bash-refine/templates.sh

563 lines
16 KiB
Bash
Executable File

#!/bin/bash
# bash-refine v1.3.4: templates.sh, Felix Lohmeier, 2020-11-04
# https://gist.github.com/felixlohmeier/d76bd27fbc4b8ab6d683822cdf61f81d
# license: MIT License https://choosealicense.com/licenses/mit/
# TODO: example for setting metadata
# TODO: example for engine config (facets)
# ======================= TEMPLATES FOR YOUR WORKFLOW ======================== #
# The following code shows several options for import, transform and export
# use the templates to write your own scripts or execute this file for a demo
# =============================== ENVIRONMENT ================================ #
# source the main script
source "${BASH_SOURCE%/*}/bash-refine.sh" || exit 1
# make script executable from another directory
cd "${BASH_SOURCE%/*}/" || exit 1
### override default config?
#endpoint="http://localhost:3333"
#memory="1400M" # increase to available RAM
#csrf=true # set to false for OpenRefine < 3.3
#date="$(date +%Y%m%d_%H%M%S)"
#workdir="${BASH_SOURCE%/*}/output/${date}"
#logfile="${BASH_SOURCE%/*}/log/${date}.log"
#jq="${BASH_SOURCE%/*}/lib/jq" # path to executable
#refine="${BASH_SOURCE%/*}/lib/openrefine/refine" # path to executable
# check requirements, set trap, create workdir and tee to logfile
init
# ================================= STARTUP ================================== #
checkpoint "Startup"; echo
# start OpenRefine server
refine_start; echo
# ============================= MOCKUP TEST DATA ============================= #
mkdir -p input
cat << "DATA" > "input/example1.csv"
a,b,c
1,2,3
0,0,0
$,\,'
DATA
cat << "DATA" > "input/example2.tsv"
a b c
' \ $
0 0 0
3 2 1
DATA
cat << "DATA" > "input/example-operations-history.json"
[
{
"op": "core/column-addition",
"engineConfig": {
"mode": "row-based"
},
"newColumnName": "apply-from-file",
"columnInsertIndex": 2,
"baseColumnName": "b",
"expression": "grel:value.replace('2','TEST')",
"onError": "set-to-blank"
}
]
DATA
# ================================== IMPORT ================================== #
checkpoint "Import"; echo
# declare input
projects["from heredoc"]=""
projects["csv file example"]="input/example1.csv"
projects["tsv file example"]="input/example2.tsv"
projects["another csv example"]="input/example1.csv"
projects["yet another csv example"]="input/example1.csv"
# --------------------------- IMPORT FROM HEREDOC ---------------------------- #
# quoted heredoc ("DATA") will not be expanded by bash (no escaping needed)
# project id will be stored in as ${projects[csv file example]}
p="from heredoc"
f="" # optional filename, will be stored in OpenRefine project metadata
echo "import heredoc..."
if curl -fs --write-out "%{redirect_url}\n" \
--form project-file="@-$(if [[ -n $f ]]; then echo ";filename=${f}"; fi)" \
--form project-name="${p}" \
--form format="text/line-based/*sv" \
--form options='{
"encoding": "UTF-8",
"separator": " "
}' \
"${endpoint}/command/core/create-project-from-upload$(refine_csrf)" \
> "${workdir}/${p}.id" \
<< "DATA"
a b c
1 2 3
0 0 0
$ \ '
DATA
then
log "imported heredoc as ${p}"
else
error "import of ${p} failed!"
fi
refine_store "${p}" "${workdir}/${p}.id" || error "import of ${p} failed!"
echo
# ---------------------------- IMPORT FROM FILE ------------------------------ #
# project id will be stored in ${projects[tsv file example]}
p="tsv file example"
echo "import file ${projects[$p]} ..."
if curl -fs --write-out "%{redirect_url}\n" \
--form project-file="@${projects[$p]}" \
--form project-name="${p}" \
--form format="text/line-based/*sv" \
--form options='{
"encoding": "UTF-8",
"separator": "\t"
}' \
"${endpoint}/command/core/create-project-from-upload$(refine_csrf)" \
> "${workdir}/${p}.id"
then
log "imported ${projects[$p]} as ${p}"
else
error "import of ${projects[$p]} failed!"
