Return to harvestmeta.pl CVS log

Up to [Repository] / foxridge-archiver

first import

#!/usr/local/bin/perl -w

use strict;
use XML::SAX;
use DBI;

use lib '/usr/local/mpiwg/archive';
use MPIWGStor;
use HarvestmetaHandler;

# make output unbuffered
$|=1;

#######################################################
# internal parameters
#

# program version
my $version = "0.1 (08.06.2004)";

# read command line parameters
my $args = MPIWGStor::parseargs;

# debug level
$debug = (exists $$args{'debug'}) ? $$args{'debug'} : 0;

# XML namespace (not really implemented!)
my $namespace = "";

# delete and rebuild database
my $purgeDB = (exists $$args{'purgedb'});

# database connection
my $dbh = DBI->connect("dbi:Pg:dbname=storage", "archiver", "");
if (! $dbh) {
    logger('ABORT', "unable to connect to database!");
    exit 1;
}
$dbh->{AutoCommit} = 0;
my $dbNextFileId;
my $dbNewFile;
my $dbNewMeta;
my $dbClearMeta;
my $dbFindFileName;
my $dbFindFilePath;
my $dbClearFile;
my $dbFindFileFlag;
my $dbFindFileFlagPath;
my $dbSetFileFlag;
my $dbClearAllFileFlag;

#######################################################
# check parameters that were passed to the program
#
my $basedir = $$args{'path'};
if (! $basedir) {
    logger("ABORT", "no document directory given!");
    exit 1;
}
# strip trailing slashes
$basedir =~ s/\/$//;
if (! -d $basedir) {
    logger("ABORT", "document directory \'$basedir\' doesn't exist!");
    exit 1;
}

my $metaParserHandler = HarvestmetaHandler->new;
my $metaParser = XML::SAX::ParserFactory->parser(Handler => $metaParserHandler);

#######################################################
# internal variables
#

# number of errors
my $errcnt = 0;
# number of warnings
my $warncnt = 0;

# number of files on fs
my $fcnt = 0;
# number of index files
my $idxcnt = 0;

#######################################################
# subroutines
#

#
# readAllFiles($realdir, $basedir, \%files, \%dirs)
#
# reads all files and directories below $realdir and puts the
# files in %files and directories in %dirs
# $basedir is only for recursion, it should be empty when called 
# from outside
#
sub readAllFiles {
    my ($directory, $basedir) = @_;    
    my $cnt = 0;

    if (! opendir DIR, $directory) {
	return 0;
    }
    my @dirfiles = readdir DIR;
    foreach my $fn (@dirfiles) {
	# ignore names starting with a dot
	next if ($fn =~ /^\./);
	# ignore other silly files
	next if ($junk_files{$fn});

	$cnt++;
	$fcnt++;
	my $f = "$directory/$fn";
	my $docf = ($basedir) ? "$basedir/$fn" : $fn;
	#logger('DEBUG', "fs_file: \"$f\"");
	if (-f $f) {
	    #logger("  is file");
	    if ($fn eq "index.meta") {
		harvestFile($fn, $directory);
	    }
	} elsif (-d _) {
	    #logger("  is dir");
	    # recurse into directory
	    $cnt += readAllFiles($f, $docf);
	}
    }
    return $cnt;
}

#
# cleanUnmarkedFiles($basepath)
#
# deletes all unflagged file and meta entries.
#
sub cleanUnmarkedFiles {
    my ($basepath) = @_;
    my $rv = $dbFindFileFlagPath->execute("${basepath}%");
    my $ids = $dbFindFileFlagPath->fetchall_arrayref;
    for my $i (@$ids) {
	my $id = $$i[0];
	logger('DEBUG', "cleaning file and meta of id: $id");
	$dbClearMeta->execute($id);
	$dbClearFile->execute($id);
	$dbh->commit;
    }
}

#
# harvestFile($filename, $filepath)
#
# reads the index file $filename at $filepath and puts the contents
# in the database.
#
sub harvestFile {
    my ($filename, $filepath) = @_;
    logger('DEBUG', "looking at file '$filename' at '$filepath'");
    # get file time
    my ($dev,$ino,$mode,$nlink,$uid,$gid,$rdev,$size,
	$atime,$mtime,$ctime,$blksize,$blocks)
	= stat("$filepath/$filename");
    my $filetime = stime($mtime);
    # register file in db
    my $fid = registerFile("$filepath/$filename", $filetime);
    if ($fid) {
	# file is new/modified
	# parse index file
	$metaParser->parse_uri("$filepath/$filename");
	my @data = $metaParserHandler->getData();
	logger('DEBUG', "parsed $#data+1 elements");
	registerMeta($fid, @data);
    }
    $idxcnt++;
    logger('INFO', "$idxcnt index files of $fcnt") if ($idxcnt % 10 == 0) ;
}

