X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Fhttp_getter%2Fhttp_getter.pl.in;h=e9d1f4a3e04609137ad68b8db5658759be797f81;hb=5d7d6bd5090f3f82279bef0b93b4b361a5b1d751;hp=40fd3118d1b1b156635587bbb49e6eaf8e4a03bf;hpb=0e5feffa1e6379b84125357ea356907a56d65c3c;p=helm.git diff --git a/helm/http_getter/http_getter.pl.in b/helm/http_getter/http_getter.pl.in index 40fd3118d..e9d1f4a3e 100755 --- a/helm/http_getter/http_getter.pl.in +++ b/helm/http_getter/http_getter.pl.in @@ -30,7 +30,7 @@ my $VERSION = "@VERSION@"; # various variables my ($HELM_LIB_PATH); my $cgi_dir = "@HELM_CGI_DIR@"; -my (%map, %rdf_map); +my (%xml_map, %rdf_map, %xslt_map); # First of all, let's load HELM configuration use Env; @@ -43,23 +43,28 @@ if (defined ($HELM_LIB_DIR)) { $HELM_LIB_PATH = $DEFAULT_HELM_LIB_DIR."/configuration.pl"; } -# : TODO temporary, move this setting to configuration file -# set the cache mode, may be "gzipped" or "normal" -my $cachemode = $ENV{'HTTP_GETTER_CACHE_MODE'} || 'gzipped'; -if (($cachemode ne 'gzipped') and ($cachemode ne 'normal')) { - die "Invalid HTTP_GETTER_CACHE_MODE environment variable, must be". - "'normal' or 'gzipped'\n"; -} -# - # next require defines: $helm_dir, $html_link, $dtd_dir, $uris_dbm, $indexname require $HELM_LIB_PATH; # TEMP: TODO put these vars in configuration file configuration.xml # -$helm_rdf_dir = "/usr/local/helm/rdf"; -$rdf_dbm = "/usr/local/helm/rdf_urls"; -$rdf_indexname = "rdf_index.txt"; +my $cachemode = $ENV{'HTTP_GETTER_CACHE_MODE'} || + 'gzipped'; +if (($cachemode ne 'gzipped') and ($cachemode ne 'normal')) { + die "Invalid HTTP_GETTER_CACHE_MODE environment variable, must be". + "'normal' or 'gzipped'\n"; +} + +$helm_rdf_dir = $ENV{'HTTP_GETTER_RDF_DIR'} || $helm_rdf_dir; +$rdf_dbm = $ENV{'HTTP_GETTER_RDF_DBM'} || $rdf_dbm; +$xslt_dbm = $ENV{'HTTP_GETTER_XSLT_DBM'} || $xslt_dbm; +$rdf_indexname = $ENV{'HTTP_GETTER_RDF_INDEXNAME'} || $rdf_indexname; +$xslt_indexname = $ENV{'HTTP_GETTER_XSLT_INDEXNAME'} || $xslt_indexname; +$servers_file = $ENV{'HTTP_GETTER_SERVERS_FILE'} || $servers_file; +$uris_dbm = $ENV{'HTTP_GETTER_URIS_DBM'} || $uris_dbm; +$dtdbaseurl = $ENV{'HTTP_GETTER_DTD_BASE_URL'} || $dtdbaseurl; +$getterport = $ENV{'HTTP_GETTER_PORT'} || $getterport; + # # Let's override the configuration file @@ -73,13 +78,15 @@ use LWP::UserAgent; use DB_File; use Compress::Zlib; use CGI; +use URI; use URI::Escape; #CSC: mancano i controlli sulle condizioni di errore di molte funzioni #CSC: ==> non e' robusto #CSC: altra roba da sistemare segnata con CSC -my $d = new HTTP::Daemon LocalPort => 8081; +my $d = new HTTP::Daemon LocalPort => $getterport + or die "Error: port $getterport not available."; my $myownurl = $d->url; # Let's patch the returned URL @@ -90,25 +97,33 @@ $myownurl =~ s/http:\/\/(.*):(.*)/$1/; ($myownurl) = gethostbyname($myownurl); $myownurl = "http://".$myownurl.":".