fi
refine_store "${p}" "${workdir}/${p}.id" || error "import of ${p} failed!"
echo
# -------------------- IMPORT MULTIPLE FILES (PARALLEL) ---------------------- #
# project ids will be stored in ${projects[another csv example]} etc.
ps=( "csv file example" "another csv example" "yet another csv example" )
echo "import files" \
"$(for p in "${ps[@]}"; do printf "%s" "${projects[$p]} "; done)..."
for p in "${ps[@]}"; do
(if curl -fs --write-out "%{redirect_url}\n" \
--form project-file="@${projects[$p]}" \
--form project-name="${p}" \
--form format="line-based" \
--form options='{
"encoding": "UTF-8",
"separator": ","
}' \
"${endpoint}/command/core/create-project-from-upload$(refine_csrf)" \
> "${workdir}/${p}.id"
then
log "imported ${projects[$p]} as ${p}"
else
error "import of ${projects[$p]} failed!"
fi) &
monitor "${p}"
done
monitoring
for p in "${ps[@]}"; do
refine_store "${p}" "${workdir}/${p}.id" || error "import of ${p} failed!"
done
echo
# ================================ TRANSFORM ================================= #
checkpoint "Transform"; echo
# ------------------------ APPLY OPERATIONS FROM FILE ------------------------ #
p="csv file example"
f="input/example-operations-history.json"
echo "apply ${f} to ${p}..."
if curl -fs \
--data project="${projects[$p]}" \
--data-urlencode operations@"${f}" \
"${endpoint}/command/core/apply-operations$(refine_csrf)" > /dev/null
then
log "transformed ${p} (${projects[$p]})"
else
error "transform ${p} (${projects[$p]}) failed!"
fi
echo
# ---------------------- APPLY OPERATIONS FROM HEREDOC ----------------------- #
# quoted heredoc ("JSON") will not be expanded by bash (no escaping needed)
p="csv file example"
echo "add column apply-from-heredoc to ${p}..."
if curl -fs \
--data project="${projects[$p]}" \
--data-urlencode "operations@-" \
"${endpoint}/command/core/apply-operations$(refine_csrf)" > /dev/null \
<< "JSON"
[
{
"op": "core/column-addition",
"engineConfig": {
"mode": "row-based"
},
"newColumnName": "apply-from-heredoc",
"columnInsertIndex": 2,
"baseColumnName": "b",
"expression": "grel:value.replace('2','TEST')",
"onError": "set-to-blank"
}
]
JSON
then
log "transformed ${p} (${projects[$p]})"
else
error "transform ${p} (${projects[$p]}) failed!"
fi
echo
# ---------------- APPLY OPERATIONS FROM HEREDOC AND VARIABLES --------------- #
# unquoted heredocs with variable and multi-line expression (requires jq)
# \ must be used to quote the characters \, $, and `.
p="csv file example"
replace='TEST'
column="apply with variables"
echo "add column ${column} to ${p}..."
read -r -d '' expression << EXPRESSION
grel:value.replace(
'2',
'${replace}'
)
EXPRESSION
if curl -fs \
--data project="${projects[$p]}" \
--data-urlencode "operations@-" \
"${endpoint}/command/core/apply-operations$(refine_csrf)" > /dev/null \
<< JSON
[
{
"op": "core/column-addition",
"engineConfig": {
"mode": "row-based"
},
"newColumnName": "${column}",
"columnInsertIndex": 2,
"baseColumnName": "b",
"expression": $(echo "${expression}" | ${jq} -s -R '.'),
"onError": "set-to-blank"
}
]
JSON
then
log "transformed ${p} (${projects[$p]})"
else
error "transform ${p} (${projects[$p]}) failed!"
fi
echo
# ------ APPLY OPERATIONS FROM HEREDOC TO MULTIPLE PROJECTS (PARALLEL) ------ #
# quoted heredoc ("JSON") will not be expanded by bash (no escaping needed)
ps=( "another csv example" "yet another csv example" )
echo "add column apply-from-heredoc to" "${ps[@]}" "..."
