#!/Utils/bin/zsh
# Maarten M. Fokkinga, January 2001
#
# Produces on standard output the INF'ers belonging to the top 10K most cited
# computer scientists, according to http://citeseer.ist.psu.edu.
# Please note these two ad-hoc corrections that affects the outcome:

DELETE=( 'Bui' 'Zhang' 'Liu' 'J. Martinez' 'J. van den Berg' 'Y. Chen' 'J. Wu' 'Y. Guo' 'L. Feng' 'J. Hu' 'H. Chen' 'S. Even' 'A. Field' 'V. Jones' 'M. Le' 'D. Muller' 'A. Kleppe')
THIS_PART_ONLY=( 'Brinksma' 'van Rein' 'Argenio' )

# The DELETIONS are not counted because of voluminous writing name-sakes,
# or because of confusion with "name-sake" 'M. Le Borgne'.
# When inititals etc are problematic, we search for THIS PART of the name ONLY.

#------ setting some parameters ------------------------------

# Where to get the data:
TOP10Ksite=http://citeseer.ist.psu.edu/allcited.html
INFPHONEBOOK=http://wwwcs.cs.utwente.nl/restr_inf/telefoonnummers/html/All.html

# What groups to take (only CS groups, no Math or EL groups)
INFGROUPS='(ASNA|CADTES|DACS|DB|DIES|FMT|HMI|IS|SE)'

# Wher to find lynx and vim
#UTILSDIR=/Utils/bin  # solaris
UTILSDIR=/usr/bin     #linux

# temporary files:
TOP10K=/tmp/allcited.txt
INFLIST=/tmp/inflist.txt
PHONEBOOK2LIST=/tmp/phonebook2list.vim

#------ doing the work ---------------------------------------

#echo Building a vim script that transforms the phone book into a list...

#MMF: cat > $PHONEBOOK2LIST <<END
#MMF: :set noic
#MMF: :g/^[^,]*\$/d
#MMF: :1,\$s/^ *//
#MMF: :1,\$s/^\(.\{-}\)\(\<[A-Z]\>\.\)/\2 \1/
#MMF: :1,\$s/^\(.\.\)\( .*, [\. a-zA-Z]*\.\)\(\( \+[a-z]\+\)*\)/\1\3\2/
#MMF: :1,\$s/,.*\$//
#MMF: :1,\$s/ \+/ /g
#MMF: END
#MMF: 
#MMF: #-- Uitleg van bovenstaande vi-commando's:
#MMF: #   :g/^[^,]*\$/d   
#MMF: #     -- haalt alle niet-personen weg (personen hebben een , achter de naam)
#MMF: #   :1,\$s/^ *//     
#MMF: #     -- haalt alle begin-spatie's weg
#MMF: #   :1,\$s/^\(.\{-}\)\(\<[A-Z]\>\.\)/\2 \1/
#MMF: #     -- zet eerste voorletter (single hoofdletter gevolgd door punt) vooraan
#MMF: #   :1,\$s/^\(.\.\)\( .*, [\. a-zA-Z]*\.\)\(\( \+[a-z]\+\)*\)/\1\3\2/
#MMF: #     -- zet tussenvoegsels (na een " xxx." iteratie) vooraan na de initiaal
#MMF: #   :1,\$s/,.*\$//
#MMF: #     -- haalt alles na de komma achter de naam weg
#MMF: #   :1,\$s/ \+/ /g
#MMF: #     -- Maakt dubbele spaties enkelvoudig
#MMF: #
#MMF: # We nemen geen moeite secretaresses en dergelijke te verwijderen.
#MMF: 
#MMF: # Deleting INF persons that have voluminous writing name-sakes:
#MMF: for name in $DELETE ; do
#MMF: ed $PHONEBOOK2LIST >/dev/null <<END
#MMF: \$a
#MMF: :g/$name/d
#MMF: .
#MMF: w
#MMF: q
#MMF: END
#MMF: ;
#MMF: done
#MMF: 
#MMF: # Some people have problematic (wrong!) initials or middle names:
#MMF: for name in $THIS_PART_ONLY ; do
#MMF: ed $PHONEBOOK2LIST >/dev/null <<END
#MMF: \$a
#MMF: :1,\$s/.*$name/$name/
#MMF: .
#MMF: w
#MMF: q
#MMF: END
#MMF: ;
#MMF: done
#MMF: 
#MMF: # Adding a last line to vim script for writing the file and quiting:
#MMF: ed $PHONEBOOK2LIST >/dev/null <<END
#MMF: \$a
#MMF: :wq
#MMF: .
#MMF: w
#MMF: q
#MMF: END
#MMF: # This completes the formation of the vim script that edits the phone book.

echo Downloading top 10K most cited computer scientists...
$UTILSDIR/lynx -dump $TOP10Ksite > $TOP10K
head -10 $TOP10K
echo ...
echo

echo Downloading INF phone book...
#MMF: $UTILSDIR/lynx -dump $INFPHONEBOOK > $INFLIST.1
cp $INFLIST.1 $INFLIST.ori

echo Transforming INF phone book into INF person list...
#MMF: /usr/bin/grep -E -w "$INFGROUPS" $INFLIST.1 > $INFLIST
$UTILSDIR/vim -s $PHONEBOOK2LIST $INFLIST > /dev/null 2>&1

echo Searching top10K most cited list for INF persons...
echo
echo Not searched because of voluminous writing name-sakes:
for name in $DELETE ; do
   echo -n $name,\ 
done
echo
echo

echo "<RANK> <PERSON>: <CITATIONS>"
/usr/bin/grep -w -f $INFLIST $TOP10K 

#MMF: rm -f $TOP10K
#rm -f $INFLIST $TOP10K $PHONEBOOK2LIST