$myownport; -tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); # open dbs -tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664); +tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # open dbs +tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664); +tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664); print "\n"; print "HTTP Getter $VERSION\n"; # print hello information print "Please contact me at: \n"; print "\n"; +print "servers_file: $servers_file\n"; print "helm_dir: $helm_dir\n"; print "helm_rdf_dir: $helm_rdf_dir\n"; print "style_dir: $style_dir\n"; print "dtd_dir: $dtd_dir\n"; -print "urls_of_uris.db: $uris_dbm.db\n"; +print "urls_of_uris db: $uris_dbm\n"; +print "rdf db: $rdf_dbm\n"; +print "xslt db: $xslt_dbm\n"; print "cache mode: $cachemode\n"; print "indexname: $indexname\n"; print "rdf_indexname: $rdf_indexname\n"; +print "xslt_indexname: $xslt_indexname\n"; +print "dtdbaseurl: $dtdbaseurl\n"; +print "getterport: $getterport\n"; print "\n"; $SIG{CHLD} = "IGNORE"; # do not accumulate defunct processes $SIG{USR1} = \&update; # sent by the child to make the parent update + while (my $c = $d->accept) { if (fork() == 0) { while (my $r = $c->get_request) { @@ -121,7 +136,7 @@ while (my $c = $d->accept) { my $inputuri = $cgi->param('uri'); print "\nRequest: ".$r->url."\n\n"; - print "\nUnescaped query: ".$http_query."\n"; + print "\nUnescaped query: ".$http_query."\n"; # "getxml" works with rdf uris if ($http_method eq 'GET' and $http_path eq "/getxml") { @@ -142,56 +157,60 @@ while (my $c = $d->accept) { } my $filename = $inputuri; + my $prefix; if (not isRdfUri($inputuri)) { # standad cic: or theory: uri $filename =~ s/^cic:(.*)/$1/; $filename =~ s/^theory:(.*)/$1/; } else { # rdf uri + # The "helm:rdf/<<...>>//cic:" prefix is kept, but quoted + # (so that a "helm:rdf/<<...>>//cic:" dir is created + $prefix = $filename; $filename =~ s/^(.*)\/\/cic:(.*)/$2/; $filename =~ s/^(.*)\/\/theory:(.*)/$2/; + $prefix =~ s/\Q$filename\E//; + $prefix =~ s/_/__/g; + $prefix =~ s/\//_/g; + $filename = $prefix.$filename; } my $url = resolve ($inputuri); # resolve uri in url - if (not defined($url)) { - die "uri \"$inputuri\" can't be resolved\n"; - } - - my $extension; # file extension - if ($url =~ /\.xml$/) { # non gzipped file - $extension = ".xml"; - } elsif ($url =~ /\.xml\.gz$/) { # gzipped file - $extension = ".xml.gz"; - } else { # error: unknown extension - die "unexpected extension in url: $url, might be '.xml'". - "or '.xml.gz'"; - } - - if (not isRdfUri ($inputuri)) { - $filename = $helm_dir.$filename.$extension; - } else { - $filename = $helm_rdf_dir.$filename.$extension; - } - - if (!defined($url)) { - print "\nNOT FOUND!!!!!\n"; + if (not defined($url)) { # uri not found in uri2url map + die "NOT FOUND: uri \"$inputuri\" can't be resolved\n"; $c->send_error(RC_NOT_FOUND) - } else { - print_request("cic",$inputuri,$url,$filename); - - # Retrieves the file - my $ciccontent = download($patch_dtd,"cic",$url,$filename,$answerformat); - - # Answering the client - if ($answerformat eq "normal") { - answer($c,$ciccontent,"text/xml",""); - } else { - answer($c,$ciccontent,"text/xml","x-gzip"); - } + } else { # uri found and mapped to url + my $extension; # file extension + if ($url =~ /\.xml$/) { # non gzipped file + $extension = ".xml"; + } elsif ($url =~ /\.xml\.gz$/) { # gzipped file + $extension = ".xml.gz"; + } else { # error: unknown extension + die "unexpected extension in url: $url, might be '.xml'". + "or '.xml.gz'"; + } + + if (not isRdfUri ($inputuri)) { # save in uri std cache dir + $filename = $helm_dir.$filename.$extension; + } else { # save in rdf metadata cache dir + $filename = $helm_rdf_dir."/".$filename.