From a8f5e66eb9aa869b3f1a4bcbe1e44953b527521c Mon Sep 17 00:00:00 2001
From: Pavel Dimens <pdimens@live.com>
Date: Tue, 11 May 2021 13:40:11 -0400
Subject: [PATCH] fix trimming rules

---
 rules/trim.smk      | 14 +++++++++-----
 scripts/LepWrapQA.r |  4 ++--
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/rules/trim.smk b/rules/trim.smk
index 147ae4d..9f42242 100644
--- a/rules/trim.smk
+++ b/rules/trim.smk
@@ -24,17 +24,21 @@ rule trim_summary:
     input:
         expand("5_Trim/ordered.{lg}.trimmed", lg = lg_range)
     output:
-        "5_Trim/trim.summary"
+        detailed = "5_Trim/trim.details",
+        summary = "5_Trim/trim.summary"
     message:
         "Summarizing trim logs >> {output}"
     priority: 1
     shell:
         """
-        echo "# this is a summary of which markers were removed from which linkage group via trimming distant edge clusters" >> {output}
-        echo -e "LG\trm_marker" >> {output}
+        echo "# this is a summary of which markers were removed from which linkage group via trimming distant edge clusters" >> {output.detailed}
+        echo -e "LG\trm_marker" >> {output.detailed}
         for each in 5_Trim/logs/ordered.*.removed ; do
             BASE=$(basename $each | cut -d "." -f1,2)
-            sed -e "s/^/$BASE /" $each >> {output}.tmp 
+            sed -e "s/^/$BASE /" $each >> {output.detailed}.tmp 
         done
-        sort -V {output}.tmp > {output} && rm {output}.tmp 
+        sort -V {output.detailed}.tmp > {output.detailed} && rm {output.detailed}.tmp
+        echo "n_removed map" > {output.summary}.tmp
+        cut -d" " -f1 {output.detailed} | uniq -c  >> {output.summary}.tmp
+        column -t {output.summary}.tmp > {output.summary} && rm {output.summary}.tmp 
         """
diff --git a/scripts/LepWrapQA.r b/scripts/LepWrapQA.r
index 601ca6f..9e6efac 100755
--- a/scripts/LepWrapQA.r
+++ b/scripts/LepWrapQA.r
@@ -30,8 +30,8 @@ filename_trunc <- paste(filename_trunc[1], filename_trunc[2], sep = ".")
 # instantiate QC columns
 lgfile$Mpass <- c(TRUE)
 lgfile$Fpass <- c(TRUE)
-outfile_base <- paste(path, "4_OrderMarkers", "best.trim", filename_trunc, sep = "/")
-outfile_log_base <- paste(path, "4_OrderMarkers", "logs/trimming", filename_trunc, sep = "/")
+outfile_base <- paste(path, "5_Trim", filename_trunc, sep = "/")
+outfile_log_base <- paste(path, "5_Trim", "logs", filename_trunc, sep = "/")
 
 #========= PDF instantiation ========#
 PDFPath <- paste(outfile_log_base, "trim.pdf", sep = ".")