b38a548970c6920b3c9d7b6d2685e88cc1797365
angie
  Wed May 27 16:38:35 2020 -0700
Remove pre-existing results before installing updates because it's possible for a clade to suddenly not exist in a new release (e.g. clade A2 disappeared in second update today).  refs #25188

diff --git src/hg/utils/otto/nextstrainNcov/doUpdate.sh src/hg/utils/otto/nextstrainNcov/doUpdate.sh
index 2ba399e..a03adc3 100755
--- src/hg/utils/otto/nextstrainNcov/doUpdate.sh
+++ src/hg/utils/otto/nextstrainNcov/doUpdate.sh
@@ -1,116 +1,121 @@
 #!/bin/bash
 set -beEu -o pipefail
 
 #	Do not modify this script, modify the source tree copy:
 #	kent/src/hg/utils/nextstrainNcov/doUpdate.sh
 
 ottoDir=/hive/data/outside/otto/nextstrainNcov
 chromSizes=/hive/data/genomes/wuhCor1/chrom.sizes
 gbdbDir=/gbdb/wuhCor1/nextstrain
 
 ncovJsonUrl='https://nextstrain.org/charon/getDataset?prefix=/ncov/global'
 
 cd $ottoDir
 
 # The file is named .json, but is actually gzipped, so gunzip it.
 rm -f ncov.json
 curl -s "$ncovJsonUrl" \
 | gunzip -c > ncov.json
 curl -s -I "$ncovJsonUrl" \
 | grep Last-Mod | sed -re 's/Last-Modified: //; s/\r//;' \
    > ncov.json.date
 latestDate=$(cat ncov.json.date)
 oldDate=$(cat old.ncov.json.date)
 if [ $(date -d "$latestDate" +%s) -le $(date -d "$oldDate" +%s)  ]; then
     exit 0
 fi
 
 today=`date +%F`
 mkdir -p $today
 ncovTime=$(date -d "$latestDate" +%F-%H:%M)
 cp -p ncov.json $today/ncov.$ncovTime.json
 mv -f ncov.json old.ncov.json
 mv ncov.json.date old.ncov.json.date
 
 runDir=$ottoDir/$today
 cd $runDir
 chmod 444 ncov.$ncovTime.json
 ln -sf ncov.$ncovTime.json ncov.json
 
-#Generate bed and VCF files
+# remove older result files in case a clade went away, as A2 did in ncov.2020-05-27-08:06.json
+rm -f *.vcf.gz* *.bedGraph *.bigWig *.nh
+
+#Generate bed, VCF etc. files
 $ottoDir/nextstrain.py
 
 # bgzip & tabix the VCF files
-bgzip -f nextstrainSamples.vcf
-tabix -p vcf nextstrainSamples.vcf.gz
-for clade in A1a A2 A2a A3 A6 A7 B B1 B2 B4; do
-  bgzip -f nextstrainSamples$clade.vcf
-  tabix -p vcf nextstrainSamples$clade.vcf.gz
+for f in nextstrain*.vcf; do
+  bgzip -f $f
+  tabix -p vcf $f.gz
 done
-bgzip -f nextstrainRecurrentBiallelic.vcf
-tabix -p vcf nextstrainRecurrentBiallelic.vcf.gz
 
 # bigBed-ify the gene names, "clades" and discarded/blacklisted/informative tracks for David
 bedToBigBed -type=bed4 -tab -verbose=0 nextstrainGene.bed $chromSizes \
     nextstrainGene.bb
 
 sort -k2n,2n nextstrainClade.bed > nextstrainClade.sorted.bed
 bedToBigBed -as=$ottoDir/nextstrainClade.as -type=bed12+7 -tab -verbose=0 \
     nextstrainClade.sorted.bed $chromSizes \
     nextstrainClade.bb
 
 bedToBigBed -type=bed4 -tab -verbose=0 nextstrainDiscarded.bed $chromSizes \
     nextstrainDiscarded.bb
 
 bedToBigBed -type=bed4 -tab -verbose=0 nextstrainBlacklisted.bed $chromSizes \
     nextstrainBlacklisted.bb
 
 bedToBigBed -type=bed4 -tab -verbose=0 nextstrainInformative.bed $chromSizes \
     nextstrainInformative.bb
 
 # bigWig for the tree parsimony scores track for David
 bedGraphToBigWig nextstrainParsimony.bedGraph $chromSizes nextstrainParsimony.bw
 
 # Max's nextstrainSamples*.bedGraph allele count bigWigs:
 for i in nextstrainSamples*.vcf.gz; do
     base=`basename $i .vcf.gz`
     zcat $i \
     | grep -v '#' \
     | perl -wne '@w=split("\t");
                  $w[7] =~ m/AC=(\d+)[\d,]*;AN=(\d+)/ ||
                           die "Cant find AC and AN in |$w[7]|";
                  print join("\t", $w[0], $w[1]-1, $w[1], (sprintf "%.06f", $1 / $2)) . "\n";' \
       > $base.bedGraph
     bedGraphToBigWig $base.bedGraph $chromSizes $base.bigWig
 done
 
-# Install
-mkdir -p $ottoDir/current
+# Install public track files
+mkdir $ottoDir/install
 cp -pf $runDir/nextstrainGene.bb $runDir/nextstrainClade.bb \
     $runDir/nextstrain*.vcf.gz{,.tbi} \
     $runDir/nextstrain*.nh \
     $runDir/nextstrainSamples*.bigWig \
-    $ottoDir/current/
+    $ottoDir/install/
+rm -f $ottoDir/current.bak
+mv -f $ottoDir/current $ottoDir/current.bak
+mv $ottoDir/install $ottoDir/current
+rm -r $gbdbDir
+mkdir $gbdbDir
 ln -sf $ottoDir/current/nextstrainGene.bb $ottoDir/current/nextstrainClade.bb \
     $ottoDir/current/nextstrain*.vcf.gz{,.tbi} \
     $ottoDir/current/nextstrain*.nh \
     $ottoDir/current/nextstrainSamples*.bigWig \
     $gbdbDir/
 
 # Install but don't archive (for now) the experimental tracks for David.
 cp -pf $runDir/nextstrain{Discarded,Blacklisted,Informative}.bb \
     $runDir/nextstrainParsimony.bw \
     $ottoDir/current/
 ln -sf $ottoDir/current/nextstrain{Discarded,Blacklisted,Informative}.bb \
     $ottoDir/current/nextstrainParsimony.bw \
     $gbdbDir/
 
 # Daily archive (may overwrite files from earlier today)
 mkdir -p $ottoDir/archive/$today
+rm -f $ottoDir/archive/$today/*
 cp -pf $runDir/nextstrainGene.bb $runDir/nextstrainClade.bb \
     $runDir/nextstrain*.vcf.gz{,.tbi} \
     $runDir/nextstrain*.nh \
     $runDir/nextstrainSamples*.bigWig \
     $ottoDir/archive/$today
 
 echo "Updated nextstrain/ncov `date` (ncov.json date $latestDate)"