$extension; + } + + print_request($inputuri,$url,$filename); + + # Retrieves the file + my $ciccontent = download($patch_dtd,$url,$filename,$answerformat); + + if ($answerformat eq "normal") { # answer the client in text/xml + answer($c,$ciccontent,"text/xml",""); + } else { # answer the client in text/xml, gzip encoding + answer($c,$ciccontent,"text/xml","x-gzip"); + } } - # "/register" does not work with rdf uris + + # "/register" does not work with rdf uris } elsif ($http_method eq 'GET' and $http_path eq "/register") { my $inputurl = $cgi->param('url'); print "Register requested...\n"; - $map{$inputuri}=$inputurl; + $xml_map{$inputuri}=$inputurl; # Now let's clean the cache my $cicfilename = $inputuri; @@ -203,12 +222,13 @@ while (my $c = $d->accept) { unlink ($helm_dir.$cicfilename.".xml.gz"); kill(USR1,getppid()); # signal changes to parent - untie %map; - print "done\n"; + untie %xml_map; + print "done.\n"; html_nice_answer($c,"Register done"); - # "/resolve" works with rdf uri + + # "/resolve" works with rdf uri } elsif ($http_method eq 'GET' and $http_path eq "/resolve") { - #my $outputurl = $map{$inputuri}; + #my $outputurl = $xml_map{$inputuri}; my $outputurl = resolve($inputuri); $outputurl = "" if (not defined($outputurl)); $cont = "\n\n"; @@ -218,7 +238,9 @@ while (my $c = $d->accept) { $cont .= "\n"; } answer($c,$cont,"text/xml",""); + } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") { + my $filename = $inputuri; $filename = $dtd_dir."/".$filename; print "DTD: $inputuri ==> ($filename)\n"; @@ -237,41 +259,51 @@ while (my $c = $d->accept) { } else { die "Could not find DTD!"; } - } elsif ($http_method eq 'GET' and $http_path eq "/getstyleconf") { - my $filename = $inputuri; - $filename = $style_dir."/config/".$filename; - if (stat($filename)) { - open(FD, $filename) or die "Cannot open $filename\n"; - $cont = ""; - while() { - s/DOCTYPE (.*) SYSTEM\s+"/DOCTYPE $1 SYSTEM "$myownurl\/getstyleconf?uri=/g; - $cont .= $_; - } - close(FD); - answer($c,$cont,"text/plain",""); - } else { - die "Could not find Style Configuration File!"; - } + } elsif ($http_method eq 'GET' and $http_path eq "/getxslt") { - my $filename = $inputuri; - $filename = $style_dir."/".$filename; - print "XSLT: $inputuri ==> ($filename)\n"; - if (stat($filename)) { - print "Using local copy\n"; - open(FD, $filename) or die "Cannot open $filename\n"; - $cont = ""; - while() { - # Vary bad heuristic for substituion of absolute URLs - # for relative ones - s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ; - s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ; - $cont .= $_; - } - close(FD); - answer($c,$cont,"text/xml",""); - } else { - die "Could not find XSLT!"; - } + + $url = $xslt_map{$inputuri}; + print "Downloading XSLT ...\n"; + helm_wget($tmp_dir, $url); + print "XSLT downloaded!\n"; + $inputuri =~ s/.*\///; + my $filename = $tmp_dir . "/" . $inputuri; + my $cont; + die "Could not find XSLT!\n" unless stat($filename); + open(FD, $filename) or die "Can't open xslt local copy: '$filename'\n"; + while() { + s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g; + s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g; + $cont .= $_; + } + close(FD); + answer($c, $cont, "text/xml", ""); + unlink $filename or + print "Can't unlink XSLT temp file '$filename'"; + +# +# my $filename = $inputuri; +# $filename = $style_dir."