formatlogs

#!/bin/bash

# this script is for extracting and formatting facebook messages between two people (party A and party B) 
# the original logs are obtained by downloading a copy of your facebook data
# the script may not output logs in perfect chronological order, it was written under the assumption that the data
# dumps contain conversations in reverse chronological order split by person, but this is not _always_ true
# party A should be the person who's last name comes LATER in the alphabet:

# usage: ./formatlogs.sh FILE "Anna" "Zed" "Joe" "Apple"

if [ $# -ne 5 ]; then
    echo "usage: $0 path/to/messages.htm AFIRST ALAST BFIRST BLAST" 1>&2
    exit 1
fi


FILE=$1 #the path to the file containing the unformatted messages
AFIRST=$2 #first name of party A (the person whose facebook data was downloaded)
ALAST="$3,"  #last name of party A 
BFIRST=$4 #first name of party B
BLAST=$5 #last name of party B
PAIR="$AFIRST $ALAST $BFIRST $BLAST" #the actual string used

echo "negotiating relineations..."
sed 's/<span class="user">/\n/g' $FILE > temp1


echo "surveying message threading... triplicating results..."
sed -n "/div class=\"thread\">$PAIR/=" temp1 > linenumbers

x=0

echo "preparing control flow..."
for line in `cat linenumbers`; do
    echo "enacting partitioning scheme..."

    tail -n +${line} temp1  > block${x}

    l=`grep -n "<div class=\"thread\">" block${x} | cut -f1 -d: | head -2 | tail -n +2`

    head -${l} block${x} > blocks${x}

    x=$(($x+1))

done

echo "temporal distillation in effect..."
tac blocks* > temp0

echo "personalizing results..."
egrep "^($BFIRST|$AFIRST)" temp0 > temp1

echo "removing meta..."
sed 's/<\/span><span class="meta">/\n/g' temp1 > temp2

echo "antispanning... reconciliating divides..."
sed 's/<\/span><\/div><\/div><p>/\n\n/g' temp2 > temp3

echo "stripping headers..."
sed 's/<\/p><div class="message"><div class="message_header">/\n/g' temp3 > temp4

echo "achieving escape... \`"
sed 's/&#039;/`/g' temp4 > temp5

echo 'achieving escape... "'
sed 's/&quot;/"/g' temp5 > temp6

echo "disabling safeguards..."
sed 's/<\/p><\/div><div class="thread">.*<div class="message"><div class="message_header">/\n/g' temp6 > temp7

echo "achieving escape... <"
sed 's/&lt;/</g' temp7 > "${AFIRST}-${BFIRST}-logs.txt"

echo "destroying evidence..."
rm linenumbers
rm temp*
rm block*