for p in "${ps[@]}"; do
(if curl -fs \
--data project="${projects[$p]}" \
--data-urlencode "operations@-" \
"${endpoint}/command/core/apply-operations$(refine_csrf)" > /dev/null \
<< "JSON"
[
{
"op": "core/column-addition",
"engineConfig": {
"mode": "row-based"
},
"newColumnName": "apply-from-heredoc",
"columnInsertIndex": 2,
"baseColumnName": "b",
"expression": "grel:value.replace('2','TEST')",
"onError": "set-to-blank"
}
]
JSON
then
log "transformed ${p} (${projects[$p]})"
else
error "transform ${p} (${projects[$p]}) failed!"
fi) &
monitor "${p}"
done
monitoring
echo
# ------------- APPLY MULTIPLE OPERATIONS GENERATED FROM HEREDOC ------------- #
# unquoted heredoc (JSON) with variables and multiplied (requires jq)
# \ must be used to quote the characters \, $, and `.
p="csv file example"
columns=( "apply-from-file" "apply-from-heredoc" )
echo "delete columns" "${columns[@]}" "in ${p}..."
for column in "${columns[@]}"; do
cat << JSON >> "${workdir}/${p}.tmp"
[
{
"op": "core/column-removal",
"columnName": "${column}"
}
]
JSON
done
if "${jq}" -s add "${workdir}/${p}.tmp" | curl -fs \
--data project="${projects[$p]}" \
--data-urlencode operations@- \
"${endpoint}/command/core/apply-operations$(refine_csrf)" > /dev/null
then
log "transformed ${p} (${projects[$p]})"
rm "${workdir}/${p}.tmp"
else
error "transform ${p} (${projects[$p]}) failed!"
fi
echo
# ================================== EXPORT ================================== #
checkpoint "Export"; echo
# ----------------------------- EXPORT TO STDOUT ----------------------------- #
p="csv file example"
format="tsv"
echo "export ${p} in ${format} format..."
if curl -fs \
--data project="${projects[$p]}" \
--data format="tsv" \
--data engine='{"facets":[],"mode":"row-based"}' \
"${endpoint}/command/core/export-rows"
then
log "exported ${p} (${projects[$p]})"
else
error "export of ${p} (${projects[$p]}) failed!"
fi
echo
# ------------------------------ EXPORT TO FILE ------------------------------ #
p="csv file example"
format="csv"
echo "export ${p} to ${format} file..."
if curl -fs \
--data project="${projects[$p]}" \
--data format="${format}" \
--data engine='{"facets":[],"mode":"row-based"}' \
"${endpoint}/command/core/export-rows" \
> "${workdir}/${p}.${format}"
then
log "exported ${p} (${projects[$p]}) to ${workdir}/${p}.${format}"
else
error "export of ${p} (${projects[$p]}) failed!"
fi
echo
# ------------------------- TEMPLATING EXPORT TO FILE ------------------------ #
p="csv file example"
format="json"
echo "export ${p} to ${format} file using template..."
IFS= read -r -d '' template << "TEMPLATE"
{
"a": {{cells['a'].value.jsonize()}},
"b": {{cells['b'].value.jsonize()}},
"c": {{cells['c'].value.jsonize()}}
}
TEMPLATE
if echo "${template}" | head -c -2 | curl -fs \
--data project="${projects[$p]}" \
--data format="template" \
--data prefix="[
" \
--data suffix="
]" \
--data separator=",
" \
--data engine='{"facets":[],"mode":"row-based"}' \
--data-urlencode template@- \
"${endpoint}/command/core/export-rows" \
> "${workdir}/${p}.${format}"
then
log "exported ${p} (${projects[$p]}) to ${workdir}/${p}.${format}"
else
error "export of ${p} (${projects[$p]}) failed!"
fi
echo
# ------------------- EXPORT TO MULTIPLE FILES (PARALLEL) -------------------- #
ps=( "another csv example" "yet another csv example" )
format="tsv"
echo "export" "${ps[@]}" "to ${format} files..."
for p in "${ps[@]}"; do
(if curl -fs \
--data project="${projects[$p]}" \
--data format="${format}" \
--data engine='{"facets":[],"mode":"row-based"}' \
"${endpoint}/command/core/export-rows" \
> "${workdir}/${p}.${format}"
then
log "exported ${p} (${projects[$p]}) to ${workdir}/${p}.${format}"
else
error "export of ${p} (${projects[$p]}) failed!"
