A repository for creating visualization of Moonboard data. There is also a page which can generate some selected figures, https://www.moonplt.com/ Note that this is in development and might be unavailable at times.
Install python dependencies via
python -m pip install matplotlib
python -m pip install numpy
python -m pip install flask
python -m pip install colorcet
python -m pip install requests
Download submodule, https://github.com/spookykat/MoonBoard via
git submodule update --init
Move into MoonBoard folder,
cd MoonBoard
For the moment, add pull request to enable fetching the logbook
git fetch origin pull/3/head:fetch_logbook
git checkout fetch_logbook
git rebase main
Edit secret.py with username and password
Fetch problems, this might take a minute or two.
python fetch_problems.py
Fetch logbook, this will probably not takes as long.
python fetch_logbook.py
The idea is having one main script for loading the data, and some modular system for different plots, statistics or other. The main script is 'run.py'. It loads data and pass this to some 'datahandlers'.
The following command will run two datahandlers. The first will produce a 'benchmark progress plot', showing the accumulated number of benchmarks as a function of time for some logbook, for all used holdsets and angles. The second will produce a scatter plot showing what week days and time benchmarks have been registered (sent?), for some logbook.
python run.py --datahandlers BenchmarkProgress Times
Another example would be
python run.py --datahandlers Benchmarks BenchmarkHoldFrequency
The first will produce a 'problem view' for all benchmark problems of the chosen dataset file (by default in the Output/Benchmarks folder), and the second will produce a visualization which holds are used the most for each grade among the benchmarks (by default in the Output/BenchmarkHoldFrequency folder).
run.py takes some arguments. --filename is the path to the dataset file, by default 'MoonBoard/problems MoonBoard 2016 .json'. --save-dir is the folder where to save data, by default 'Output'. --datahandlers take one or several datahandlers, as seen above. Using --settings one can provide arguments to the datahandlers. These can be general or specific. General settings are provided as e.g. --settings overwrite:True and will be given to all datahandlers. Specific settings are provided as e.g. --settings BenchmarkHoldFrequency:"dict(overwrite=True)" and will only apply to the specified datahandler.
Datahandlers can return data in the form of dicts, and which is then passed to following datahandlers. If someone e.g. developed a 'generate Moonboard beta' algorithm, this could be implemented as a datahandler, and then used by other datahandlers.
The web interface can be started by running
python run_flask.py
which will start a server listening on localhost:5000. As it will inform, this is a development server not to be used in deployment. One can anyway make it reachable from outside by instead running
flask --app run_flask.py run --host=0.0.0.0
This might be useful for trying a mobile interface e.g.
Contributions are appreciated! Examples could be
- ideas for plots
- implementation of plots
- web design (...)