#!/bin/bash
# Working: 2001-Jan-22
# revision: 2002-May-02
# The purpose is to take chat logs generated from curfloo
#  ( http://www.jfcinnovations.com/curfloo )
#  and produce BSD fortune files from them.
#  curfloo produces chatlogs in the format:
#   [DD/MM/YYYY hh:mm:ss] name: comment
#   where DD=day, MM=month, YYYY=year, hh=hour, mm=minutes, ss=seconds,
#    name=chat name, comment=text from name in chat
#  fortune files have the format:
#   %
#   text
#           -- author
#   The .dat file that accompanies each fortune file is created
#    by executing strfile [fortune file]
#  Note that massive manual editing of the logs may be needed 
#   to avoid having fortunes that output nothing more than,
#   for example, "lol".
# ----------------------------------------------------------------------
# default action:
#  parse $logd/$logn* logs and create 1 file per name under $wdir/$saved/?/name
# possible action:
#  grep 'pattern' in $dir/$name, pipe to awk to parse into 1 filename

# first, create a working directory 'wdir'
wdir=/opt/test/curfloo/logs/test3
mkdir -p ${wdir}
cd ${wdir}
saved=atest
mkdir -p ${wdir}/${saved}
# edit these variables to suit your setup
# log directory
logd=~/.curfloo/logs
# log filenames, note the name will have a '*' tacked on
# in the routine that looks for files, so that all files
# that _begin_ with what you put here will be processed
logn=chatlog.
# there's gotta be a better way to get these directories created
dlist="junk 0 1 2 3 4 5 6 7 8 9 a b c d e f g h i j k l m n o p q r s t \
u v w x y z A B C D E F G H I J K L M N O P Q R S T U V W X Y Z"
for d in ${dlist}
do
    mkdir -p ${saved}/${d}
done
# let the crunching begin
cnt=0
cnt=$(($cnt))
for n in ${logd}/${logn}*
do
    # pad a counter with zeros, to use in filenames later
    count=${cnt}
    if [ ${cnt} -lt 100 ]
      then
        count=0${cnt}
    fi
    if [ ${cnt} -lt 10 ]
      then
        count=0${count}
    fi
    # give some user feedback: the count and the full path and filename
    echo ${cnt}
    echo ${n}
    # process with sed to create records separated by %%%, which helps awk later
    # sed checks for lines that begin with [DD/MM/YYYY hh:mm:ss] and
    #  substitutes %%%DD/MM/YYYY hh:mm:ss then pipes it to awk,
    #  which uses %%% as a record separator because many chat comments
    #  extend across multiple lines. normally, awk uses newline for a
    #  record separator
    sed 's/^\[\([0-9]\{2\}\/[0-9]\{2\}\/[0-9]\{4\} [0-9]\{2\}:[0-9]\{2\}:[0-9]\{2\}\)\]/%%%\1/g' ${n} | \
    awk -v names=${saved} 'BEGIN {
        FS=" "
        RS="%%%"
        OFS=""
        ORS=""
        # maximum allowed line length for output
        linelen=72
        # date field, time field, name field, comment field
        #  datef needs no processing to yield datevar
        #  timef needs no processing to yield timevar
        #  namef gets : stripped from it to yield namevar,
        #  com_f is parsed to produce comment
        #
        # fields:
        datef=1
        timef=2
        namef=3
        com_f=4
        # bugs? what bugs?
        # you mean "features" or "issues", of course.
        debug=1
        }
        {
        if (NF >= com_f) {
            comment=""
            nodate=substr($0, index($0, $namef))
            # strip the : off the name in namef
            namevar=substr($0, index($0,$timef) + length($timef) + 1, match(nodate,/:/) - 1)
            namedir=substr(namevar, 1, 1)
            if (namedir < "0" || ((namedir > "Z") && (namedir < "a")) || namedir > "z") {
                namedir="junk"
            }
            namev_f=namevar
            gsub(/ /,"_",namev_f)
            gsub(/\[/,"_",namev_f)
            gsub(/\//,"_",namev_f)
           # gsub(/]/,"_",namev_f)
            gsub(/</,"_",namev_f)
           # gsub(/"/,"_",namev_f)
           # gsub(/../,"_",namev_f)
            if ((namev_f == "") || (namev_f == "..")) {
                namev_f="ERROR"
                print "namev_f=" namev_f "\n"
            }
#            namev_f="manula"
            f_name=names "/" namedir "/" namev_f
            if (debug){
                print "namef=" $namef "  com_f=" $com_f "\n"
                print "namevar=" namevar "  namedir=" namedir "  dest=" f_name "\n"
            }
            datevar=$datef
            timevar=$timef
            # everything after the name field is
            #  the variable length comment
            lastvar=substr($0, index($0, namevar) + length(namevar) + 1, length($0))
            # check it for length
            if (length(lastvar) <= (linelen + 1)) {
                comment=lastvar
            } else {
                # it is too long, parse it
                linecnt=0
                pos=1
                while (pos <= length(lastvar)) {
                    charcnt=1
                    while ((charcnt <= linelen) && (pos <= length(lastvar))) {
                        curchar=substr(lastvar,pos,1)
                        if (curchar == "\n") {
                            charcnt=0
                            linecnt++
                        }
                        comment=comment curchar
                        ++charcnt
                        ++pos
                    }
                    curchar=substr(lastvar,pos,1)
                    if ((curchar != "\n") && (pos <= length(lastvar))) {
                        startline=pos - linelen
                        testpos=pos
                        testcom=comment
                        while ((curchar != " ") && (testpos > startline)) {
                            testpos--
                            curchar=substr(lastvar,testpos,1)
                            testcom=substr(comment,1,testpos)
                            if (curchar == " ") {
                                pos=testpos
                                comment=testcom
                                comment=substr(comment,1,pos)
                                pos++
                            }
                        }
                        if (substr(lastvar,pos,1)==" ") {
                            pos++
                        }
                        comment=comment "\n"
                        ++linecnt
                    }
                }
            }
            if(debug){
                print "%\n" comment "\t\t -- " namevar "\n\t\t    " datevar " " timevar "\n"
            }
            p_out = "%\n" comment "\t\t -- " namevar "\n\t\t    " datevar " " timevar "\n"
            print p_out >> f_name
            close(f_name)
        }
    }'
    cnt=$(($cnt+1))
done