/".$filename; +# print "XSLT: $inputuri ==> ($filename)\n"; +# if (stat($filename)) { +# print "Using local copy\n"; +# open(FD, $filename) or die "Cannot open $filename\n"; +# $cont = ""; +# while() { +# # Vary bad heuristic for substituion of absolute URLs +# # for relative ones +# s/xsl:import\s+href="/xsl:import href="$myownurl\/getxslt?uri=/g ; +# s/xsl:include\s+href="/xsl:include href="$myownurl\/getxslt?uri=/g ; +# $cont .= $_; +# } +# close(FD); +# answer($c,$cont,"text/xml",""); +# } else { +# die "Could not find XSLT!"; +# } +# + + # "/update" works with rdf uri } elsif ($http_method eq 'GET' and $http_path eq "/update") { # rebuild urls_of_uris db and rdf uris db print "Update requested...\n"; @@ -279,13 +311,29 @@ while (my $c = $d->accept) { kill(USR1,getppid()); # signal changes to parent print " done\n"; html_nice_answer($c,"Update done"); - # "/ls" does not work with rdf uris + + # "/getalluris" does not work with rdf uris + } elsif ($http_method eq 'GET' and $http_path eq "/getalluris") { + # send back all the keys in xml + print "BASEURI $baseuri, FORMAT $outype\n"; + $cont = getalluris(); + answer($c,$cont,"text/xml",""); + + # "/getallrdfuris" + } elsif ($http_method eq 'GET' and $http_path eq "/getallrdfuris") { + # send back all the keys in xml + my $class = $cgi->param('class'); + print "BASEURI $baseuri, FORMAT $outype, CLASS $class\n"; + $cont = getallrdfuris($class); + answer($c,$cont,"text/xml",""); + } elsif ($http_method eq 'GET' and $http_path eq "/ls") { # send back keys that begin with a given uri my ($uritype,$uripattern,$uriprefix); my $baseuri = $cgi->param('baseuri'); chop $baseuri if ($baseuri =~ /.*\/$/); # remove trailing "/" - my $outype = $cgi->param('format'); # output type, might be 'txt' or 'xml' + # output type, might be 'txt' or 'xml': + my $outype = $cgi->param('format'); $uripattern = $baseuri; $uripattern =~ s/^.*:(.*)/$1/; if ($baseuri =~ /^cic:/) { @@ -317,10 +365,23 @@ while (my $c = $d->accept) { html_nice_answer($c,"Invalid uri , must begin with ". "'cic:' or 'theory:'"); } + } elsif ($http_method eq 'GET' and $http_path eq "/help") { # help request print "Help requested!\n"; html_nice_answer($c,"HTTP Getter Version: $VERSION"); + + } elsif ($http_method eq 'GET' and $http_path eq "/getempty") { + # return an empty, but valid, xml document + my $emptyxml = < + +]> + +EOXML + answer($c, $emptyxml, "", ""); + } elsif ($http_method eq 'GET' and $http_path =~ /\.cgi$/) { # cgi handling print "CGI requested!\n"; @@ -343,6 +404,7 @@ while (my $c = $d->accept) { html_nice_answer($c,"CGI '$http_path' not found ". "in CGI dir '$cgi_dir'"); } + } else { # unsupported request print "\n"; print "INVALID REQUEST!!!!!\n"; @@ -350,8 +412,10 @@ while (my $c = $d->accept) { print "QUERY: ",$http_query,")\n"; $c->send_error(RC_FORBIDDEN) } + print "\nRequest solved: ".$r->url."\n\n"; } + $c->close; undef($c); print "\nCONNECTION CLOSED\n\n"; @@ -369,7 +433,7 @@ sub isRdfUri { # return true if the uri is an rdf uri, false otherwise # the format is "helm:rdf/://" # my ($uri) = @_; - if ($uri =~ /^helm:rdf(.*):?(.*)\/\/(.*)/) { + if ($uri =~ /^helm:rdf(.*):(.*)\/\/(.