the new site has been published, uploading data to spreadsheets is no longer required (except for tournament sheets)

Preprocess should update players and initial ranking only if required

It should check for differences before attempting to update

Package upload instructions and Readme are not up to date

Revise Readme and package build and upload instructions.

Consider moving database to BigQuery

Currently, database is created locally every time scripts are executed. They could be stored as BigQuery tables, which would enable creation of published sheets much easier and less manual.

Add a blank line between categories

As a visual aid, a blank line can be added in published tables where applies the category limit

Add a rating helper on unknowns players

Show results of unknown player with known players (and their ratings) so it is easier to define the initial rating points

Historic per player should publish data up to given tid

Evolution plots

A few ideas to be implemented in the future:

Number of participants vs tournaments (overall and per category)
Number of active/inactive players vs tournaments (overall and per category)

PEP8 is not encouraged nor checked

Consider adding tests and hooks to improve PEP8 compliance.

Statistics should publish data up to given tournament

Nowadays is publishing stats of future tournaments. For example, T03 is showing results for >= T04.

Config cannot change during a championship

Currently configs are considered fixed during the championship and when it was required, hacks have been added to make changes during a championship. Evaluate an alternative approach to consider these kind of changes, which would allow to reuse configs from one year to the other.

add unidecode to requirements

Add it to setup.py for the next deployment

Move config to yaml or similar

move to a config file

spanish phrases
hard written text
headers