#!/usr/bin/perl -w
BEGIN
{
if (!defined $ENV{'GSDLHOME'}) {
print STDERR "Environment variable GSDLHOME not set.\n";
print STDERR " Have you sourced Greenstone's 'setup.bash' file?\n";
exit 1;
}
if (!defined $ENV{'JAVA_HOME'}) {
print STDERR "Environment variable JAVA_HOME not set.\n";
print STDERR "Needed by Fedora command line scripts.\n";
exit 1;
}
$ENV{'FEDORA_HOSTNAME'} = "localhost" if (!defined $ENV{'FEDORA_HOSTNAME'});
$ENV{'FEDORA_SERVER_PORT'} = "8080" if (!defined $ENV{'FEDORA_SERVER_PORT'});
$ENV{'FEDORA_USER'} = "fedoraAdmin" if (!defined $ENV{'FEDORA_USER'});
$ENV{'FEDORA_PASS'} = "fedoraAdmin" if (!defined $ENV{'FEDORA_PASS'});
$ENV{'FEDORA_PROTOCOL'} = "http" if (!defined $ENV{'FEDORA_PROTOCOL'});
$ENV{'FEDORA_PID_NAMESPACE'} = "greenstone" if (!defined $ENV{'FEDORA_PID_NAMESPACE'});
$ENV{'FEDORA_PREFIX'} = "/fedora" if (!defined $ENV{'FEDORA_PREFIX'});
unshift (@INC, "$ENV{'GSDLHOME'}/perllib/");
}
use strict;
no strict 'refs'; # allow filehandles to be variables and vice versa
no strict 'subs'; # allow barewords (e.g. STDERR) as function arguments
use util;
use gsprintf 'gsprintf';
use printusage;
use parse2;
use cfgread;
use colcfg;
use g2futil;
use dbutil;
my $arguments =
[
{ 'name' => "verbosity",
'desc' => "Level of verbosity generated",
'type' => "string",
'deft' => '1',
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "hostname",
'desc' => "Domain hostname of Fedora server",
'type' => "string",
'deft' => $ENV{'FEDORA_HOSTNAME'},
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "port",
'desc' => "Port that the Fedora server is running on.",
'type' => "string",
'deft' => $ENV{'FEDORA_SERVER_PORT'},
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "username",
'desc' => "Fedora admin username",
'type' => "string",
'deft' => $ENV{'FEDORA_USER'},
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "password",
'desc' => "Fedora admin password",
'type' => "string",
'deft' => $ENV{'FEDORA_PASS'},
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "protocol",
'desc' => "Fedora protocol, e.g. 'http' or 'https'",
'type' => "string",
'deft' => $ENV{'FEDORA_PROTOCOL'},
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "pidnamespace",
'desc' => "Fedora prefix for PIDs",
'type' => "string",
'deft' => $ENV{'FEDORA_PID_NAMESPACE'},
'reqd' => "no",
'hiddengli' => "no" },
{ 'name' => "gli",
'desc' => "",
'type' => "flag",
'reqd' => "no",
'hiddengli' => "yes" },
{ 'name' => "xml",
'desc' => "{scripts.xml}",
'type' => "flag",
'reqd' => "no",
'hiddengli' => "yes" },
{ 'name' => "removeold",
'desc' => "{import.removeold}",
'type' => "flag",
'reqd' => "no",
'modegli' => "3" },
{ 'name' => "language",
'desc' => "{scripts.language}",
'type' => "string",
'reqd' => "no",
'modegli' => "3" },
{ 'name' => "collectdir",
'desc' => "{import.collectdir}",
'type' => "string",
'deft' => "",
'reqd' => "no",
'hiddengli' => "yes" }
];
my $prog_options
= { 'name' => "g2fbuildcol.pl",
'desc' => "Ingest Greenstone directory of FedoraMETS documents into Fedora",
'args' => $arguments };
sub main
{
my (@ARGV) = @_;
my $GSDLHOME = $ENV{'GSDLHOME'};
my $options = {};
# general options available to all plugins
my $intArgLeftinAfterParsing = parse2::parse(\@ARGV,$arguments,$options,"allow_extra_options");
# Something went wrong with parsing
if ($intArgLeftinAfterParsing ==-1)
{
&PrintUsage::print_txt_usage($prog_options, "[options] greenstone-col");
die "\n";
}
my $xml = $options->{'xml'};
my $gli = $options->{'gli'};
if ($intArgLeftinAfterParsing != 1)
{
if ($xml) {
&PrintUsage::print_xml_usage($prog_options);
print "\n";
return;
}
else {
&PrintUsage::print_txt_usage($prog_options, "[options] greenstone-col");
print "\n";
return;
}
}
my $gs_col = $ARGV[0];
my $verbosity = $options->{'verbosity'};
my $hostname = $options->{'hostname'};
my $port = $options->{'port'};
my $username = $options->{'username'};
my $password = $options->{'password'};
my $protocol = $options->{'protocol'};
