Merge pull request #29 from SchlossLab/report

Support Snakemake's report feature

.snakemake-workflow-catalog.yml

@@ -5,3 +5,4 @@ usage:
     conda: true 
     singularity: true 
     singularity+conda: true
+  report: true

workflow/Snakefile

@@ -32,6 +32,9 @@ include: "rules/plot.smk"
 include: "rules/example-report.smk"
 
 
+report: "report/workflow.rst"
+
+
 rule targets:
     input:
         f"report_{dataset}.md",

workflow/report/benchmarks.rst

@@ -0,0 +1,2 @@
+Each model training run was given {{ snakemake.config['ncores'] }} cores for 
+parallelization.

workflow/report/workflow.rst

@@ -0,0 +1,6 @@
+Machine learning algorithms used were: {{ snakemake.config['ml_methods'] }}.
+Models were trained with {{ snakemake.config['nseeds'] }} 
+different random partitions of the 
+{{ snakemake.config['dataset_name'] }} dataset 
+into training and testing sets using 
+{{ snakemake.config['kfold'] }}-fold cross validation.

workflow/rules/plot.smk

@@ -2,7 +2,11 @@ rule plot_performance:
     input:
         csv="results/{dataset}/performance_results.csv",
     output:
-        plot="figures/{dataset}/performance.png",
+        plot=report(
+            "figures/{dataset}/performance.png",
+            category="Performance",
+            subcategory="Model Performance",
+        ),
     log:
         "log/{dataset}/plot_performance.txt",
     conda:
@@ -17,7 +21,10 @@ if find_feature_importance:
         input:
             csv="results/{dataset}/feature-importance_results.csv",
         output:
-            plot="figures/{dataset}/feature_importance.png",
+            plot=report(
+                "figures/{dataset}/feature_importance.png",
+                category="Feature Importance",
+            ),
         log:
             "log/{dataset}/plot_feature_importance.txt",
         conda:
@@ -29,7 +36,10 @@ else:
 
     rule make_blank_feature_plot:
         output:
-            plot="figures/{dataset}/feature_importance.png",
+            plot=report(
+                "figures/{dataset}/feature_importance.png",
+                category="Feature Importance",
+            ),
         log:
             "log/{dataset}/make_blank_plot.txt",
         conda:
@@ -42,7 +52,11 @@ rule plot_hp_performance:
     input:
         rds="results/{dataset}/hp_performance_results_{method}.Rds",
     output:
-        plot="figures/{dataset}/hp_performance_{method}.png",
+        plot=report(
+            "figures/{dataset}/hp_performance_{method}.png",
+            category="Performance",
+            subcategory="Hyperparameter Tuning",
+        ),
     log:
         "log/{dataset}/plot_hp_perf_{method}.txt",
     conda:
@@ -55,7 +69,12 @@ rule plot_benchmarks:
     input:
         csv="results/{dataset}/benchmarks_results.csv",
     output:
-        plot="figures/{dataset}/benchmarks.png",
+        plot=report(
+            "figures/{dataset}/benchmarks.png",
+            category="Performance",
+            subcategory="Runtime & Memory Usage",
+            caption="../report/benchmarks.rst",
+        ),
     log:
         "log/{dataset}/plot_benchmarks.txt",
     conda: