hpr-tools/Database/make_tsu_blank

333 lines
8.4 KiB
Plaintext
Raw Permalink Normal View History

#!/bin/bash -
#===============================================================================
#
# FILE: make_tsu_blank
#
# USAGE: ./make_tsu_blank [-h] [-D] start count
#
# DESCRIPTION: Make a template for generating a tag and summary update email.
#
# OPTIONS: ---
# REQUIREMENTS: ---
# BUGS: ---
# NOTES: Now obsolete but retained for reference purposes
# AUTHOR: Dave Morriss (djm), Dave.Morriss@gmail.com
# VERSION: 0.0.8
# CREATED: 2016-05-28 16:21:22
# REVISION: 2021-06-23 13:03:31
#
#===============================================================================
set -o nounset # Treat unset variables as an error
SCRIPT=${0##*/}
VERSION="0.0.8"
STDOUT="/dev/fd/2"
#
# Load library functions
#
LIB="$HOME/bin/function_lib.sh"
[ -e "$LIB" ] || { echo "$SCRIPT: Unable to source functions"; exit 1; }
# shellcheck source=/home/cendjm/bin/function_lib.sh
source "$LIB"
#
# Colour codes
#
define_colours
#
# We need the SSH tunnel (The script to test this and to open it, open_tunnel,
# are in ~/bin. This needs to be set up if running this stuff somewhere else)
#
if ! tunnel_is_open; then
echo "$SCRIPT: ${red}The SSH tunnel must be open to do this${reset}"
exit 1
fi
#=== FUNCTION ================================================================
# NAME: _usage
# DESCRIPTION: Report usage
# PARAMETERS: None
# RETURNS: Nothing
#===============================================================================
_usage () {
cat >$STDOUT <<-endusage
Usage: ./${SCRIPT} [-h] [-d] [-D] start count
Version: $VERSION
Generates a file of tag and summary updates for shows in the given range which
can be edited and submitted to tags@hackerpublicradio.org in order to update
the relevant shows.
Options:
-h Print this help
-D Select debug mode (works the same; more output)
Arguments:
start starting show number
count number of shows (shouldn't exceed 20)
Examples
./${SCRIPT} -h
./${SCRIPT} -D 700 10
endusage
exit
}
#=== FUNCTION ================================================================
# NAME: _DEBUG
# DESCRIPTION: Writes a message if in DEBUG mode
# PARAMETERS: List of messages
# RETURNS: Nothing
#===============================================================================
_DEBUG () {
[ "$DEBUG" == 0 ] && return
for msg in "$@"; do
printf 'D> %s\n' "$msg"
done
}
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# Option defaults
#
DEBUG=0
#
# Process options
#
while getopts :hdD opt
do
case "${opt}" in
h) _usage;;
D) DEBUG=1;;
?) echo "$SCRIPT: Invalid option; aborting"; exit 1;;
esac
done
shift $((OPTIND - 1))
#
# Directories and files
#
BASEDIR="$HOME/HPR/Database"
TSU="$BASEDIR/tsu"
PREFIX="tag_summary_updates_"
GENERATOR="$BASEDIR/query2tt2"
LIVECFG="$BASEDIR/.hpr_livedb.cfg"
TEMPLATE="$BASEDIR/query2tt2_taglist.tpl"
#
# Sanity checks
#
[ -d "$BASEDIR" ] || { echo "Unable to find directory $BASEDIR"; exit 1; }
[ -d "$TSU" ] || { echo "Unable to find directory $TSU"; exit 1; }
for item in $GENERATOR $LIVECFG $TEMPLATE; do
[ -e "$item" ] || {
echo "Unable to find component: $item"
exit 1
}
done
#
# Maximum number of shows to scan. This is advisory since we might want to
# scan 40 and only get 3 which need work!
#
LIMIT=20
#
# Check arguments
#
if [[ $# -ne 2 ]]; then
_usage
fi
#
# Validate arguments and make the END variable
#
START="$1"
COUNT="$2"
RE='^[0-9]+$'
if ! [[ $START =~ $RE ]]; then
echo "${red}Invalid starting value: $1${reset}"
exit 1
fi
if ! [[ $COUNT =~ $RE ]]; then
echo "${red}Invalid count value: $2${reset}"
exit 1
fi
#
# Deal with leading zeroes if any by forcing such numbers to base 10
#
START=$((10#$START))
COUNT=$((10#$COUNT))
((END = START + COUNT - 1))
_DEBUG "Start: $START" "Count: $COUNT" "End: $END"
#
# Argument sanity checks
#
if [[ $COUNT -gt $LIMIT ]]; then
echo "${yellow}Range: $START..$END ($COUNT)${reset}"
echo "${yellow}You are asking for a count greater than 20.${reset}"
echo "${red}Beware! This could be unmanageable!${reset}"
if ! yes_no 'Are you sure you want this? %s ' 'N'; then
echo "${red}Request ignored. Please try again.${reset}"
exit
fi
fi
#
# Generate the output file path
#
printf -v OUTFILE "%s/%s%04d-%04d.txt" "$TSU" "$PREFIX" "$START" "$END"
_DEBUG "Output: $OUTFILE"
