📝 Update README

- 🐛 use larger image to test tutorial on colab

.github/workflows/test_pkg_on_colab.yaml

@@ -11,7 +11,7 @@ on:
 jobs:
     test-tutorial-on-colab:
         name: Test tutorial on latest colab image
-        runs-on: ubuntu-latest-2core # increase disk space  
+        runs-on: ubuntu-latest-4core # increase disk space  
         # https://console.cloud.google.com/artifacts/docker/colab-images/europe/public/runtime
         container:
             image: europe-docker.pkg.dev/colab-images/public/runtime:latest

project/README.md

@@ -204,3 +204,17 @@ jupytext 00_5_training_data_exploration.py --to ipynb -o - | papermill - runs/ex
 - finds files originationg from fractionation experiments
 - plot mask indicating presence/abscence of peptide measurement in an experiment
 - intensity log-transformation: 
+
+## Adhoc analysis
+
+Compare performance splitting samples into train, validation and test set.
+Use scikit-learn `KNN_IMPUTER` as it's easiest to tweak and understand.
+
+```bash
+# classic:
+jupytext --to ipynb -k - -o - 01_1_train_KNN.py | papermill - runs/rev3/01_1_train_KNN.ipynb
+# train only on samples without simulated missing values, add simulated missing values to test and validation samples
+jupytext --to ipynb -k - -o - 01_1_train_KNN_unique_samples.py | papermill - runs/rev3/01_1_train_KNN_unique_samples.ipynb
+# new comparison (check if the old nb could be used for this purpose)
+jupytext --to ipynb -k - -o - 01_3_revision3.py | papermill - runs/rev3/01_3_revision3.ipynb
+```

project/config/single_dev_dataset/proteinGroups_N50/README.md

@@ -0,0 +1,23 @@
+# Config files
+
+## Version 1 imputation workflow
+
+For [`worflow/Snakefile`](https://github.com/RasmussenLab/pimms/blob/HEAD/project/workflow/Snakefile)
+
+```bash
+config.yaml # main config
+split.yaml # split data config referenced in config.yaml
+train_CF.yaml # CF train config referenced in config.yaml
+train_DAE.yaml # DAE train config referenced in config.yaml
+train_KNN.yaml # KNN train config referenced in config.yaml
+train_Median.yaml # Median train config referenced in config.yaml
+train_VAE.yaml # VAE train config referenced in config.yaml
+```
+
+## Version 2 impuation workflow
+
+For [`workflow/Snakefile_v2.yaml`](https://github.com/RasmussenLab/pimms/blob/HEAD/project/workflow/Snakefile_v2.smk) only one config file is needed:
+
+```bash
+config_v2.yaml
+```