Merge pull request #46 from SchlossLab/iss-44

Fix seed type & feature importance column

Author: kelly-sovacool

.github/workflows/build.yml

@@ -9,19 +9,15 @@ jobs:
   build:
     name: build
     runs-on: ubuntu-latest
-    strategy:
-      matrix:
-        python-version: ["3.11"]
 
     steps:
       - uses: actions/checkout@v3
         with:
           persist-credentials: false
           fetch-depth: 0
-      - uses: conda-incubator/setup-miniconda@v2
+      - uses: conda-incubator/setup-miniconda@v3
         with:
           python-version: 3.11
-          miniforge-variant: Mambaforge
           miniforge-version: latest
           activate-environment: github-actions
           environment-file: workflow/envs/github-actions.yml

Dockerfile

@@ -1,8 +1,8 @@
-FROM condaforge/mambaforge:latest
+FROM condaforge/miniforge3:latest
 LABEL io.github.snakemake.containerized="true"
-LABEL io.github.snakemake.conda_env_hash="6aa289536136aae2d34bac6dce9ce47d037da888ed09e2c8ada989c90ef10658"
+LABEL io.github.snakemake.conda_env_hash="c1edb6a4917211d511a661c54768e671f7f067b1ea473011e8fdcbc485178d2c"
 
-# Step 1: Retrieve conda environments
+# Step 2: Retrieve conda environments
 
 # Conda environment:
 #   source: workflow/envs/graphviz.yml
@@ -17,7 +17,7 @@ COPY workflow/envs/graphviz.yml /conda-envs/b42323b0ffd5d034544511c9db1bdead/env
 
 # Conda environment:
 #   source: workflow/envs/mikropml.yml
-#   prefix: /conda-envs/3f83a46ff5ea715a12fde6ee46136b0b
+#   prefix: /conda-envs/e7c23e20e8aab7662ae81be2ad57d998
 #   name: mikropml
 #   channels:
 #     - conda-forge
@@ -30,15 +30,15 @@ COPY workflow/envs/graphviz.yml /conda-envs/b42323b0ffd5d034544511c9db1bdead/env
 #     - r-future
 #     - r-future.apply
 #     - r-import
-#     - r-mikropml>=1.5.0
+#     - r-mikropml>=1.6.0
 #     - r-patchwork
 #     - r-rmarkdown
 #     - r-rpart
 #     - r-purrr
 #     - r-schtools>=0.4.0
 #     - r-tidyverse
-RUN mkdir -p /conda-envs/3f83a46ff5ea715a12fde6ee46136b0b
-COPY workflow/envs/mikropml.yml /conda-envs/3f83a46ff5ea715a12fde6ee46136b0b/environment.yaml
+RUN mkdir -p /conda-envs/e7c23e20e8aab7662ae81be2ad57d998
+COPY workflow/envs/mikropml.yml /conda-envs/e7c23e20e8aab7662ae81be2ad57d998/environment.yaml
 
 # Conda environment:
 #   source: workflow/envs/smk.yml
@@ -54,9 +54,9 @@ COPY workflow/envs/mikropml.yml /conda-envs/3f83a46ff5ea715a12fde6ee46136b0b/env
 RUN mkdir -p /conda-envs/457b7b75191d44b96e5086432876e333
 COPY workflow/envs/smk.yml /conda-envs/457b7b75191d44b96e5086432876e333/environment.yaml
 
-# Step 2: Generate conda environments
+# Step 3: Generate conda environments
 
-RUN mamba env create --prefix /conda-envs/b42323b0ffd5d034544511c9db1bdead --file /conda-envs/b42323b0ffd5d034544511c9db1bdead/environment.yaml && \
-    mamba env create --prefix /conda-envs/3f83a46ff5ea715a12fde6ee46136b0b --file /conda-envs/3f83a46ff5ea715a12fde6ee46136b0b/environment.yaml && \
-    mamba env create --prefix /conda-envs/457b7b75191d44b96e5086432876e333 --file /conda-envs/457b7b75191d44b96e5086432876e333/environment.yaml && \
-    mamba clean --all -y
+RUN conda env create --prefix /conda-envs/b42323b0ffd5d034544511c9db1bdead --file /conda-envs/b42323b0ffd5d034544511c9db1bdead/environment.yaml && \
+    conda env create --prefix /conda-envs/e7c23e20e8aab7662ae81be2ad57d998 --file /conda-envs/e7c23e20e8aab7662ae81be2ad57d998/environment.yaml && \
+    conda env create --prefix /conda-envs/457b7b75191d44b96e5086432876e333 --file /conda-envs/457b7b75191d44b96e5086432876e333/environment.yaml && \
+    conda clean --all -y

workflow/envs/mikropml.yml

@@ -10,7 +10,7 @@ dependencies:
   - r-future
   - r-future.apply
   - r-import
-  - r-mikropml>=1.5.0
+  - r-mikropml>=1.6.0
   - r-patchwork
   - r-rmarkdown
   - r-rpart

workflow/scripts/find_feature_importance.R

@@ -36,7 +36,8 @@ feat_imp <- mikropml::get_feature_importance(
   seed = seed,
 )
 
-wildcards <- schtools::get_wildcards_tbl()
+wildcards <- schtools::get_wildcards_tbl() %>%
+  dplyr::mutate(seed = as.numeric(seed))
 
 readr::write_csv(
   feat_imp %>%

workflow/scripts/plot_feature_importance.R

@@ -7,13 +7,13 @@ feat_df <- readr::read_csv(snakemake@input[["csv"]])
 top_n <- as.numeric(snakemake@params[["top_n"]])
 
 top_feats <- feat_df %>%
-  group_by(method, names) %>%
+  group_by(method, feat) %>%
   summarize(median_diff = median(perf_metric_diff)) %>%
   slice_max(order_by = median_diff, n = top_n)
 
 feat_plot <- feat_df %>%
-  right_join(top_feats, by = c("method", "names")) %>%
-  mutate(features = factor(names, levels = unique(top_feats$names))) %>%
+  right_join(top_feats, by = c("method", "feat")) %>%
+  mutate(features = factor(feat, levels = unique(top_feats$feat))) %>%
   ggplot(aes(x = perf_metric_diff, y = features, color = method)) +
   geom_boxplot() +
   facet_wrap(~method) +

workflow/scripts/train_ml.R

@@ -18,7 +18,8 @@ ml_results <- mikropml::run_ml(
   hyperparameters = hyperparams
 )
 
-wildcards <- schtools::get_wildcards_tbl()
+wildcards <- schtools::get_wildcards_tbl() %>%
+  dplyr::mutate(seed = as.numeric(seed))
 
 readr::write_csv(
   ml_results$performance %>%