723e217665f3d885134264840b7176f129af860c
jcasper
  Fri Nov 15 13:32:58 2013 -0800
findLevel now handles alternate html files, plus has a hopefully more extensible design
diff --git src/utils/qa/findLevel src/utils/qa/findLevel
index ea34a40..38f2cc5 100755
--- src/utils/qa/findLevel
+++ src/utils/qa/findLevel
@@ -12,109 +12,184 @@
 #  find out which level in the trackDb directory
 #  a track is on, and which level the corresponding
 #  .html file is on.
 #
 #  02-25-2010
 #  Brooke Rhead
 #  Changed to a bash script to simplify dealing with stderr
 #  and stdout.
 #
 ################################
 
 db=""
 tableName=""
 currDir=""
 error="no"
+releases="alpha beta public"
 
 # get arguments or print usage
 if [ $# -ne 2 ] 
 then
   echo -e "
   searches trackDb hierarchy for your table and corresponding .html file
   also returns the value of the priority and visibility entries
   and the .ra file location for each\n
     usage:  database tableName\n" >&2
   exit 1
 else
   db="$1"
   tableName="$2"
 fi
 
 # check for kent source dir at $USER root
 if ! [ -d ~/kent/src/hg/makeDb/trackDb/ ]
 then
   echo -e "\n  ERROR: `basename $0` presumes you have the kent source
   source tree in your home dir, in a dir called \"kent\"\n" >&2
   exit 1
 fi
 
 # make sure this is a valid database name
 if ! hgsql -Ne "show databases" | grep -qw $db
 then
   echo -e "\nERROR: database $db not found.  Try running on hgwdev?\n" >&2
   exit 1
 fi
 
-echo
+###########################################
+# Function definitions
 
 ###########################################
-# find the level of the associated .html file
+# htmlPrint
+# Usage: htmlPrint database html_filename
+# Prints the path to the html file for the specified database and
+# filename.  Prints an error if the file is not found.  Note that the
+# .html suffix should be omitted from the second argument.
+# Side-effects: $currDir is smashed, and the current directory may be changed
+function htmlPrint ()
+{
+  # find the level of the associated .html file(s)
   # start at the assembly level
-cd ~/kent/src/hg/makeDb/trackDb/*/$db
+  cd ~/kent/src/hg/makeDb/trackDb/*/$1
   currDir=""
 
-if [ -e $tableName.html ]
+  if [ -e $2.html ]
   then # the .html file is found at the assembly level
     currDir=`pwd -P`
   else
     cd ..
-  if [ -e $tableName.html ]
+    if [ -e $2.html ]
     then # the .html file is found at the organism level
       currDir=`pwd -P`
     else
       cd ..
-    if [ -e $tableName.html ]
+      if [ -e $2.html ]
       then # the .html file is found at the top level
         currDir=`pwd -P`
       fi
     fi
   fi
 
   if [ "$currDir" == "" ]
   then
-  echo -e " * the $tableName.html file does not exist at any level\n"
+    echo -e "the $2.html file does not exist at any level"
   else
-  echo -e " * html file: $currDir/$tableName.html\n"
+    echo -e "$currDir/$2.html"
   fi
+}
 
+# End of function definitions
 ###########################################
-# find the trackDb.ra entry (using tdbQuery)
-# the tdbQuery default is -release=alpha
-echo " * trackDb:"
-alpha=`tdbQuery "select track,release,priority,visibility,filePos from \
-     $db where track='$tableName'" 2> /dev/null` || error=yes
 
-beta=`tdbQuery -release=beta "select track,release,priority,visibility,filePos from \
-        $db where track='$tableName'" 2> /dev/null` || error=yes
 
-public=`tdbQuery -release=public "select track,release,priority,visibility,filePos from \
-       $db where track='$tableName'" 2> /dev/null` || error=yes
+# Pull the track data for each release.
+# Done with tdbQuery (default release is alpha, but we override that
+# with the release list defined at the top of this file).
+# Releases are grouped together if they have the same data.
+releaseGroups=( )
+releaseData=( )
+for release in $releases
+do
+  # note: built in assumption here that we'll only get one record back.
+  thisRecord=`tdbQuery -release=$release "select track,table,release,html, \
+          filePos from $db where track = '$tableName' or table = \
+          '$tableName'" 2> /dev/null || error=yes`
+  if [ "$thisRecord" == "" ]
+  then  # the release does not include this track - skip it
+    continue
+  fi
+
+  # check if it's identical to the track data for another release we've seen
+  recordIsNew=1
+  limit=${#releaseGroups[*]}
+  for ((i=0; i < ${limit}; i++))
+  do
+    if [ "${releaseData[$i]}" == "$thisRecord" ]
+    then   # record matches an existing record - join them
+      recordIsNew=0
+      releaseGroups[$i]="${releaseGroups[$i]},$release"
+    fi
+  done
+  if [ "$recordIsNew" -eq 1 ]
+  then  # record did not match an existing record
+    releaseData[$i]="$thisRecord"
+    releaseGroups[$i]="$release"
+  fi
+done
 
-# always print "alpha"; only print beta or public if they differ from alpha
-echo -e "$alpha\n"
-if [ "$alpha" != "$beta" ]
+if [ "${#releaseGroups[*]}" -eq 0 ]
 then
-  echo -e "$beta\n"
+  echo -e "\nERROR: No tracks exist in $db with track or table name ${tableName}.\n"
+  exit 1;
 fi
-if [ "$alpha" != "$public" ] && [ "$beta" != "$public" ] 
+
+# Now to parse the data we've pulled (for now, just pull out the html files)
+# Keeping this separate from the printing for now, as more parsing may
+# be added later for composite tracks, etc.
+limit=${#releaseGroups[*]}
+htmlFileList=( )
+for ((i=0; i < ${limit}; i++))
+do
+  htmlFile=`echo -e "${releaseData[$i]}" | awk '$1 ~ /^html/ {print $2}'`
+  # if the entry is blank, just use the track name to find the html file
+  if [ "$htmlFile" == "" ]
   then
-  echo -e "$public\n"
+    htmlFile=`echo -e "${releaseData[$i]}" | awk '$1 ~ /^track/ {print $2}'`
   fi
+  htmlFileList[$i]="$htmlFile"
+done
+
+# for each set of matching releases, print the corresponding html file location
+echo
+echo " * html files:"
+limit=${#releaseGroups[*]}
+for ((i=0; i < ${limit}; i++))
+do
+  # if there's only one entry, then all releases match.  Skip printing their names
+  if [ "$limit" -ne 1 ]
+  then
+    echo -n "${releaseGroups[$i]}: "
+  fi
+  htmlPrint $db ${htmlFileList[$i]}
+done
+
+
+###########################################
+# print the trackDb.ra entries minus the html field
+echo
+echo " * trackDb:"
+
+limit=${#releaseGroups[*]}
+for ((i=0; i < ${limit}; i++))
+do
+  echo -e "${releaseData[$i]}\n" | awk '$1 !~ /^html/ {print}'
+done
+
 
 # if tdbQuery is only giving errors, variables will be empty and nothing printed
 if [ "$error" = "yes" ]
 then
   echo -e "\nERROR: tdbQuery returned errors\n" >&2
   exit 1
 fi
 
 exit 0