X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Fhttp_getter%2Fhttp_getter.pl.in;h=44001955c7570652434640874f17bc1fd6ddefec;hb=e1d232bab1b061d9098fd666ca24bed84b38f99e;hp=85ed00a55e684786b2bb85a92788b3a83d9eccb4;hpb=ef8d99e27438595b185d7a31c7bc01859067b5b8;p=helm.git diff --git a/helm/http_getter/http_getter.pl.in b/helm/http_getter/http_getter.pl.in index 85ed00a55..44001955c 100755 --- a/helm/http_getter/http_getter.pl.in +++ b/helm/http_getter/http_getter.pl.in @@ -23,8 +23,14 @@ # For details, see the HELM World-Wide-Web page, # http://cs.unibo.it/helm/. +#use strict; + my $VERSION = "@VERSION@"; +# various variables +my ($HELM_LIB_PATH); +my (%map); + # First of all, let's load HELM configuration use Env; my $HELM_LIB_DIR = $ENV{"HELM_LIB_DIR"}; @@ -58,6 +64,8 @@ use HTTP::Request; use LWP::UserAgent; use DB_File; use Compress::Zlib; +use CGI; +use URI::Escape; #CSC: mancano i controlli sulle condizioni di errore di molte funzioni #CSC: ==> non e' robusto @@ -89,15 +97,26 @@ while (my $c = $d->accept) { while (my $r = $c->get_request) { #CSC: mancano i controlli di sicurezza - my $inputuri = $r->url; - $inputuri =~ s/^[^?]*\?uri=(.*)/$1/; - print "\nRequest: ".$r->url."\n\n"; my $http_method = $r->method; my $http_path = $r->url->path; + my $http_query = uri_unescape($r->url->query); + my $cgi = new CGI("$http_query"); + my $inputuri = $cgi->param('uri'); + print "\nRequest: ".$r->url."\n\n"; + + print "\nUnescaped query: ".$http_query."\n"; if ($http_method eq 'GET' and $http_path eq "/getciconly") { # finds the uri, url and filename my $cicuri = $inputuri; + my $answerformat = $cgi->param('format'); + $answerformat = "" if (not defined($answerformat)); + + if (($answerformat ne "gz") and ($answerformat ne "normal") + and ($answerformat ne "")) { + die "Wrong output format: $answerformat, must be 'normal' ". + "or 'gz'\n"; + } my $cicfilename = $cicuri; $cicfilename =~ s/cic:(.*)/$1/; @@ -122,110 +141,12 @@ while (my $c = $d->accept) { print_request("cic",$cicuri,$cicurl,$cicfilename); # Retrieves the file - my $ciccontent = download(0,"cic",$cicurl,$cicfilename); + my $ciccontent = download(0,"cic",$cicurl,$cicfilename,$answerformat); # Answering the client answer($c,$ciccontent); } - } elsif ($http_method eq 'GET' and $http_path eq "/get") { - # finds the uris, urls and filenames - my $cicuri = $inputuri, - $typesuri = $inputuri, - $annuri = $inputuri; - my $annsuffix; - if ($inputuri =~ /\.types$/) { - $cicuri =~ s/(.*)\.types$/$1/; - undef($annuri); - } elsif ($inputuri =~ /\.types\.ann$/) { - $cicuri =~ s/(.*)\.types\.ann$/$1/; - $typesuri =~ s/(.*)\.ann$/$1/; - $annsuffix = ".types.ann"; - } elsif ($inputuri =~ /\.ann$/) { - $cicuri =~ s/(.*)\.ann$/$1/; - undef($typesuri); - $annsuffix = ".ann"; - } else { - undef($typesuri); - undef($annuri); - } - - my $cicfilename = $cicuri; - $cicfilename =~ s/cic:(.*)/$1/; - $cicfilename =~ s/theory:(.*)/$1/; - $cicfilename = $helm_dir.$cicfilename; - - my $cicurl = $map{$cicuri}; - my $typesurl = $map{$typesuri} if (defined($typesuri)); - my $annurl = $map{$annuri} if (defined($annuri)); - my ($cicext, $typesext, $annext); - if ($cicurl =~ /\.xml$/) { # normal file - $cicext = ".xml"; - } elsif ($cicurl =~ /\.xml\.gz$/) { # gzipped file - $cicext = ".xml.gz"; - } else { - die "unexpected extension in url: $cicurl;". - "might be '.xml' or '.xml.gz'"; - } - if (defined($typesuri)) { # extension selection for types file - if ($typesurl =~ /\.xml$/) { # normal file - $typesext = ".types.xml"; - } elsif ($typesurl =~ /\.xml\.gz$/) { # gzipped file - $typesext = ".types.xml.gz"; - } else { - die "unexpected extension in url: $typesurl;". - "might be '.xml' or '.xml.gz'"; - } - } - if (defined($annuri)) { # extension selection for annotation file - if ($annurl =~ /\.xml$/) { # normal file - $annext = ".xml"; - } elsif ($annurl =~ /\.xml\.gz$/) { # gzipped file - $annext = ".xml.gz"; - } else { - die "unexpected extension in url: $annurl". - "might be '.xml' or '.xml.gz'"; - } - } - my $typesfilename = $cicfilename.$typesext if $typesuri; - my $annfilename = $cicfilename.$annsuffix.$annext if $annuri; - $cicfilename .= $cicext; - - if (!defined($cicurl) || - (!defined($typesurl) && $typesuri) || - (!defined($annuri) && $annuri)) - { - print "\nNOT FOUND!!!!!\n"; - $c->send_error(RC_NOT_FOUND) - } else { - print_request("cic",$cicuri,$cicurl,$cicfilename); - print_request("types",$typesuri,$typesurl,$typesfilename) - if ($typesuri); - print_request("ann",$annuri,$annurl,$annfilename) - if ($annuri); - - # Retrieves the files - - my $ciccontent = download(1,"cic",$cicurl,$cicfilename); - my $typescontent = - download(1,"types",$typesurl,$typesfilename) if ($typesuri); - my $anncontent = - download(1,"ann",$annurl,$annfilename) if ($annuri); - - # Merging the files together - - my $merged = < - -$ciccontent -$typescontent -$anncontent - -EOT - - # Answering the client - answer($c,$merged); - } - } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") { + } elsif ($http_method eq 'GET' and $http_path eq "/getdtd") { my $filename = $inputuri; $filename = $dtd_dir."/".$filename; print "DTD: $inputuri ==> ($filename)\n"; @@ -290,16 +211,52 @@ EOT $cont = "$quoted_html_link"; answer($c,$cont); } elsif ($http_method eq 'GET' and $http_path eq "/update") { + # rebuild urls_of_uris.db print "Update requested...\n"; mk_urls_of_uris(); kill(USR1,getppid()); # signal changes to parent print " done\n"; answer($c,"