#
# Does the output file exist? If so, can we detect any work having been done
# to it?
#
overwrite=0
if [[ -e $OUTFILE ]]; then
if [[ -s $OUTFILE ]]; then
echo "${yellow}${OUTFILE/$HOME/\~} already exists.${reset}"
if grep -E -q "^(summary|tags): ?\w+" "$OUTFILE"; then
echo -n "${yellow}** Work has been done on this file"
missing=$(grep -E -c "^(summary|tags): *$" "$OUTFILE")
if ((missing)); then
echo " (there are still tags/summaries to be added).${reset}"
else
echo ".${reset}"
fi
else
echo "${yellow}This file has not had tags or summaries added.${reset}"
fi
if ! yes_no 'Are you sure you want to replace it? %s ' 'N'; then
echo "${red}File not overwritten${reset}"
exit
else
overwrite=1
fi
else
#
# This shouldn't happen. An empty file caused by a failed query or
# because there's nothing to do should be cleared away immediately
# rather than here where the file has been left hanging around.
#
echo "${yellow}${OUTFILE/$HOME/\~} exists but is empty. Deleting it.${reset}"
rm -f "$OUTFILE"
fi
fi
_DEBUG "Overwrite: $overwrite"
#
# If we're overwriting no collision check otherwise check check check!
#
if [[ $overwrite -eq 0 ]]; then
#
# Check for collisions.
#
# Look for individual files already created, taking the FROM and TO values
# from their names. Look to see if the range START-END is in the range FROM-TO
# or the other way round. Print all collisions. Any found mean the script
# can't continue.
#
# Note that we have to force numbers to base 10 in case they have leading
# zeroes (and will therefore be treated as octal).
#
collisions=0
FILERE="${PREFIX}([0-9]{4})-([0-9]{4})\\.txt$"
for f in "$TSU"/"${PREFIX}"*; do
if [[ $f =~ $FILERE ]]; then
FROM="${BASH_REMATCH[1]}"
FROM=$((10#$FROM))
TO="${BASH_REMATCH[2]}"
TO=$((10#$TO))
if [[ (( $START -ge $FROM && $START -le $TO ) ||\
( $END -ge $FROM && $END -le $TO )) || \
(( $FROM -ge $START && $FROM -le $END ) ||\
( $TO -ge $START && $TO -le $END )) ]]; then
printf \
'%sCollision: range %04d-%04d overlaps the range %04d-%04d (in '%s')%s\n' \
"${red}" "$START" "$END" "$FROM" "$TO" "${f##*/}" "${reset}"
((collisions++))
fi
fi
done
if [[ $collisions -gt 0 ]]; then
echo "${red}Found $collisions collisions; aborting${reset}"
exit 1
fi
fi
#
# Define the SQL.
# 2021-06-20: Now we make a simpler query and rely on a script and template to
# format everything.
#
SQL=$(cat <<ENDSQL
SELECT
id, summary, tags
FROM eps
WHERE id BETWEEN $START AND $END
AND (length(summary) = 0 OR length(tags) = 0)
ORDER BY id
ENDSQL
)
_DEBUG "----" "$SQL" "----"
#
# Run MySQL with the query.
# 2021-06-20: The script below does all we want using the predefined template
#
$GENERATOR -config="$LIVECFG" -template="$TEMPLATE" "$SQL" > "$OUTFILE"
RES=$?
#
# Die if the query failed, and clear up the empty output file if found
#
[ $RES -eq 0 ] || {
echo "${red}Query failed; aborting${reset}"
if [[ -e $OUTFILE && ! -s $OUTFILE ]]; then
rm -f "$OUTFILE"
fi
exit 1
}
#
# An empty file could be "successfully" created. If so we delete it
#
if [[ -s $OUTFILE ]]; then
#
# Report the file created.
#
# 2021-06-20: The original sed call is not needed any more because the
# script we ran made the file in the form we want.
#
echo "${green}Output is in ${OUTFILE/$HOME/\~}${reset}"
else
rm -f "$OUTFILE"
echo "${yellow}No episodes need work in that range${reset}"
fi
exit
# vim: syntax=sh:ts=8:sw=4:ai:et:tw=78:fo=tcrqn21