#!/bin/sh # @(#$Id: rclpurple,v 1.1 2008-09-15 08:00:17 dockes Exp $ (C) 2004 J.F.Dockes # Parts taken from Estraier: #================================================================ # Estraier: a personal full-text search system # Copyright (C) 2003-2004 Mikio Hirabayashi #================================================================ #================================================================ # Extract text and other information from gaim logs # #================================================================ # set variables LANG=C ; export LANG LC_ALL=C ; export LC_ALL progname="rclpurple" filetype="purple/pidgin log" #RECFILTCOMMONCODE ############################################################################## # !! Leave the previous line unmodified!! Code imported from the # recfiltcommon file # Utility code common to all shell filters. This could be sourced at run # time, but it's slightly more efficient to include the code in the # filters at build time (with a sed script). # Describe error in a way that can be interpreted by our caller senderror() { echo RECFILTERROR $* # Also alert on stderr just in case echo ":2:$progname::: $*" 1>&2 exit 1 } iscmd() { cmd=$1 case $cmd in */*) if test -x $cmd -a ! -d $cmd ; then return 0; else return 1; fi ;; *) oldifs=$IFS; IFS=":"; set -- $PATH; IFS=$oldifs for d in $*;do test -x $d/$cmd -a ! -d $d/$cmd && return 0;done return 1 ;; esac } checkcmds() { for cmd in $*;do if iscmd $cmd then a=1 else senderror HELPERNOTFOUND $cmd fi done } # show help message if test $# -ne 1 -o "$1" = "--help" then echo "Convert a $filetype file to HTML text for Recoll indexing." echo "Usage: $progname [infile]" exit 1 fi infile="$1" # check the input file existence (may be '-' for stdin) if test "X$infile" != X- -a ! -f "$infile" then senderror INPUTNOSUCHFILE "$infile" fi # protect access to our temp files and directories umask 77 ############################################################################## # !! Leave the following line unmodified ! #ENDRECFILTCOMMONCODE checkcmds awk awk ' # First line: parse from, to , output html header NR == 1 { if (NF != 14 && NF != 13 && NF != 9) { printf("Bad format: (NF %d) %s\n", NF, $0) exit 1 } to = $3 if (NF == 14 || NF == 13) { mon_i["Jan"] = "01" mon_i["Feb"] = "02" mon_i["Mar"] = "03" mon_i["Apr"] = "04" mon_i["May"] = "05" mon_i["Jun"] = "06" mon_i["Jul"] = "07" mon_i["Aug"] = "08" mon_i["Sep"] = "09" mon_i["Oct"] = "10" mon_i["Nov"] = "11" mon_i["Dec"] = "12" date = $8 "-" mon_i[$7] "-" $6 "T" $9 if (NF == 14) { from = $13 } if (NF == 13) { from = $12 } } if (NF == 9) { from = $8 date = $5 } #printf("from [%s] to [%s] date [%s]\n", from, to, date) print "
" print ""
if (ENVIRON["RECOLL_FILTER_FORPREVIEW"] == "yes") {
printf("%s\n", $0)
}
# Remember who the main persons are. This is so that we output
# them once while indexing the conversation body, to avoid giving
# excessive weight by repeated indexing to the term.
authors[from] = "yes"
authors[to] = "yes"
next
}
/^\([0-2][0-9]:[0-5][0-9]:[0-5][0-9]\)/ {
# Conversation element 1st line. We strip from/to (except 1st
# occurrence) and time when indexing. Time is not interesting and
# repeated from/to indexing would give excessive weight
if (ENVIRON["RECOLL_FILTER_FORPREVIEW"] == "yes") {
# Preview: output everything
print $0
} else {
# Index: output only text, except each new author once. Unfortunately,
# it is too late to add them to the "author" field.
from = $2
sub(":$", "", from);
if (authors[from] == "") {
authors[from] = "yes"
printf("%s : ", from);
}
for (idx = 3; idx <= NR; idx++) {
printf("%s ", $idx)
}
printf("\n")
}
next
}
# Conversation element continuation line: print it
{
printf("%s\n", $0)
}
END {
printf("\n")
}
' < "$infile"
# exit normally
exit 0