Update done

"); - } elsif ($http_method eq 'GET' and $http_path eq "/version") { - print "Version requested!"; - answer($c,"

HTTP Getter Version ".$VERSION."

"); + } 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' + $uripattern = $baseuri; + $uripattern =~ s/^.*:(.*)/$1/; + if ($baseuri =~ /^cic:/) { + $uritype = "cic"; + } elsif ($baseuri =~ /^theory:/) { + $uritype = "theory"; + } else { + $uritype = "invalid"; + } + if ($uritype ne "invalid") { # uri is valid + if (($outype ne 'txt') and ($outype ne 'xml')) { # invalid out type + print "Invalid output type specified: $outype\n"; + answer($c,"

Invalid output type, may be ". + "\"txt\" or \"xml\"

"); + } else { # valid output + print "BASEURI $baseuri, FORMAT $outype\n"; + $cont = finduris($uritype,$uripattern,$outype); + answer($c,$cont); + } + } else { # invalid uri + print "Invalid uri: $baseuri, may begin with 'cic:', ". + "'theory:' or '*:'\n"; + answer($c,"

Invalid uri , may begin with ". + "\"cic:\", \"theory:\" or \"*:\"

"); + } + } elsif ($http_method eq 'GET' and $http_path eq "/help") { + print "Help requested!"; + answer($c,"

HTTP Getter Version ". + $VERSION."

