✨ Integrate data splitting config into main config

- allow to set frac_mnar from commandline using:
   --config frac_mnar=.5
- dump created data config using separate rule (into experiment folder)

project/config/knn_comparison/ald_pgs_all/config.yaml

@@ -1,9 +1,15 @@
-config_split: config/knn_comparison/ald_pgs_all/split.yaml
 config_train: runs/knn_comparison/ald_pgs_all/configs_train/train_{model}.yaml
 folder_experiment: runs/knn_comparison/ald_pgs_all
 fn_rawfile_metadata: data/ALD_study/processed/raw_meta.csv
 file_format: pkl
 cuda: False
+split_data:
+    FN_INTENSITIES: data/ALD_study/processed/ald_plasma_proteinGroups.pkl
+    sample_completeness: 0.5
+    min_RT_time: 20
+    column_names:
+        - PG.ProteinAccessions
+# frac_mnar: 0.0
 models:
     - Median:
           model: Median

project/workflow/Snakefile_v2

@@ -32,8 +32,10 @@ rule all:
 
 nb = "01_2_performance_plots.ipynb"
 
-print(config["models"])
+if "frac_mnar" in config:
+    config["split_data"]["frac_mnar"] = config["frac_mnar"]
 
+# print(config['split_data'])
 # MODELS = config["models"].copy()
 
 MODELS = list()
@@ -164,7 +166,9 @@ rule transform_data_to_wide_format:
 # train models in python
 rule train_models:
     input:
-        nb=lambda wildcards: "01_1_train_{}.ipynb".format(model_configs[wildcards.model]["model"]),
+        nb=lambda wildcards: "01_1_train_{}.ipynb".format(
+            model_configs[wildcards.model]["model"]
+        ),
         train_split="{folder_experiment}/data/train_X.csv",
         configfile=config["config_train"],
     output:
@@ -224,4 +228,18 @@ rule create_splits:
         " -f {input.configfile}"
         " -r folder_experiment {params.folder_experiment}"
         " -p fn_rawfile_metadata {params.meta_data}"
-        " && jupyter nbconvert --to html {output.nb}"
+        " && jupyter nbconvert --to html {output.nb}"
+
+
+##########################################################################################
+# create config file dumps for each model
+
+
+rule dump_split_config:
+    output:
+        configfile=f"{folder_experiment}/{nb_stem}.yaml",
+    run:
+        import yaml
+
+        with open(output.configfile, "w") as f:
+            yaml.dump(config["split_data"], f)