Evaluation - plotting_utils

[JanBenisek]

• Change names and labels. Based on the input, the function should know when use "AUC" or "RMSE".

  cobra/cobra/evaluation/plotting_utils.py

  Line 9 in 4da81c6

  def plot_univariate_predictor_quality(df_auc: pd.DataFrame,

• Change names and labels here. Based on the input, the function should know when use "AUC" or "RMSE".

  cobra/cobra/evaluation/plotting_utils.py

  Line 77 in 4da81c6

  def plot_performance_curves(model_performance: pd.DataFrame,

• Document and write/modify unit tests

[hendrikdewinter8]

When testing the performance curves plot, for the tutorial of linear regression, I had following bug:

[sborms]

Hi @hendrikdewinter8, I tried to set a reasonable automatic range for the y-axis, but seems I was a bit too quick. Goal is to have a bit of extra graph space on top. Maybe np.arange(0, max_metric+1) is more than sufficient, even for very large values (but might be too big for small values...). Feel free to see what works and push a branch with a fix to the repo.

[hendrikdewinter8]

@sborms , I used your suggestion and now it works fine! I'll commit it

[sandervh14]

Reopening shortly to upload a fix for the y axis in the model performance plot for forward feature selection: