#!/bin/tcsh -efx
# Script to create a relational version of UniProt database.  Should be run on
# hgwdev.

# NOTE: the next person to build this should set a variable DB, and then
# use $DBDATE instead of substituting $DBDATE with whatever the next data is!

set DBDATE=121210
set DB=sp$DBDATE

# Set up working directory
mkdir -p /hive/data/outside/uniProt/$DBDATE/build

# Download uniProt. This will take about 12 hours
cd /hive/data/outside/uniProt/$DBDATE/build
wget ftp://ftp.expasy.org/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.dat.gz
# 473,685,422  462K/s   in 13m 48s
wget ftp://ftp.expasy.org/databases/uniprot/current_release/knowledgebase/complete/uniprot_trembl.dat.gz
# 12,088,824,987  342K/s   in 11h 28m
wget ftp://ftp.expasy.org/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot_varsplic.fasta.gz
# 6,932,064   1.01M/s   in 10s

# Turn flat file into relational tab-separated files.
time zcat *.dat.gz | spToDb stdin ../tabFiles
# real    32m21.930s

# Create the database.  
hgsql mysql <<end
create database sp$DBDATE;
end

# Load it up with table definitions from source directory
hgsql sp$DBDATE < ~/kent/src/hg/protein/spToDb/spDb.sql

# Load up the data from tab files.  This takes about an hour.
set s=`date +%s`
cd /hive/data/outside/uniProt/$DBDATE/tabFiles
foreach i (*.txt)
    hgsqlimport --local sp$DBDATE $i
end
set e=`date +%s`
expr $e - $s
# 7252

# Add varsplice info 
zcat ../build/uniprot_sprot_varsplic.fasta.gz | spDbAddVarSplice sp$DBDATE stdin .
hgLoadSqlTab sp$DBDATE -append varProtein /dev/null varProtein.txt
hgLoadSqlTab sp$DBDATE -append protein /dev/null varProtein.txt
hgLoadSqlTab sp$DBDATE -append varAcc /dev/null varAcc.txt
hgLoadSqlTab sp$DBDATE -append displayId /dev/null varDisplayId.txt
hgLoadSqlTab sp$DBDATE -append accToTaxon /dev/null varAccToTaxon.txt
hgLoadSqlTab sp$DBDATE -append geneLogic /dev/null varGeneLogic.txt
hgLoadSqlTab sp$DBDATE -append gene /dev/null varGene.txt
hgLoadSqlTab sp$DBDATE -append description /dev/null varDescription.txt

# Add table descriptions
makeTableDescriptions sp$DBDATE ~/kent/src/hg/protein/spToDb/spDbTables.as

# Zip up tab files for people who prefer them to database.
gzip *.txt