*)/) { return 1; } else { return 0; @@ -378,28 +442,77 @@ sub isRdfUri { # return true if the uri is an rdf uri, false otherwise sub resolve { # resolve an uri in a url, work both with standard cic: or theory: # uris and rdf uris + print "RESOLVE subroutine\n"; my ($uri) = @_; + print "GIVEN URI: \"$uri\"\n"; if (isRdfUri ($uri)) { # rdf uri, resolve using rdf db + print "IS A RDF URI\n"; + print "I WILL RETURN '$rdf_map{$uri}'\n"; return ($rdf_map{$uri}); } else { # standard cic: or theory: uri, resolve using std uri db - return ($map{$uri}); + print "IS NOT A RDF URI\n"; + print "I WILL RETURN '$xml_map{$uri}'\n"; + return ($xml_map{$uri}); } } +sub getalluris { # get all the keys whose prefix is cic + my $content = ""; + my ($uri); + my $debug=1; # for debug + + $content .= '' . "\n"; + $content .= "" . "\n\n"; + $content .= "\n"; + foreach $uri (sort (keys(%xml_map))) { # select matching uris + if ($uri =~ /^cic:/ && not $uri =~ /.types$/) { + print "GETALLURI: $uri\n" if defined($debug); + $content .= "\t\n"; + } + } + $content .= "\n"; + return $content; +} + +sub getallrdfuris { + my $class = $_[0]; + my $content = ""; + my ($uri); + my $debug=1; # for debug + + $content .= '' . "\n"; + $content .= "" . "\n\n"; + $content .= "\n"; + foreach $uri (sort (keys(%rdf_map))) { + if ($class eq "forward" && + $uri =~ /^helm:rdf:www.cs.unibo.it\/helm\/rdf\/forward/ + || + $class eq "backward" && + $uri =~ /^helm:rdf:www.cs.unibo.it\/helm\/rdf\/backward/) { + print "GETALLRDFURI: $uri\n" if defined($debug); + $content .= "\t\n"; + } + } + $content .= "\n"; + return $content; +} + sub finduris { # find uris for cic and theory trees generation my ($uritype,$uripattern,$format) = @_; my $content = ""; my ($uri,$localpart,$basepart,$dirname,$suffix,$flags,$key); my (@itemz,@already_pushed_dir); my (%objects,%dirs); # map uris to suffixes' flags - #my $debug=1; # for debug + my $debug=1; # for debug print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ". "format: $format\n\n" if defined($debug); if (($uritype eq "cic") or ($uritype eq "theory")) { # get info only of one type: cic or theory - foreach (keys(%map)) { # select matching uris + foreach (keys(%xml_map)) { # select matching uris $uri = $_; if ($uri =~ /^$uritype:$uripattern(\/|$|\.)/) { if ($uri =~ /^$uritype:$uripattern\//) { # directory match @@ -416,9 +529,9 @@ sub finduris { # find uris for cic and theory trees generation if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT $basepart = $localpart; - $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/; - # remove exts .types or - # .types.ann + $basepart =~ s/^([^.]*\.[^.]*)((\.body)|(\.types))?(\.ann)?/$1/; + # remove exts .types, .body, + # .types.ann or .body.ann $flags = $objects{$basepart}; # get old flags if ($localpart =~ /\.ann$/) { $flags = add_flag("ann","YES",$flags); @@ -432,6 +545,13 @@ sub finduris { # find uris for cic and theory trees generation } else { $flags = add_flag("types","NO",$flags); } + if ($localpart =~ /\.body$/) { + $flags = add_flag("body","YES",$flags); + } elsif ($localpart =~ /\.body\.ann$/) { + $flags = add_flag("body","ANN",$flags); + } else { + $flags = add_flag("body","NO",$flags); + } $objects{$basepart} = $flags; # save new flags } else { # exists at least one slash, a DIR ($dirname) = split (/\//, $localpart); @@ -463,9 +583,10 @@ sub finduris { # find uris for cic and theory trees generation $content .= "\t\n"; $flags = $objects{$key}; $flags =~ s/^<(.*)>$/$1/; - my ($annflag,$typesflag) = split /,/,$flags; + my ($annflag,$typesflag,$bodyflag) = split /,/,$flags; $content .= "\t\t\n"; $content .= "\t\t\n"; + $content .= "\t\t\n"; $content .= "\t\n"; } $content .= "\n"; @@ -476,17 +597,18 @@ sub finduris { # find uris for cic and theory trees generation } sub add_flag { -# manage string like: "" +# manage string like: "" # "ann_flag" may be one of "ann_YES", "ann_NO" # "type_flag" may be one of "types_NO", "types_YES", "types_ANN" +# "body_flag" may be one of "body_NO", "body_YES", "body_ANN" # when adding a flag the max between the current flag and the new flag -# is taken, the orders are ann_NO < ann_YES and types_NO < types_YES < -# types_ANN +# is taken, the orders are ann_NO < ann_YES, types_NO < types_YES < +# types_ANN and body_NO < body_YES < body_ANN my ($flagtype,$newflag,$str) = @_; - $str = "<,>" if ($str eq ""); - ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ". + $str = "<,,>" if ($str eq ""); + ($str =~ s/^<(.*,.*,.*)>$/$1/) or die "Internal error: ". "wrong string format for flag adding in $str"; - my ($annflag,$typeflag) = split /,/,$str; + my ($annflag,$typeflag,$bodyflag) = split /,/,$str; if ($flagtype eq "ann") { # ANN flag handling if ($newflag eq "YES") { $annflag = "YES"; @@ -506,10 +628,21 @@ sub add_flag { } else { die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\""; } + } elsif ($flagtype eq "body") { # BODY flag handling + if ($newflag eq "ANN") { + $bodyflag = "ANN"; + } elsif ($newflag eq "YES") { + $bodyflag = "YES" unless ($bodyflag eq "ANN"); + } elsif ($newflag eq "NO") { + $bodyflag = "NO" + unless (($bodyflag eq "ANN") or ($bodyflag eq "YES")); + } else { + die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\""; + } } else { die "Internal error: unsupported flagtype \"$flagtype\""; } - $str = "<$annflag,$typeflag>"; + $str = "<$annflag,$typeflag,$bodyflag>"; } #CSC: Too much powerful: creates even /home, /home/users/, ... @@ -528,16 +661,12 @@ sub mkdirs sub print_request { - my ($str,$uri,$url,$filename) = @_; - print $str."uri: $uri\n"; - print $str."url: $url\n"; - print $str."filename: $filename\n\n"; -} - -sub callback -{ - my ($data) = @_; - $cont .= $data; + my ($uri,$url,$filename) = @_; + print "\n"; + print "uri: $uri\n"; + print "url: $url\n"; + print "filename: $filename\n\n"; + print "\n"; } sub gunzip { # gunzip a file and return the deflated content @@ -572,10 +701,11 @@ sub gzip { # gzip the content argument and save it to filename argument } sub download { - my ($patch_dtd,$str,$url,$filename,$format) = @_; + my ($patch_dtd,$url,$filename,$format) = @_; my ($gz, $buffer); - #my $debug=1; # for DEBUG only +# print "DOWNLOAD subs receives url: \"$url\"\n"; +# print "DOWNLOAD subs receives filename: \"$filename\"\n"; my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz) if ($filename =~ /\.xml$/) { # set retrieve mode @@ -597,7 +727,7 @@ sub download { } if ($localfname ne "") { # we already have local copy of requested file # check both possible cache entry: gzipped or normal - print "Using local copy for the $str file\n"; + print "Using local copy.