Merge pull request #48 from eseiler/misc/util_fix

[MISC] Adjust util scripts
seqan · Aug 10, 2021 · 02312c4 · 02312c4
2 parents 5f1588a + 0f3044d
commit 02312c4
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 5 deletions.
diff --git a/util/src/bash_scripts/benchmark.sh b/util/src/bash_scripts/benchmark.sh
@@ -19,6 +19,8 @@ for i in $(seq -f "$BIN_DIR/$BIN_NUMBER/bins/bin_%0${#BIN_NUMBER}g.fasta" 0 1 $(
 do
     cp $i $working_directory/bins/
 done
+seq -f "$working_directory/bins/bin_%0${#BIN_NUMBER}g.fasta" 0 1 $((BIN_NUMBER-1)) > $working_directory/bins.list
+
 cp $BIN_DIR/$BIN_NUMBER/reads_e$ERRORS\_$READ_LENGTH/all.fastq $working_directory/reads/
 
 do_task () {
@@ -33,7 +35,7 @@ do_task () {
             --size $SIZE \
             --threads $THREADS \
             --hash $HASH \
-            $(seq -f "$working_directory/bins/bin_%0${#BIN_NUMBER}g.fasta" 0 1 $((BIN_NUMBER-1)))
+            $working_directory/bins.list
 
     query_log=$working_directory/$w\_$k\_$SIZE\_query.log # Does not contain HASH
     query_out=$working_directory/$w\_$k\_$SIZE.out
@@ -43,8 +45,6 @@ do_task () {
             --query $working_directory/reads/all.fastq \
             --index $ibf_filename \
             --output $query_out \
-            --kmer $k \
-            --window $w \
             --threads $THREADS \
             --error $ERRORS \
             --pattern $READ_LENGTH \

diff --git a/util/src/evaluation_scripts/eval_benchmark.py b/util/src/evaluation_scripts/eval_benchmark.py
@@ -20,8 +20,13 @@ def process_output(path, bin_count, read_count, path_to_fp, path_to_fn):
                     try:
                         [x, y, thr, count] = line.strip().split('\t')
                     except ValueError:
-                        fn += 1
-                        continue
+                        try:
+                            [x, y] = line.strip().split('\t')
+                            thr = -1
+                            count = -1
+                        except ValueError:
+                            fn += 1
+                            continue
                     [read_id, bins] = [int(x), [int(e) for e in y[:-1].split(',') if e != '']]
                     true_id = (read_id % read_count) // (read_count // bin_count)
                     if true_id in bins: