This commit is contained in:
Philippe PITTOLI 2024-05-23 22:58:57 +02:00
parent 09678965e0
commit dc34e5b881
3 changed files with 26 additions and 42 deletions

View File

@ -1,5 +1,4 @@
#!/bin/sh #!/bin/sh
if [ $# -ne 1 ] if [ $# -ne 1 ]
then then
echo "usage: $0 result-directory" echo "usage: $0 result-directory"
@ -8,14 +7,7 @@ fi
d=$1 d=$1
echo "from data (.d) to truncated data (.t)" awk '{ print $1 }' < $d/ram_index.d | sort -n | uniq > it
for i in $d/*.d
do
fname=$(echo $i | sed "s/[.]d$/.t/")
awk '{ print $2, $3, $5 }' < $i > $fname
done
awk '{ print $1 }' < $d/ram_index.d > it
mkdir data mkdir data
echo "from truncated data (.t) to graphed data data/XXX.d" echo "from truncated data (.t) to graphed data data/XXX.d"

View File

@ -1,8 +1,10 @@
#!/bin/sh #!/bin/sh
extract="./bin/extract-final-data.sh" # .raw -> bad format -> .summary (great format)
summary="./bin/summary.r" raw2sum="./bin/raw-to-summary.sh"
summary_to_line="./bin/rsum2line.awk" # .summary (with too much data) -> truncated data (.t)
truncate_data="./bin/summary-to-truncated-data.sh"
# ./bin/extract-data-*.sh: .t -> data/XXX.d (paste an index + *.t)
if [ $# -ne 1 ] if [ $# -ne 1 ]
then then
@ -12,29 +14,7 @@ fi
dir="$1" dir="$1"
raw_to_summary() { $raw2sum "${dir}"
for i in $dir/*.raw
do
summary_with_bad_format=$(echo $i | sed "s/.raw$/.unconveniently_formated_summary/")
target=$(echo $i | sed "s/.raw$/.summary/")
if [ -f $summary_with_bad_format ]; then
echo -n "\r$summary_with_bad_format already exists: skipping "
else
Rscript $summary $i > $summary_with_bad_format
fi
if [ -f $target ]; then
echo -n "\r$target already exists: skipping "
else
$summary_to_line $summary_with_bad_format > $target
fi
done
echo ""
# Beyond a certain number of entries, retrieving data from partitions and tags isn't tested anymore.
# This leads to create "fake entries" with a duration of 0, resulting to causing some problems with
# statistical analysis. So, we need to replace "NaN" by "0" in summaries.
sed -i "s/NaN/0/g" $dir/*.summary
}
# List raw files with the number of iterations as a prefix so they can then be sorted. # List raw files with the number of iterations as a prefix so they can then be sorted.
sort_summary_files() { sort_summary_files() {
@ -46,6 +26,8 @@ f() {
} }
fill() { fill() {
# Remove previous computations.
rm ${dir}/*.d
while read LINE; do while read LINE; do
nb_it=$(echo $LINE | awk '{ print $1 }') nb_it=$(echo $LINE | awk '{ print $1 }')
target=$(echo $LINE | awk '{ print $2 }') target=$(echo $LINE | awk '{ print $2 }')
@ -55,12 +37,6 @@ fill() {
done done
} }
raw_to_summary
sort_summary_files | fill sort_summary_files | fill
extract_final_data() { $truncate_data "${dir}"
$extract $dir
}
extract_final_data

View File

@ -0,0 +1,16 @@
#!/bin/sh
if [ $# -ne 1 ]
then
echo "usage: $0 result-directory"
exit 0
fi
dir=$1
echo "from data (.d) to truncated data (.t)"
for i in $dir/*.d
do
fname=$(echo $i | sed "s/[.]d$/.t/")
awk '{ print $2, $3, $5 }' < $i > $fname
done