\n"; if ($localfname =~ /\.xml\.gz$/) { # deflating cached file and return it $cont = gunzip($localfname); } elsif ($localfname =~ /\.xml$/) { # just return cached file @@ -609,7 +739,7 @@ sub download { ."must end with '.gz' or '.xml.gz'\n"; } } else { # download file from net - print "Downloading the $str file\n"; # download file + print "Downloading file ...\n"; # download file my $ua = LWP::UserAgent->new; my $request = HTTP::Request->new(GET => "$url"); my $response = $ua->request($request, \&callback); @@ -617,18 +747,19 @@ sub download { # cache retrieved file to disk # TODO: inefficent, I haven't yet undestood how to deflate # in memory gzipped file, without call "gzopen" - print "Storing the $str file\n"; - print "Making dirs ...\n" if (defined($debug)); + print "Storing file ...\n"; +# print "Making dirs ...\n"; mkdirs($filename); - print "Opening tmp file for writing ...\n" if (defined($debug)); +# print "Opening tmp file for writing ...\n"; open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n"; - print "Writing on tmp file ...\n" if (defined($debug)); +# print "Writing on tmp file ...\n"; print FD $cont; - print "Closing tmp file ...\n" if (defined($debug)); +# print "Closing tmp file ...\n"; close(FD); # handle cache conversion normal->gzipped or gzipped->normal as user choice - print "cachemode:$cachemode, resourcetype:$resourcetype\n" if (defined($debug)); + print "cachemode:$cachemode, resourcetype:$resourcetype\n" + if (defined($debug)); if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) { # cache the file as is rename "$filename.tmp", $filename; @@ -656,7 +787,7 @@ sub download { # $cont now contained uncompressed data } if ($patch_dtd eq "yes") { - $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g; + $cont =~ s/DOCTYPE (.*) SYSTEM\s+"$dtdbaseurl\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g; } if ($format eq "gz") { gzip($cont,"$basefname.tmp"); @@ -688,6 +819,7 @@ sub answer { $res->push_header("Pragma" => "no-cache"); $res->push_header("Expires" => "0"); $c->send_response($res); + $c->force_last_request(); } sub html_nice_answer { @@ -697,9 +829,14 @@ sub html_nice_answer { answer($c,$content,"text/html",""); } +sub callback { + my ($data) = @_; + $cont .= $data; +} + sub helm_wget { -#retrieve a file from an url and write it to a temp dir -#used for retrieve resource index from servers +# retrieve a file from an url and write it to a temp dir +# used for retrieve resource index from servers $cont = ""; my ($prefix, $URL) = @_; my $ua = LWP::UserAgent->new; @@ -713,30 +850,35 @@ sub helm_wget { } sub update { - untie %map; - tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); +# retie dbs + untie %xml_map; + tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); untie %rdf_map; - tie(%rdf_map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); + tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664); + untie %xslt_map; + tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664); } sub update_dbs { # rebuild dbs fetching resource indexes from servers. -# Rebuild urls_of_uris db (1) and rdf db (2) +# Rebuild urls_of_uris db (1), rdf db (2) and xslt db (3) my ( $server, $idxfile, $rdf_idxfile, $uri, $url, $comp, $line, @servers, - %urls_of_uris, %rdf # local proxy of $map and $rdf_map + %urls_of_uris, %rdf, %xslt # local proxy of dbs ); - untie %map; + untie %xml_map; untie %rdf_map; - foreach my $file ($uris_dbm, $rdf_dbm) { # remove old db file - if (stat $file.".db") { # remove if exists - unlink ($file.".db") or die "can't unlink old db file: $file.db\n"; + untie %xslt_map; + foreach my $file ($uris_dbm, $rdf_dbm, $xslt_dbm) { # remove old db file + if (stat $file) { # remove if exists + unlink ($file) or die "can't unlink old db file: $file\n"; } } - tie(%urls_of_uris, 'DB_File', $uris_dbm.".db", O_RDWR|O_CREAT, 0664); - tie(%rdf, 'DB_File', $rdf_dbm.".db", O_RDWR|O_CREAT, 0664); + tie(%urls_of_uris, 'DB_File', $uris_dbm, O_RDWR|O_CREAT, 0664); + tie(%rdf, 'DB_File', $rdf_dbm, O_RDWR|O_CREAT, 0664); + tie(%xslt, 'DB_File', $xslt_dbm, O_RDWR|O_CREAT, 0664); open (SRVS, "< $servers_file") or die "cannot open servers file: $servers_file\n"; @@ -748,12 +890,15 @@ sub update_dbs { chomp $server; helm_wget($tmp_dir, $server."/".$indexname); # get index helm_wget($tmp_dir, $server."/".$rdf_indexname); # get rdf index + helm_wget($tmp_dir, $server."/".$xslt_indexname); # get xslt index $idxfile = $tmp_dir."/".$indexname; $rdf_idxfile = $tmp_dir."/".$rdf_indexname; + $xslt_idxfile = $tmp_dir . "/" . $xslt_indexname; - open (INDEX, "< $idxfile") or # (1) REBUILD URLS_OF_URIS DB + # (1) REBUILD URLS_OF_URIS DB + open (INDEX, "< $idxfile") or die "cannot open temporary index file: $idxfile\n"; - while ($line = ) { # parse index and add entry to urls_of_uris db + while (my $line = ) { # parse index and add entry to urls_of_uris db chomp $line; ($uri,$comp) = split /[ \t]+/, $line; # build url: @@ -767,9 +912,10 @@ sub update_dbs { $urls_of_uris{$uri} = $url; } - open (RDF_INDEX, "< $rdf_idxfile") or # (2) REBUILD RDF DB - die "cannot open temporary rdf index file: $rdf_idxfile\n"; - while ($line = ) { # parse index and add entry to rdf db + # (2) REBUILD RDF DB + open (RDF_INDEX, "< $rdf_idxfile") or + die "cannot open temporary RDF index file: $rdf_idxfile\n"; + while (my $line = ) { # parse index and add entry to rdf db chomp $line; ($uri,$comp) = split /[ \t]+/, $line; # comp is "gz" or nothing # build url: @@ -783,18 +929,36 @@ sub update_dbs { $rdf{$uri} = $url; } + # (3) REBUILD XSLT DB + open (XSLT_INDEX, "< $xslt_idxfile") or + die "cannot open temporary XSLT index file: $xslt_idxfile\n"; + while (my $line = ) { # parse index and add entry to xslt db + chomp $line; + my $stylesheet = $line; + my $url = $server . "/" . $stylesheet; + $xslt{$stylesheet} = $url; + } + close INDEX; # close indexes close RDF_INDEX; - die "cannot unlink temporary file: $idxfile\n" # remove temp files - if (unlink $idxfile) != 1; - die "cannot unlink temporary file: $rdf_idxfile\n" - if (unlink $rdf_idxfile) != 1; + close XSLT_INDEX; + + # remove temp files + foreach my $file ($idxfile, $rdf_idxfile, $xslt_idxfile) { + print "cannot unlink temporary file: $file\n" if (unlink $file) != 1; + } } untie(%urls_of_uris); # untie local proxies untie(%rdf); - tie(%map, 'DB_File', $uris_dbm.".db", O_RDWR, 0664); # retie global ones - tie(%rdf_map, 'DB_File', $rdf_dbm.".db", O_RDWR, 0664); + untie(%xslt); + tie(%xml_map, 'DB_File', $uris_dbm, O_RDWR, 0664); # retie global ones + tie(%rdf_map, 'DB_File', $rdf_dbm, O_RDWR, 0664); + tie(%xslt_map, 'DB_File', $xslt_dbm, O_RDWR, 0664); + } # update_dbs +# vim modline: do not remove! +# vim: set ft=perl: +