Automatically generates "usage" charts, from a collection of "results" screenshots from Pokemon Masters EX. A time-saver for a handful of people in Pokemon Masters EX communities.
Currently, it works well with some limitations. The image recognition accuracy is 100% on identical Sync Pairs, so no false positives. But, realistically, it needs to match "mechanically identical Sync Pairs" (e.g. EX-upgraded vs not, costume/pose variations). Brainstorming ways to avoid manually tagging groups.
Progress:
- Extract trainer faces
- Group identical Sync Pairs
- Generate charts
- Group variations of the same Sync Pairs (E.g. pose variations, EX vs non-EX)
- Install python 1.12
- Open the command line and run this to install
git clone [repo url]
cd pomas_chart
pip install -r requirements
- Add images to the
infolder - Go to the
pomas_chartfolder and run this commandpython main.py - The chart should be in
out\chart.png
out has intermediate outputs. zip them up (along with the in folder) and send them along if you have a bug to report.
If you want to tweak the chart, you can also modify out\groups; each group should contain the collected screenshots of Sync Pair. You can regenerate the chart after by running python generate_chart.py.