Remove old cluster batching code

ark/phenotyping/run_trained_som.R

@@ -8,7 +8,6 @@
 # - normValsPath: path to the 99.9% normalized values file
 # - pixelWeightsPath: path to the SOM weights file
 # - pixelClusterDir: path to directory where the clustered data will be written to
-# - batchSize: number of fovs to cluster at once
 
 library(arrow)
 library(data.table)
@@ -32,9 +31,6 @@ pixelWeightsPath <- args[4]
 # get the cluster write path directory
 pixelClusterDir <- args[5]
 
-# get the batch size
-batchSize <- strtoi(args[6])
-
 # read the weights
 somWeights <- as.matrix(arrow::read_feather(pixelWeightsPath))
 
@@ -44,59 +40,8 @@ normVals <- as.matrix(arrow::read_feather(normValsPath))
 # get the marker names from the weights matrix
 markers <- colnames(somWeights)
 
-# divide the fovs into batches
-fovBatches <- split(fovs, cut(seq_along(fovs), length(fovs) / batchSize, labels=FALSE))
-
-print("Mapping data to cluster labels")
-start <- proc.time()
-batchNum <- 1
-for (fovs in fovBatches) {
-    # create pixel data to cluster
-    batchPixelData <- matrix(nrow=0, ncol=length(markers))
-    colnames(batchPixelData) <- markers
-
-    batchFileName <- sprintf("cluster_%s.feather", batchNum)
-
-    for (fov in fovs) {
-        fovFileName <- paste(fov, ".feather", sep="")
-        matPath <- paste(pixelMatDir, fovFileName, sep="/")
-        fovPixelData <- as.matrix(arrow::read_feather(matPath, col_select=all_of(markers)))
-
-        batchPixelData <- rbind(batchPixelData, fovPixelData)
-    }
-
-    # 99.9% normalize pixel data
-    for (marker in markers) {
-        # this prevents all- or mostly-zero columns from getting normalized and becoming NA/Inf
-        if (normVals[1, marker] != 0) {
-            batchPixelData[, marker] = batchPixelData[, marker] / normVals[1, marker]
-        }
-    }
-
-    print(batchPixelData)
-
-    # map FlowSOM data
-    clusters <- FlowSOM:::MapDataToCodes(somWeights, batchPixelData)
-
-    # assign cluster labels column to pixel data
-    batchPixelData <- as.matrix(cbind(as.matrix(batchPixelData), cluster=clusters[,1]))
-
-    # write to feather
-    clusterPath <- paste(pixelClusterDir, batchFileName, sep="/")
-    arrow::write_feather(as.data.table(batchPixelData), clusterPath)
-
-    batchNum <- batchNum + 1
-
-    # print an update every 10 fovs
-    if (batchNum %% 10 == 0) {
-        sprintf("Finished clustering %s fovs", i)
-    }
-}
-print(proc.time() - start)
-
 # using trained SOM, batch cluster the original dataset by fov
 print("Mapping data to cluster labels")
-start <- proc.time()
 for (i in 1:length(fovs)) {
     # read in pixel data
     fileName <- paste(fovs[i], ".feather", sep="")
@@ -126,6 +71,5 @@ for (i in 1:length(fovs)) {
         sprintf("Finished clustering %s fovs", i)
     }
 }
-print(proc.time() - start)
 
 print("Done!")

ark/phenotyping/som_utils.py

@@ -288,7 +288,7 @@ def train_som(fovs, channels, base_dir,
 
 def cluster_pixels(fovs, base_dir, pre_dir='pixel_mat_preprocessed',
                    norm_vals_name='norm_vals.feather', weights_name='weights.feather',
-                   cluster_dir='pixel_mat_clustered', batch_size=5):
+                   cluster_dir='pixel_mat_clustered'):
     """Uses trained weights to assign cluster labels on full pixel data
 
     Saves data with cluster labels to cluster_dir
@@ -307,8 +307,6 @@ def cluster_pixels(fovs, base_dir, pre_dir='pixel_mat_preprocessed',
             The name of the weights file
         cluster_dir (str):
             The name of the directory to write the clustered data
-        batch_size (int):
-            Number of fovs to cluster at once
     """
 
     # define the paths to the data
@@ -353,8 +351,7 @@ def cluster_pixels(fovs, base_dir, pre_dir='pixel_mat_preprocessed',
 
     # run the trained SOM on the dataset, assigning clusters
     process_args = ['Rscript', '/run_trained_som.R', ','.join(fovs),
-                    preprocessed_path, norm_vals_path, weights_path,
-                    clustered_path, str(batch_size)]
+                    preprocessed_path, norm_vals_path, weights_path, clustered_path]
 
     process = subprocess.Popen(process_args, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)