"); } else { - print "\nINVALID REQUEST!!!!!\n"; + print "\n"; + print "INVALID REQUEST!!!!!\n"; + print "(PATH: ",$http_path,", "; + print "QUERY: ",$http_query,")\n"; $c->send_error(RC_FORBIDDEN) } print "\nRequest solved: ".$r->url."\n\n"; @@ -313,6 +270,120 @@ EOT #================================ +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 + + print "FINDURIS, uritype: $uritype, uripattern: $uripattern, ". + "format: $format\n\n"; + + if (($uritype eq "cic") or ($uritype eq "theory")) { + # get info only of one type: cic or theory + foreach (keys(%map)) { # select matching uris + $uri = $_; + if ($uri =~ /^$uritype:$uripattern\//) { + $localpart = $uri; + $localpart =~ s/^$uritype:$uripattern\/(.*)/$1/; + + if ($localpart =~ /^[^\/]*$/) { # no slash, an OBJECT + $basepart = $localpart; + $basepart =~ s/^([^.]*\.[^.]*)(\.types)?(\.ann)?/$1/; # remove exts .types or + # .types.ann + $flags = $objects{$basepart}; # get old flags + if ($localpart =~ /\.ann$/) { + $flags = add_flag("ann","YES",$flags); + } else { + $flags = add_flag("ann","NO",$flags); + } + if ($localpart =~ /\.types$/) { + $flags = add_flag("types","YES",$flags); + } elsif ($localpart =~ /\.types\.ann$/) { + $flags = add_flag("types","ANN",$flags); + } else { + $flags = add_flag("types","NO",$flags); + } + $objects{$basepart} = $flags; # save new flags + } else { # exists at least one slash, a DIR + ($dirname) = split (/\//, $localpart); + $dirs{$dirname} = ""; # no flags requirement for dir + } + } + } + } else { + die "getter internal error: unsupported uritype: \"$uritype\""; + } + # now generate OUTPUT: + # output will be saved in $content + if ($format eq "txt") { # TXT output + foreach $key (sort(keys %dirs)) { + $content .= "dir, " . $key . "\n"; + } + foreach $key (sort(keys %objects)) { + $content .= "object, $key, " . $objects{$key} . "\n"; + } + } elsif ($format eq "xml") { # XML output + $content .= '' . "\n"; + $content .= '' . "\n\n"; + $content .= "\n"; + foreach $key (sort(keys %dirs)) { + $content .= "\t
$key
\n"; + } + foreach $key (sort(keys %objects)) { + $content .= "\t\n"; + $flags = $objects{$key}; + $flags =~ s/^<(.*)>$/$1/; + my ($annflag,$typesflag) = split /,/,$flags; + $content .= "\t\t\n"; + $content .= "\t\t\n"; + $content .= "\t\n"; + } + $content .= "\n"; + } else { # may not enter this branch + die "Getter internal error: invalid format received by finduris sub"; + } + return $content; +} + +sub add_flag { +# manage string like: "" +# "ann_flag" may be one of "ann_YES", "ann_NO" +# "type_flag" may be one of "types_NO", "types_YES", "types_ANN" +# when adding a flag the max between the current flag and the new flag +# is taken, the orders are ann_NO < ann_YES and types_NO < types_YES < +# types_ANN + my ($flagtype,$newflag,$str) = @_; + $str = "<,>" if ($str eq ""); + ($str =~ s/^<(.*,.*)>$/$1/) or die "Internal error: ". + "wrong string format for flag adding in $str"; + my ($annflag,$typeflag) = split /,/,$str; + if ($flagtype eq "ann") { # ANN flag handling + if ($newflag eq "YES") { + $annflag = "YES"; + } elsif ($newflag eq "NO") { + $annflag = "NO" unless ($annflag eq "YES"); + } else { + die "Internal error: annflag must be \"YES\" or \"NO\""; + } + } elsif ($flagtype eq "types") { # TYPES flag handling + if ($newflag eq "ANN") { + $typeflag = "ANN"; + } elsif ($newflag eq "YES") { + $typeflag = "YES" unless ($typeflag eq "ANN"); + } elsif ($newflag eq "NO") { + $typeflag = "NO" + unless (($typeflag eq "ANN") or ($typeflag eq "YES")); + } else { + die "Internal error: typeflag must be \"YES\", \"NO\" or \"ANN\""; + } + } else { + die "Internal error: unsupported flagtype \"$flagtype\""; + } + $str = "<$annflag,$typeflag>"; +} #CSC: Too much powerful: creates even /home, /home/users/, ... #CSC: Does not raise errors if could not create dirs/files @@ -362,20 +433,24 @@ sub gunzip { # gunzip a file and return the deflated content sub gzip { # gzip the content argument and save it to filename argument my ($cont, $filename) = @_; + #my $debug=1; # debug only - my ($gz, $cont); - - $gz = gzopen($filename, "w") + print "gzopening $filename ...\n" if (defined($debug)); + my $gz = gzopen($filename, "w") or die "Cannot gzopen for writing file $filename: $gzerrno"; - $gz->gzwrite($cont) or die "error writing: $gzerrno\n" ; + print "gzwriting ...\n" if (defined($debug)); + $gz->gzwrite($cont) or die "error writing: $gzerrno, exiting!\n"; + print "gzclosing ...