my $pid_namespace = $options->{'pidnamespace'};
# The following are needed in the FedoraMETS plugout
$ENV{'FEDORA_HOSTNAME'} = $hostname;
$ENV{'FEDORA_SERVER_PORT'} = $port;
my $collectdir = $options->{'collectdir'};
if (!$collectdir) {
if($ENV{'GSDL3HOME'}) {
$collectdir = &util::filename_cat($ENV{'GSDL3HOME'},"sites","localsite","collect");
} else {
$collectdir = &util::filename_cat($ENV{'GSDLHOME'},"collect");
}
}
my $full_gs_col = &util::filename_cat($collectdir,$gs_col);
if (!-e $full_gs_col ) {
print STDERR "Unable to find Greenstone collection $full_gs_col\n";
exit 1;
}
## my $archives_dir = &util::filename_cat($full_gs_col,"archives");
my $export_dir = &util::filename_cat($full_gs_col,"export");
print "***\n";
print "* Ingesting Greenstone processed files into Fedora $pid_namespace\n";
print "***\n";
# Following falls foul of Schematron rule checking
my $fd_add_prog = "fedora-ingest";
# my $fd_add_cmd;
# $fd_add_args = "dir $export_dir O metslikefedora1 $hostname:$port $username $password \\\n";
# $fd_add_args .= " \"Automated_ingest_by_gs2fed.pl\"";
# &g2futil::run_cmd($fd_add_prog,$fd_add_args,$options);
# => Ingest individually!
# set up fedoragsearch for updating the index upon ingesting documents
my $fedoragsearch_webapp = &g2futil::gsearch_webapp_folder();
# need the username and password preset in order to run fedoraGSearch's RESTClient script
# this assumes that the fedoragsearch authentication details are the same as for fedora
if (defined $fedoragsearch_webapp) {
$ENV{'fgsUserName'} = $options->{'username'};
$ENV{'fgsPassword'} = $options->{'password'};
}
if (opendir(DIR, $export_dir)) {
closedir DIR;
## my @hash_dirs = grep { /\.dir$/ } readdir(DIR);
my @hash_dirs = &g2futil::get_all_hash_dirs($export_dir);
# for each hash dir, purge its respective PID
foreach my $hd (@hash_dirs) {
my $hash_id = &g2futil::get_hash_id($hd);
if (defined $hash_id) {
my $pid = "$pid_namespace:$gs_col-$hash_id";
my $dsinfo_status = &g2futil::run_datastore_info($pid,$options);
if ($dsinfo_status == 0) {
# first remove the doc from the gsearch index before removing it from the fedora repository
print " deleting $pid from GSearch index\n";
&g2futil::run_delete_from_index($fedoragsearch_webapp,$pid,$options) if defined $fedoragsearch_webapp;
print " $pid being updated.\n";
&g2futil::run_purge($pid,$options);
}
else {
print " $pid not present.\n";
}
}
my $docmets_filename
= &util::filename_cat($hd,"docmets.xml");
print STDERR "\n" if $gli;
print "Ingesting $docmets_filename\n";
my $status = &g2futil::run_ingest($docmets_filename,$options);
# if the document was ingested into Fedora successfully, index it with GSearch next
if($status == 0) {
if(defined $hash_id) {
my $pid = "$pid_namespace:$gs_col-$hash_id";
# now update the fedoragsearch index with the newly ingested document
&g2futil::run_update_index($fedoragsearch_webapp,$pid,$options) if defined $fedoragsearch_webapp;
}
}
print STDERR "\n" if $gli;
}
}
else {
print STDERR "Error: Unable to open directory $export_dir: $!\n";
exit 1;
}
# can possibly use inexport instead of running buildcol.pl through system()
print STDERR "**** Just for now, also run Greenstone's buildcol.pl\n";
my $gs_opts = " -verbosity $verbosity";
$gs_opts .= " -gli" if ($gli);
$gs_opts .= " -collectdir \"$collectdir\"" if ($collectdir);
$gs_opts .= " -mode infodb";
my $gs_buildcol_arguments = "$gs_opts $gs_col";
&g2futil::run_cmd("buildcol.pl", $gs_buildcol_arguments, $options);
# read in collect cfg file to work out db type
my $collectcfg = &util::filename_cat ($collectdir, $gs_col, "etc", "collectionConfig.xml");
#print STDERR "**** collectcfg file: $collectcfg\n";
unless(open(FIN, "<$collectcfg")) {
print STDERR "g2f-buildcol.pl: Unable to open $collectcfg...ERROR: $!\n";
exit 1;
}
close(FIN);
# for now we assume GS3, since that's what the following gets implemented for
my $collect_cfg = &colcfg::read_collection_cfg ($collectcfg, "gs3");
