Predicting the lanes for League of Legends Participants with a CNN

Setup

Download the Git repository and install the requirements

pip install -r requirements.txt

How to use

Training the CNN

I really recommend the use tensorflow-gpu for the training or it will take some time to train it To run the training open the train.py file in you favorite python editor. You may change the name:

from tools.train import CnnTrain
cnn = CnnTrain(data_name='data')
cnn.run('test')

You can also look up the Comment i tried to explain everything If you want to see the graphs from you model while its training you can run in the console. When you are in the League-of-Legends-Predict-Lane's folder (--logdir="Path to logs")

tensorboard --logdir=logs

Predicting Lanes

There is all ready a pre trained model in the models folder which you can use. To predict something you import the predict function from the predict.py file and run it with a Summoner Name. The Summoner have to be in a game to call the spectator api.

from tools.predict import Lane
lane = Lane("RGAPI-XXX-XXX", model_name='test4_conv1D_153_one_dense_153_1557335309_1557335316')
prediction = lane.predict('SaItySurprise')
prediction = lane.predict(4023710639)
print(prediction)
{
 'blue':
    {
    'Lux': {'Bottom': 92.0, 'Middle': 7.4, 'Top': 0.73, 'Jungle': 0.34}, 
    'Ashe': {'Bottom': 92.0, 'Middle': 7.2, 'Top': 0.71, 'Jungle': 0.34}, 
    'Zed': {'Middle': 90.0, 'Top': 7.0, 'Bottom': 2.7, 'Jungle': 0.72}, 
    'Urgot': {'Top': 83.0, 'Jungle': 8.6, 'Middle': 6.4, 'Bottom': 1.5}, 
    'Evelynn': {'Jungle': 99.0, 'Bottom': 0.29,'Top': 0.21, 'Middle': 0.031}
    }, 
 'red': 
    {
    'Illaoi': {'Top': 83.0, 'Jungle': 8.0, 'Middle': 7.4, 'Bottom': 1.8}, 
    'Ekko': {'Middle': 91.0, 'Top': 6.7, 'Bottom': 2.0, 'Jungle': 0.64}, 
    'Soraka': {'Bottom': 92.0, 'Middle': 7.2, 'Top': 0.71, 'Jungle': 0.34}, 
    'Kaisa': {'Bottom': 92.0, 'Middle': 7.2, 'Top': 0.71, 'Jungle': 0.34}, 
    'Trundle': {'Jungle': 99.0, 'Bottom': 0.29, 'Top': 0.21, 'Middle': 0.031}
    }
}

If you want to know more about the prediction feel free to read the Comments in the prediction.py file. I tried to explain everything. If there a Questions or any suggestions feel free to Label the issues and make a pull requests

The Flask Module

Install the requirements run the main.py script the the start and the first request need need some time but the next's request are much faster. To predict something send a get request to

localhost:5000/predict/NameOrMatchId

If you want to change the port just open the main.py file and chang there:

if __name__ == '__main__':
    app.run(port=YourPreferredPort)