\n" if (defined($debug)); $gz->gzclose(); } sub download { - my ($remove_headers,$str,$url,$filename) = @_; + my ($remove_headers,$str,$url,$filename,$format) = @_; my ($gz, $buffer); + #my $debug=1; # for DEBUG only + my $resourcetype; # retrieve mode: "normal" (.xml) or "gzipped" (.xml.gz) if ($filename =~ /\.xml$/) { # set retrieve mode $resourcetype = "normal"; @@ -408,51 +483,71 @@ sub download ."must end with '.gz' or '.xml.gz'\n"; } } else { # download file from net - print "Downloading the $str file\n"; # download file - $ua = LWP::UserAgent->new; - $request = HTTP::Request->new(GET => "$url"); - $response = $ua->request($request, \&callback); - - # 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 $str file\n"; - mkdirs($filename); - open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n"; - print FD $cont; - close(FD); - - # handle cache conversion normal->gzipped or gzipped->normal as user choice - if (($cachemode eq 'normal') and ($resourcetype eq 'normal')) { - # cache the file as is - rename "$filename.tmp", $filename; - } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) { - # cache the file as is - # and update the $cont variabile with deflated content - rename "$filename.tmp", $filename; - $cont = gunzip($filename); - } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) { - # deflate cache entry - # and update $cont - open(FD, "> $basefname") or die "cannot open $basefname\n"; - $cont = gunzip($filename.".tmp"); - print FD $cont; - close(FD); - unlink "$filename.tmp"; # delete old gzipped file - } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) { - # compress cache entry - gzip($cont, $basefname.".gz"); - unlink "$filename.tmp"; # delete old uncompressed file - } else { - die "Internal error, unsopported cachemode, resourcetype couple\n"; - } - # $cont now contained uncompressed data + print "Downloading the $str file\n"; # download file + my $ua = LWP::UserAgent->new; + my $request = HTTP::Request->new(GET => "$url"); + my $response = $ua->request($request, \&callback); + + # 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)); + mkdirs($filename); + print "Opening tmp file for writing ...\n" if (defined($debug)); + open(FD, ">".$filename.".tmp") or die "Cannot open $filename.tmp\n"; + print "Writing on tmp file ...\n" if (defined($debug)); + print FD $cont; + print "Closing tmp file ...\n" if (defined($debug)); + close(FD); + # handle cache conversion normal->gzipped or gzipped->normal as user choice + 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", $filename; + } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'gzipped')) { + # cache the file as is + # and update the $cont variabile with deflated content + rename "$filename.tmp", $filename; + $cont = gunzip($filename); + } elsif (($cachemode eq 'normal') and ($resourcetype eq 'gzipped')) { + # deflate cache entry + # and update $cont + open(FD, "> $basefname") or die "cannot open $basefname\n"; + $cont = gunzip($filename.".tmp"); + print FD $cont; + close(FD); + unlink "$filename.tmp"; # delete old gzipped file + } elsif (($cachemode eq 'gzipped') and ($resourcetype eq 'normal')) { + # compress cache entry + print "gzipping ...\n" if (defined($debug)); + gzip($cont, $basefname.".gz"); + unlink "$filename.tmp"; # delete old uncompressed file + } else { + die "Internal error, unsopported cachemode, resourcetype couple\n"; + } + # $cont now contained uncompressed data } if ($remove_headers) { $cont =~ s/<\?xml [^?]*\?>//sg; $cont =~ s/]*>//sg; + } else { + $cont =~ s/DOCTYPE (.*) SYSTEM\s+"http:\/\/www.cs.unibo.it\/helm\/dtd\//DOCTYPE $1 SYSTEM "$myownurl\/getdtd?uri=/g; + } + if ($format eq "gz") { + gzip($cont,"$basefname.tmp"); + open (TMP, "< $basefname.tmp") + or die "Can't open tempfile: $filename.tmp, exiting!\n"; + $cont = ""; + while() { + $cont .= $_; + } + close TMP; + unlink ($basefname . ".tmp") or + die "cannot unlink temporary file: $basefname.tmp\n"; } + return $cont; } @@ -465,21 +560,18 @@ sub answer } sub helm_wget { -#ported from helm_wget executable, TODO: rewrite without wget. #retrieve a file from an url and write it to a temp dir -#used for retrieve resource indexe from servers +#used for retrieve resource index from servers + $cont = ""; my ($prefix, $URL) = @_; - if ($URL =~ /^file:\//) { - $URL =~ s/^file:\///; - my $command = "mkdir -p $prefix ; cp $URL $prefix"; - print "$command\n"; - system($command) == 0 - or die "\"$command\" error"; - } else { - my $command = "wget -c -P $prefix $URL"; - system($command) == 0 - or die "\"$command\" error"; - } + my $ua = LWP::UserAgent->new; + my $request = HTTP::Request->new(GET => "$URL"); + my $response = $ua->request($request, \&callback); + my ($filename) = reverse (split "/", $URL); # get filename part of the URL + open (TEMP, "> $prefix/$filename") + || die "Cannot open temporary file: $prefix/$filename\n"; + print TEMP $cont; + close TEMP; } sub update { @@ -509,7 +601,6 @@ sub mk_urls_of_uris { while ($server = pop @servers) { #cicle on servers in reverse order print "processing server: $server ...\n"; chomp $server; - $debugserver = "http://dalamar.krynn.it/helm"; # FOR DEBUG ONLY: REMOVE ME !! helm_wget($tmp_dir, $server."/".$indexname); #get index $idxfile = $tmp_dir."/".$indexname; open (INDEX, "< $idxfile") or