changeset 9:b3ca3ece8a8b draft

WIP detect/trap "bad primer product range" errors.

--- a/README.rst	Tue Jun 06 05:56:12 2017 -0400
+++ b/README.rst	Fri Jun 16 09:12:09 2017 -0400
@@ -61,6 +61,9 @@
 Version    Changes
 ---------- ----------------------------------------------------------------------
 
+0.02.04.6  - Trap for errors in ``pal_finder_v0.02.04.pl`` resulting in bad
+             ranges being supplied to ``primer3_core`` for some reads via
+             ``PRIMER_PRODUCT_RANGE_SIZE``.
 0.02.04.5  - Update to handle large output files which can sometimes be generated
              by the ``pal_finder_v0.02.04.pl`` or ``pal_filter.py`` scripts (logs
              of hundreds of Gb's have been observed in production): log files

--- a/pal_finder_wrapper.sh	Tue Jun 06 05:56:12 2017 -0400
+++ b/pal_finder_wrapper.sh	Fri Jun 16 09:12:09 2017 -0400
@@ -53,6 +53,9 @@
 # Maximum size reporting log file contents
 MAX_LINES=500
 #
+# Get helper functions
+. $(dirname $0)/pal_finder_wrapper_utils.sh
+#
 # Initialise locations of scripts, data and executables
 #
 # Set these in the environment to overide at execution time
@@ -63,31 +66,18 @@
 # Filter script is in the same directory as this script
 PALFINDER_FILTER=$(dirname $0)/pal_filter.py
 if [ ! -f $PALFINDER_FILTER ] ; then
-    echo No $PALFINDER_FILTER script >&2
-    exit 1
+    fatal No $PALFINDER_FILTER script
 fi
 #
 # Check that we have all the components
-function have_program() {
-    local program=$1
-    local got_program=$(which $program 2>&1 | grep "no $(basename $program) in")
-    if [ -z "$got_program" ] ; then
-	echo yes
-    else
-	echo no
-    fi	
-}
 if [ "$(have_program $PRIMER3_CORE_EXE)" == "no" ] ; then
-    echo "ERROR primer3_core missing: ${PRIMER3_CORE_EXE} not found" >&2
-    exit 1
+    fatal "primer3_core missing: ${PRIMER3_CORE_EXE} not found"
 fi
 if [ ! -f "${PALFINDER_DATA_DIR}/config.txt" ] ; then
-    echo "ERROR pal_finder config.txt not found in ${PALFINDER_DATA_DIR}" >&2
-    exit 1
+    fatal "pal_finder config.txt not found in ${PALFINDER_DATA_DIR}"
 fi
 if [ ! -f "${PALFINDER_SCRIPT_DIR}/pal_finder_v0.02.04.pl" ] ; then
-    echo "ERROR pal_finder_v0.02.04.pl not found in ${PALFINDER_SCRIPT_DIR}" >&2
-    exit 1
+    fatal "pal_finder_v0.02.04.pl not found in ${PALFINDER_SCRIPT_DIR}"
 fi
 #
 # Initialise parameters used in the config.txt file
@@ -118,7 +108,7 @@
 if [ $# -lt 2 ] ; then
   echo "Usage: $0 FASTQ_R1 FASTQ_R2 MICROSAT_SUMMARY PAL_SUMMARY [OPTIONS]"
   echo "       $0 --454    FASTA    MICROSAT_SUMMARY PAL_SUMMARY [OPTIONS]"
-  exits
+  fatal "Bad command line"
 fi
 if [ "$1" == "--454" ] ; then
     PLATFORM="454"
@@ -235,16 +225,14 @@
 # Check that primer3_core is available
 got_primer3=`which $PRIMER3_CORE_EXE 2>&1 | grep -v "no primer3_core in"`
 if [ -z "$got_primer3" ] ; then
-  echo ERROR primer3_core not found >&2
-  exit 1
+  fatal "primer3_core not found"
 fi
 #
 # Set up the working dir
 if [ "$PLATFORM" == "Illumina" ] ; then
     # Paired end Illumina data as input
     if [ $FASTQ_R1 == $FASTQ_R2 ] ; then
-	echo ERROR R1 and R2 fastqs are the same file >&2
-	exit 1
+	fatal ERROR R1 and R2 fastqs are the same file
     fi
     ln -s $FASTQ_R1
     ln -s $FASTQ_R2
@@ -264,17 +252,6 @@
 /bin/cp $PALFINDER_DATA_DIR/config.txt .
 #
 # Update the config.txt file with new values
-function set_config_value() {
-    local key=$1
-    local value=$2