#
# $fileid = registerFile($filepath, $filetime)
#
# returns the file ID for the file $filepath. If necessary it
# will be added to the database. returns 0 if an update is not necessary.
#
sub registerFile {
    my ($filepath, $filetime) = @_;
    my $fileid = 0;
    # look if file is in db
    my $rv = $dbFindFileName->execute($filepath);
    my $mtime;
    ($fileid, $mtime) = $dbFindFileName->fetchrow_array;
    if ($fileid) {
	# file is in db
	# update flag
	$dbSetFileFlag->execute($fileid, 1);
	$dbh->commit;
	my $stime = s2stime($mtime);
	if ($stime ge $filetime) {
	    # if its current return 0
	    logger('DEBUG', "file: $fileid is old! time: '$stime' (vs '$filetime')");
	    return 0;
	} else {
	    logger('DEBUG', "file: $fileid is new! time: '$stime' (vs '$filetime')");
	}
    }
    if (! $fileid) {
	# get a new file id
	my $rv = $dbNextFileId->execute;
	($fileid) = $dbNextFileId->fetchrow_array;
	logger('DEBUG', "DB newfile: id=$fileid filename=$filepath mtime=$filetime");
	$dbNewFile->execute($fileid, $filepath, $filetime);
	# update flag
	$dbSetFileFlag->execute($fileid, 1);
	$dbh->commit;
    }
    return $fileid;
}

#
# registerMeta($fileid, @meta)
#
# adds the metadata information @meta for $fileid to the database.
#
sub registerMeta {
    my ($fileid, @meta) = @_;
    logger('DEBUG', "DB newmeta: fileid=$fileid ($#meta)");
    my $idx = 0;
    foreach my $keyval (@meta) {
	#logger('DEBUG', "  DB meta: $$keyval[0]=$$keyval[1]");
	$dbNewMeta->execute($fileid, $idx++, $$keyval[0], $$keyval[2], $$keyval[1]);
    }
    $dbh->commit;
    logger('INFO', "added $idx elements (file $fileid)");
}

#
# initdb()
#
# initialises the database connection.
#
sub initDB {
    my $rv;
    # clean tables
    if ($purgeDB) {
	$rv = $dbh->do("delete from files");
	$rv = $dbh->do("delete from meta");
	if ($dbh->err) {
	    logger('ABORT', "unable to clean table!");
	    exit 1;
	}
	$dbh->commit;
    }

    # clear flags
    $rv = $dbh->do("create temporary table file_flags ( fileid integer primary key, flag integer )");
    $dbh->commit;

    # prepare statements
    $dbNextFileId = $dbh->prepare("select nextval('files_id_seq')");
    $dbNewFile = $dbh->prepare("insert into files (id, filename, mtime) values (?,?,?)");
    $dbFindFileName = $dbh->prepare("select id,mtime from files where filename=?");
    $dbFindFilePath = $dbh->prepare("select id,filename,flag from files where filename like ?");
    $dbClearFile = $dbh->prepare("delete from files where id=?");
    $dbFindFileFlag = $dbh->prepare("select fileid from file_flags where flag=?");
    $dbFindFileFlagPath = $dbh->prepare("select id from files left outer join file_flags on files.id=file_flags.fileid where filename like ? and flag is null");
    $dbSetFileFlag = $dbh->prepare("insert into file_flags (fileid, flag) values (?,?)");
    $dbNewMeta = $dbh->prepare("insert into meta (fileid, idx, tags, attributes, content) values (?,?,?,?,?)");
    $dbClearMeta = $dbh->prepare("delete from meta where fileid=?");

}

#######################################################
# main
#

logger("INFO", "harvestmeta $version");
 
initDB();

# read and process all files under $basedir
my $fnum = readAllFiles($basedir, "");
# delete orphaned data (under $basedir)
cleanUnmarkedFiles($basedir);

logger("INFO", "analysed $idxcnt of $fnum files!");
logger("INFO", "$warncnt warnings");
logger("INFO", "$errcnt errors");
if ($errcnt > 0) {
    logger("ABORT", "there were errors!");
    exit 1;
} else {
    logger("DONE", "all index files read successfully!");
}