Merge pull request #28 from uclahs-cds/jarbet-cluster-paths

remove cluster paths

Author: jarbet

.github/workflows/R-CMD-check.yaml

@@ -24,7 +24,7 @@ jobs:
 
       # Step 3: Cache R packages
       - name: Cache R packages
-        uses: actions/cache@v2
+        uses: actions/cache@v3
         with:
           path: ~/.cache/R
           key: ${{ runner.os }}-R-${{ hashFiles('**/*.R') }}

inst/example.data.R

@@ -1,8 +1,8 @@
 devtools::load_all();
+source('config.R') # see project PRAD-000101-MethySubtypes/PrCaMethy/config.R
 library(impute);
-path.data <- '/hot/project/disease/ProstateTumor/PRAD-000101-MethySubtypes/data/2024-07-16_pooled_tumour_normal_all_cpgs_all_cohorts.rds';
 
-example.data <- readRDS(path.data);
+example.data <- readRDS(arg$path.example.methy.data);
 rownames(example.data) <- example.data$patient.id;
 example.data <- subset(example.data, subset = cohort == 'TCGA', select = - c(patient.id, cohort));
 

inst/gene.promoter.cpgi.R

@@ -1,10 +1,10 @@
 devtools::load_all();
+source('config.R') # see project PRAD-000101-MethySubtypes/PrCaMethy/config.R
 
 data(cpg.annotation);
 #data(example.data);
 #methy <- example.data; # methylation data with patient ids as rownames, CpGs as columns
-path.data <- '/hot/project/disease/ProstateTumor/PRAD-000101-MethySubtypes/data/2024-07-16_pooled_tumour_normal_all_cpgs_all_cohorts.rds';
-methy <- readRDS(path.data);
+methy <- readRDS(arg$path.example.methy.data);
 print.progress <- TRUE;
 methy.cpgs <- colnames(methy)[grep('^cg', colnames(methy))];
 methy <- NULL;

inst/save.models.R

@@ -1,13 +1,12 @@
+source('config.R') # see project PRAD-000101-MethySubtypes/PrCaMethy/config.R
 compress <- 'xz';
 test.mode <- FALSE;
 
-path.ml.res <- '/hot/project/disease/ProstateTumor/PRAD-000101-MethySubtypes/output/prediction/2025-02-05_F72-predict-clinical-and-drivers_gene-methy_association-filter_discrete-methyFALSE_updated.RData';
-path.final.models <- '/hot/project/disease/ProstateTumor/PRAD-000101-MethySubtypes/output/prediction/2025-02-05_PrCaMethy_final_models.rds';
-load(path.ml.res);
-final.models <- readRDS(path.final.models);
+load(arg$path.ml.res);
+final.models <- readRDS(arg$path.final.models);
 
 # extract date from path.ml.res
-res.date <- gsub('.*/output/prediction/(\\d{4}-\\d{2}-\\d{2}).*', '\\1', path.ml.res);
+res.date <- gsub('.*/output/prediction/(\\d{4}-\\d{2}-\\d{2}).*', '\\1', arg$path.ml.res);
 res.date;
 
 outcomes <- unique(ml.res.params$outcome);
@@ -67,11 +66,11 @@ models <- lapply(
             ml.res.params$top.features == final.model$top.features
             );
 
-        file <- file.path(dirname(path.ml.res), paste0(res.date, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', final.model$top.features, '.RData'));
+        file <- file.path(dirname(arg$path.ml.res), paste0(res.date, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', final.model$top.features, '.RData'));
 
         if (!file.exists(file)) {
             res.date2 <- as.Date(res.date) - 1;
-            file <- file.path(dirname(path.ml.res), paste0(res.date2, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', final.model$top.features, '.RData'));
+            file <- file.path(dirname(arg$path.ml.res), paste0(res.date2, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', final.model$top.features, '.RData'));
             }
         stopifnot(file.exists(file));
 

inst/save.models.explore.model.size.R

@@ -1,19 +1,19 @@
 devtools::load_all();
-res.date <- '2025-01-13';
+source('config.R') # see project PRAD-000101-MethySubtypes/PrCaMethy/config.R
 
-discrete.methy <- FALSE;
 compress <- 'xz';
 test.mode <- FALSE;
 
-path.ml.res <- paste0('/hot/project/disease/ProstateTumor/PRAD-000101-MethySubtypes/output/prediction/', res.date, '_F72-predict-clinical-and-drivers_gene-methy_association-filter_discrete-methy', discrete.methy, '.RData');
+arg$path.ml.res <- arg$path.ml.res;
+res.date <- regmatches(arg$path.ml.res, regexpr('\\d{4}-\\d{2}-\\d{2}', arg$path.ml.res));
 tolerances <- c(0, 0.01, 0.02, 0.03, 0.04, 0.05) # use smallest model within __ of best model
 
 model.size <- lapply(
     X = seq_along(tolerances),
     FUN = function(z) {
         tolerance <- tolerances[z];
 
-        load(path.ml.res);
+        load(arg$path.ml.res);
 
         outcomes <- unique(ml.res.params$outcome);
 
@@ -80,11 +80,11 @@ model.size <- lapply(
 
                 mod.id <- which(ml.res.params$outcome == outcome & ml.res.params$top.features == res$top.features);
 
-                file <- file.path(dirname(path.ml.res), paste0(res.date, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', res$top.features, '.RData'));
+                file <- file.path(dirname(arg$path.ml.res), paste0(res.date, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', res$top.features, '.RData'));
 
                 if (!file.exists(file)) {
                     res.date2 <- as.Date(res.date) - 1;
-                    file <- file.path(dirname(path.ml.res), paste0(res.date2, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', res$top.features, '.RData'));
+                    file <- file.path(dirname(arg$path.ml.res), paste0(res.date2, '_F72-predict-clinical-and-drivers_discrete-methyFALSE_models-', mod.id, '-', outcome, '-', res$top.features, '.RData'));
                     }
                 stopifnot(file.exists(file));
 

inst/subtype.model.R

@@ -1,6 +1,5 @@
-path.subtype.model <- '/hot/project/disease/ProstateTumor/PRAD-000101-MethySubtypes/data/2024-11-13_example-data_for-predicting-subtypes.Rdata';
-
-load(path.subtype.model);
+source('config.R') # see project PRAD-000101-MethySubtypes/PrCaMethy/config.R
+load(arg$path.subtype.model);
 
 subtype.model <- model.for.predicting.subtypes;
 subtype.model.required.cpgs <- rownames(subtype.model$centroids);