-    local config_txt=$3
-    if [ -z "$value" ] ; then
-	echo "No value for $key, left as default"
-    else
-	echo Setting "$key" to "$value"
-	sed -i 's,^'"$key"' .*,'"$key"'  '"$value"',' $config_txt
-    fi
-}
 # Input files
 set_config_value platform $PLATFORM config.txt
 if [ "$PLATFORM" == "Illumina" ] ; then
@@ -299,6 +276,7 @@
 # Primer3 settings
 set_config_value primer3input Output/pr3in.txt config.txt
 set_config_value primer3output Output/pr3out.txt config.txt
+set_config_value keepPrimer3files 1 config.txt
 set_config_value primer3executable $PRIMER3_CORE_EXE config.txt
 set_config_value prNamePrefix ${PRIMER_PREFIX}_ config.txt
 set_config_value PRIMER_MISPRIMING_LIBRARY "$PRIMER_MISPRIMING_LIBRARY" config.txt
@@ -329,8 +307,28 @@
 #
 # Check that log ends with "Done!!" message
 if [ -z "$(tail -n 1 pal_finder.log | grep Done!!)" ] ; then
-    echo ERROR pal_finder failed to complete successfully >&2
-    exit 1
+    fatal ERROR pal_finder failed to complete successfully
+fi
+echo "### pal_finder finished ###"
+#
+# Check for errors in pal_finder output
+echo "### Checking for errors ###"
+if [ ! -z "$(grep 'primer3_core: Illegal element in PRIMER_PRODUCT_SIZE_RANGE' pal_finder.log)" ] ; then
+    echo ERROR primer3 terminated prematurely due to bad product size ranges
+    cat >&2 <<EOF
+ERROR primer3 terminated prematurely due to bad product size ranges
+
+Pal_finder generated bad ranges for the following read IDs:
+EOF
+    $(find_bad_primer_ranges Output/pr3in.txt) >&2
+    cat >&2 <<EOF
+
+This error can occur when input data contains short R1 reads and has
+has not been properly trimmed and filtered.
+
+EOF
+    fatal pal_finder failed to complete successfully
+EOF
 fi
 #
 # Sort outputs into a consistent order regardless of Perl version
@@ -354,11 +352,9 @@
     fi
     tail -$MAX_LINES pal_filter.log
     if [ $? -ne 0 ] ; then
-	echo ERROR $PALFINDER_FILTER exited with non-zero status >&2
-	exit 1
+	fatal $PALFINDER_FILTER exited with non-zero status
     elif [ ! -f PAL_summary.filtered ] ; then
-	echo ERROR no output from $PALFINDER_FILTER >&2
-	exit 1
+	fatal no output from $PALFINDER_FILTER
     fi
 fi
 #
@@ -378,8 +374,7 @@
     if [ -f "$assembly" ] ; then
 	/bin/mv $assembly "$OUTPUT_ASSEMBLY"
     else
-	echo ERROR no assembly output found >&2
-	exit 1
+	fatal no assembly output found
     fi
 fi
 if [ ! -z "$OUTPUT_CONFIG_FILE" ] && [ -f config.txt ] ; then

--- a/pal_finder_wrapper.xml	Tue Jun 06 05:56:12 2017 -0400
+++ b/pal_finder_wrapper.xml	Fri Jun 16 09:12:09 2017 -0400
@@ -274,6 +274,29 @@
 
 -------------
 
+.. class:: warning
+
+**Known problems**
+
+.. class:: infomark
+
+**Bad primer product size ranges**
+
+For some datasets pal_finder may generate 'bad' product size ranges (where the
+lower limit exceeds the upper limit) for one or more reads, for input into
+primer3_core.
+
+If this occurs then the tool will terminate with an error. A list of the reads
+for which the bad ranges were generated can be found in the error message
+which can be accessed via the 'bug' icon from a failed dataset.
+
+The conditions which cause this error are unclear. However we believe it to be
+associated with short or low quality reads. It is recommended that the input
+data are sufficiently trimmed and filtered (using e.g. the Trimmomatic tool)
+before rerunning pal_finder.
+
+-------------
+
 .. class:: infomark
 
 **Credits**