# get the database type for this collection from its configuration file (may be undefined)
my $infodbtype = $collect_cfg->{'infodbtype'} || &dbutil::get_default_infodb_type();
# open .gdbm database file in building/text/$colname.gdb, using dbutil
my $colname = $gs_col;
$colname =~ s/(:?\\|\/)(.*)$/$1/; # remove any collect group from collection name to get tailname
my $building_txt_dir = &util::filename_cat ($collectdir, $gs_col, "building", "text");
my $building_txt_db = &dbutil::get_infodb_file_path($infodbtype, "$colname", $building_txt_dir);
# foreach key that matches http://dir1/dir2/....file.xxx
my $db_keys = {};
&dbutil::read_infodb_keys($infodbtype,$building_txt_db, $db_keys);
foreach my $key (keys %$db_keys) {
if($key =~ m@^http://@) {
# get value for the key
my $src_rec = &dbutil::read_infodb_entry($infodbtype,$building_txt_db, $key);
my $OID_hash_value = $src_rec->{'section'}->[0];
$OID_hash_value = "$pid_namespace:$gs_col-".$OID_hash_value; # convert to fedoraPID
# its fedora pid = "greenstone-http:$colname-http:||dir|file.xxx"
# except that fedorapids don't like extra colons and don't like |
my $fedora_identifier = "$pid_namespace-http:$gs_col-$key";
# CAN'T HAVE | OR : (as in "http:||one|two.html") in fedoraPID
$key =~ s@/@_@g;
$key =~ s@:@-@g;
my $fedora_pid = "$pid_namespace-http:$gs_col-$key";
# To run fedora ingest on the new file need to have sensible
# filenames that won't offend windows
my $fedora_key_file_name = "$fedora_pid";
$fedora_key_file_name =~ s@\.@-@g;
$fedora_key_file_name =~ s/\:/=/g;
$fedora_key_file_name .= ".xml";
# print STDERR "+++++ fpid: $fedora_pid, fedora-key filename: $fedora_key_file_name\n";
# write out a FedoraMets File for this key (in /tmp)
# -> it has one metadata value, which is 'dc:title' = HASHxxxxxx
# The HASHID shouldn't be the title: then will have
# duplicate titles and it will be hard to search for
# unique ones. What about making the filename the
# dc.title and the HASHID the dc.identifier
my $contents = "\n";
$contents .= "\n";
$contents .= "\n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= " $OID_hash_value\n";
# $contents .= " $fedora_identifier\n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= "\n";
# write out the file and then run fedora ingest on that file
# The file gets purged in g2f-import.pl, so don't remove it from export dir now
my $fedora_key_file_path = &util::filename_cat($export_dir, $fedora_key_file_name);
unless(open(FOUT, ">$fedora_key_file_path")) {
print STDERR "g2f-buildcol.pl: Unable to open $fedora_key_file_path...ERROR: $!\n";
exit 1;
}
print FOUT $contents;
close(FOUT);
print STDERR "\n" if $gli;
print STDERR "Ingesting $fedora_key_file_name\n";
# print STDERR "#### ".join(",", %$options)."\n";
&g2futil::run_ingest($fedora_key_file_path,$options);
print STDERR "\n" if $gli;
}
}
# If successful!!! Then need to think about:
# [CLX] nodes
# Doing this with FedoraMETSPlugin
# for the Greenstone reader interface to make the new Fedora collection available,
# need to write out buildConfig.xml with FedoraServiceProxy as a new ServiceRack element
# Kathy thinks it's better to create a buildConfig.xml than put it in collectionConfig.xml
my $indexdir = &util::filename_cat ($collectdir, $gs_col, "index");
&util::mk_dir($indexdir) unless &util::dir_exists($indexdir);
my $buildcfg = &util::filename_cat ($indexdir, "buildConfig.xml");
if(-e $buildcfg) {
print STDERR "***** $buildcfg already exists for this fedora collection.\n";
print STDERR "***** Not modifying it to insert a FedoraServiceProxy ServiceRack.\n";
}
else { # or do I just have a template buildConfig.xml that I copy over?
my $contents = "\n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= " \n";
$contents .= "\n";
#print STDERR "**** buildcfg file: $buildcfg\n";
unless(open(FOUT, ">$buildcfg")) {
print STDERR "g2f-buildcol.pl: Unable to open $buildcfg...ERROR: $!\n";
exit 1;
}
print FOUT $contents;
close(FOUT);
}
}
&main(@ARGV);