#!/bin/sh

# INSTALLATION:
# change the parameters below

# where we find the PDB structures hierarchy
export PDBDIR=/DATA/PDB/PDB     	# <<<<  <<<<  <<<<

# source GNUstep environment
source /usr/local/GNUstep/System/Makefiles/GNUstep.sh	# <<<< <<<< <<<<	

# PostgreSQL database
export PGDATABASE=mydb     		# <<<<  <<<<  <<<<
export PGHOST=localhost     		# <<<<  <<<<  <<<<
export PGPORT=5432     			# <<<<  <<<<  <<<<

# The script
export PDBCHAINSAWST=$HOME/PDBChainSaw.s	# <<<< <<<< <<<<t

# Where you expect output files to be created
BASEDIR=$HOME/PDBChainSaw     		# <<<<  <<<<  <<<<






# nothing to be changed beyound this point!



if [ ! -d "$BASEDIR" ]; then
	mkdir $BASEDIR
fi


LOGFILE=$BASEDIR/`date -I`.log
FTABLE1=$BASEDIR/chsaw_chain.sql.gz
FTABLE2=$BASEDIR/chsaw_pdb.sql.gz
FSEQUENCES=$BASEDIR/chsaw.fas

cd $BASEDIR


if [ ! -e $PDBDIR ]; then
	echo "PDB database tree not accessible. Abort."
	echo
	exit -1
fi

# temporary files
TMPPDBLIST=/tmp/pdblist.$$
TMPSTRX=/tmp/TEMPSTRX
TMPFASTA=pdbchainsaw.fas


#### 2.1. enter file list of PDB codes in table t_pdbnew
{
psql -q << ENDSQL
DROP VIEW pdb_sneu;
DROP VIEW pdb_sgeloescht;
DROP TABLE t_pdbnew;
CREATE TABLE t_pdbnew ( code char(4) not null unique );
ENDSQL

echo "BEGIN WORK;" > $TMPPDBLIST
find ${PDBDIR}/[0-9A-Z][0-9A-Z] -name "*.pdb" -type f  | sed -ne 's/.*\([1-9][0-9A-Z]\{3\}\)\.pdb$/insert into t_pdbnew (code) values ('\''\1'\'');/p;' >>  $TMPPDBLIST
echo "END WORK;" >> $TMPPDBLIST
psql -q < $TMPPDBLIST

#### 2.2. create differences to table of known pdb codes
psql -q << ENDSQL
CREATE VIEW pdb_sneu as (SELECT code from t_pdbnew except SELECT code from chsaw_pdb);
CREATE VIEW pdb_sgeloescht as (SELECT code from chsaw_pdb except SELECT code from t_pdbnew);
ENDSQL

#### 2.3. delete old structures (from table pdb_sgeloescht)
psql -qt << ENDSQL
SELECT 'total of removed structures: '||count(*) from pdb_sgeloescht;
DELETE from chsaw_chain where (pdb IN (select code from pdb_sgeloescht));
DELETE from chsaw_pdb where (code IN (select code from pdb_sgeloescht));
ENDSQL

#### 2.4. call program ChainSaw on every new structure which must be added to db
psql -qt -c "SELECT 'total of new structures: '||count(*) from pdb_sneu;"
for S in `psql -q -t -c 'select code from pdb_sneu' `; do
	echo "*******************************************************************************"
	echo "Running ChainSaw on PDB file $S.pdb"
	MolTalk $PDBCHAINSAWST -i $S
done

#### 2.5. remove temporary stuff
if [ -e $TMPPDBLIST ]; then rm $TMPPDBLIST; fi

psql -q << ENDSQL
DROP VIEW pdb_sneu;
DROP VIEW pdb_sgeloescht;
DROP TABLE t_pdbnew;
ENDSQL

} > $LOGFILE

# compress log file
gzip $LOGFILE

# dump tables chsaw_chain and chsaw_pdb
psql -c 'copy chsaw_chain (pdb,code,chainid,chainid2,residues,aminoacids,heterogens,solvent,sequence,organism,compound,eccode) to stdout;' | gzip - > $FTABLE1
psql -c 'copy chsaw_pdb (code,deposited,lastrev,experiment,resolution,header,title) to stdout;' | gzip - > $FTABLE2

chmod 644 $FTABLE1
chmod 644 $FTABLE2

# dump sequences in FASTA format (short peptides are omitted)
psql -qt -c "SELECT '>'||chainid||' ('||length(sequence)||')\n'||sequence from chsaw_chain where length(sequence)>=10;" | sed -ne 's/^ *\(.*\)$/\1/p;' > $FSEQUENCES


echo "all done."

exit 0;