fi) &
monitor "${p}"
done
monitoring
echo
# ------------------------------ EXPORT PROJECT ------------------------------ #
p="csv file example"
format="openrefine.tar.gz"
echo "export ${p} to ${format} file..."
if curl -fs \
--data project="${projects[$p]}" \
"${endpoint}/command/core/export-project" \
> "${workdir}/${p}.${format}"
then
log "exported ${p} (${projects[$p]}) to ${workdir}/${p}.${format}"
else
error "export of ${p} (${projects[$p]}) failed!"
fi
echo
# ================================ UTILITIES ================================= #
checkpoint "Utilities"; echo
# ------------------------------ LIST PROJECTS ------------------------------- #
# get all project metadata and reshape json to print a list (requires jq)
echo "list projects..."
if curl -fs --get \
"${endpoint}/command/core/get-all-project-metadata" \
| "${jq}" -r '.projects | keys[] as $k | "\($k): \(.[$k] | .name)"'
then
: #log "printed list of projects"
else
error "getting list of projects failed!"
fi
echo
# ------------------------------- GET METADATA ------------------------------- #
# get project metadata and reshape json to include project id (requires jq)
p="csv file example"
echo "metadata for ${p}..."
if curl -fs --get \
--data project="${projects[$p]}" \
"${endpoint}/command/core/get-project-metadata" \
| "${jq}" "{ id: ${projects[$p]} } + ."
then
: #log "printed metadata of ${p} (${projects[$p]})"
else
error "getting metadata of ${p} (${projects[$p]}) failed!"
fi
echo
# ------------------------------ GET ROW COUNT ------------------------------- #
# get total number of rows
p="csv file example"
echo "total number of rows in ${p}..."
if curl -fs --get \
--data project="${projects[$p]}" \
--data limit=0 \
"${endpoint}/command/core/get-rows" \
| tr "," "\n" | grep total | cut -d ":" -f 2
then
: #log "printed row count of ${p} (${projects[$p]})"
else
error "getting row count of ${p} (${projects[$p]}) failed!"
fi
echo
# ------------------------------- GET COLUMNS -------------------------------- #
# get column names from project model (requires jq)
p="csv file example"
echo "column names of ${p}..."
if curl -fs --get \
--data project="${projects[$p]}" \
"${endpoint}/command/core/get-models" \
| "${jq}" -r '.columnModel | .columns[] | .name'
then
: #log "printed column names of ${p} (${projects[$p]})"
else
error "getting column names of ${p} (${projects[$p]}) failed!"
fi
echo
# -------------------------- GET OPERATIONS HISTORY -------------------------- #
# get operations history and reshape json to make it applicable (requires jq)
p="csv file example"
f="${workdir}/${p}_history.json"
echo "history of operations for ${p}..."
if curl -fs --get \
--data project="${projects[$p]}" \
"${endpoint}/command/core/get-operations" \
| "${jq}" '[ .entries[] | .operation ]' \
> "${f}"
then
log "saved ops history of ${p} (${projects[$p]}) to ${f}"
else
error "getting ops history of ${p} (${projects[$p]}) failed!"
fi
echo
# ---------------------------- GET IMPORT HISTORY ---------------------------- #
# get project metadata and filter import options history (requires jq)
p="csv file example"
echo "history of import for ${p}..."
if curl -fs --get \
--data project="${projects[$p]}" \
"${endpoint}/command/core/get-project-metadata" \
| "${jq}" ".importOptionMetadata[0]"
then
: #log "printed import history of ${p} (${projects[$p]})"
else
error "getting import history of ${p} (${projects[$p]}) failed!"
fi
echo
# ------------------------------ DELETE PROJECT ------------------------------ #
# delete a project (rarely needed for batch processing)
p="yet another csv example"
echo "delete project ${p}..."
if curl -fs \
--data project="${projects[$p]}" \
"${endpoint}/command/core/delete-project$(refine_csrf)" > /dev/null
then
log "deleted ${p} (${projects[$p]})"
else
error "deletion of ${p} (${projects[$p]}) failed!"
fi
echo
# ================================== FINISH ================================== #
checkpoint "Finish"; echo
# stop OpenRefine server
refine_stop; echo
# calculate run time based on checkpoints
checkpoint_stats; echo
# word count on all files in workdir
count_output