Perl Script to Insert DBpedia Infobox Data into a MySQL Database
I’m in the middle of a project involving Wikipedia and ResearchCyc, and I needed to get the data contained in DBpedia’s Infobox RDF Triple file into a MySQL database so I could use it in conjunction with the database created by the excellent Wikipedia Miner.
This script parses out the Wikipedia page, DBpedia Infobox Predicate and Infobox subject, and inserts them into a MySQL table. I thought I’d share it with The Internet in case someone else wanted to work with DBpedia infobox data in the same way.
One ends up with a table filled with entries like:
+----------+-----------------+---------------+------------------------------------------+ | id | predicate | subject | page | +----------+-----------------+---------------+------------------------------------------+ | 22378459 | foundationDate | 1918-10-18 | Fortaleza_Esporte_Clube | | 21460098 | areaCode | 07404 | Bösingen,_Baden-Württemberg | | 21536062 | demonym | Caronelli | Carona_(BG) | | 23913919 | draftyear | 1991 | Mike_Pritchard | | 22278400 | producer | Bow_Wow | Face_Off_(Bow_Wow_&_Omarion_album) | | 22320735 | bodyStyle | Coup | Fiat_1200 | | 22245790 | classis | Magnoliopsida | Euphorbia_atropurpurea | | 23710420 | order | Sapindales | Malleastrum_leroyi | | 25463380 | starring | Lena_Headey | The_Contractor | | 26021756 | latitudeMinutes | 27 | Walnut_Township,_Atchison_County,_Kansas | +----------+-----------------+---------------+------------------------------------------+
And a table containing the frequencies of each predicate:
+-------------------------+-----------+ | predicate | frequency | +-------------------------+-----------+ | background | 30863 | | homeTown | 49083 | | genre | 245669 | | label | 132028 | | associatedBand | 35371 | | associatedMusicalArtist | 35371 | | artist | 98762 | | producer | 96997 | | reviews | 35207 | | recordDate | 24784 | +-------------------------+-----------+
I used the DBpedia infobox data from this post: “DBpedia – Rethinking Wikipedia infobox extraction“, kindly provided by Georgi Kobilarov. The data linked to in that post is much more suitable for my requirements than previously available DBpedia infobox data – instead of multiple predicates for birth date – birthDate, dateBirth, dateOfBirth, birth…, they’ve been mapped to dbpedia:Person#birthdate. Wonderful!
To handle the occasional /uXXXX, I used Unicode::String and Unicode::Escape, by Gisle Aas and Hitoshi Amano respectively.
The script could be improved: though the Wikipedia Pages and DBpedia Infobox Predicates are fine, some of the subjects are rather … interesting. As I’m currently only interested in the 200 most frequently occuring predicates, I haven’t put more time into smoothing out some of the more interesting subject data. If anyone makes changes to the script, please let me know and I’ll update this post.
Here it is:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 | #!/usr/bin/perl ################################################################################### # # Parse out the Wikipedia page title, the DBpedia Infobox Relation and the subject from the infobox.nt file. # # Inserts data into a MySQL database tables with the structures: # # infoboxes # # +-----------+--------------+------+-----+---------+----------------+ # | Field | Type | Null | Key | Default | Extra | # +-----------+--------------+------+-----+---------+----------------+ # | id | int(11) | NO | PRI | NULL | auto_increment | # | predicate | varchar(255) | YES | | NULL | | # | subject | varchar(255) | YES | | NULL | | # | page | varchar(255) | YES | | NULL | | # | type | int(1) | YES | | NULL | | # +-----------+--------------+------+-----+---------+----------------+ # # predicateFrequency # # +-----------+--------------+------+-----+---------+-------+ # | Field | Type | Null | Key | Default | Extra | # +-----------+--------------+------+-----+---------+-------+ # | predicate | varchar(255) | NO | PRI | | | # | frequency | int(6) | YES | | NULL | | # +-----------+--------------+------+-----+---------+-------+ # # # To output to a CSV file, uncomment lines beginning with "#@#" # # DBpedia infobox.nt downloaded from: http://blog.georgikobilarov.com/2008/10/dbpedia-rethinking-wikipedia-infobox-extraction/ # # Script written by Michael Robinson - faceleg - http://pagesofinterest.net/ # # Contact: mike@pagesofinterest.net # #################################################################################### use strict; use warnings; use DBI; use DBD::mysql; #################################################################################### # # Unicode: http://search.cpan.org/~gaas/Unicode-String-2.09/String.pm # Unicode::Escape: http://search.cpan.org/~itwarrior/Unicode-Escape-0.0.2/lib/Unicode/Escape.pm # #################################################################################### use Unicode::Escape qw(escape unescape); #Database details my $username = ''; my $pass = ''; my $db = 'wikipedia'; my $querystring; #connect to database my $dbh = DBI->connect( "dbi:mysql:$db", $username, $pass, { 'PrintError' => 1, 'RaiseError' => 1 } ); my $lost = 0; # amount of entries that are "lost" due to crazy formatting. #File & directory declarations my $dir = "/path/to/": my $DBpedia = "infobox.nt"; #@#my $CSV = "cleanedRDF.csv"; ################################################################################### # # Decode URL encodings # # thanks to: http://glennf.com/writing/hexadecimal.url.encoding.html # ################################################################################### sub URLDecode { my $theURL = $_[0]; $theURL =~ tr/+/ /; $theURL =~ s/%([a-fA-F0-9]{2,2})/chr(hex($1))/eg; $theURL =~ s/<!–(.|\n)*–>//g; return $theURL; } ################################################################################### # # If a string begins with 0 or more white spaces and ", and ends with " and 0 or more white spaces, remove the white spaces and ". # ################################################################################### sub deQuote { my $deQuoted = $_[0]; if($deQuoted =~ /^\s*".*"\s*$/){ $deQuoted =~ /\s*"(.*)"\s*$/; return $1; } else{ return $deQuoted; } } ################################################################################### # # Inserts data into MySQL table # ################################################################################### sub doInsert { if($3 && $3 !~ /^\s*$/ && $3 ne ""){ #make sure the subject isn't empty... in theory - doesn't seem to work for radio station: frequency...?! #insert triple into infoboxes table $querystring = "insert into infoboxes (page,predicate,subject) values("; $querystring.= $dbh->quote(URLDecode($1)).",".$dbh->quote(deQuote(URLDecode($2))).",".$dbh->quote(deQuote(unescape(URLDecode($3)))).")"; $dbh->do($querystring); #insert or update data in predicateFrequency table my $predicate = $dbh->quote(deQuote(URLDecode($2))); $querystring = "INSERT INTO predicateFrequency(predicate,frequency) VALUES($predicate,0) ON duplicate KEY UPDATE frequency=frequency+1"; $dbh->do($querystring); } } # go through dbpedia infobox relations open (IB, $dir.$DBpedia); #@#open (CSV,'<'.$dir.$DBpedia); while(<IB>) { chomp; #offer a byte sized sacrifice to the great compiler ################################################################################### # # Wikipedia page title should not be dequoted! # # Split into elements # $1 = Wikipedia Page # $2 = Predicate # $3 = Object # ################################################################################### # # # standard triple, no extra screwing about required. # ################################################################################### if($_ =~ /<http:\/\/dbpedia\.org\/resource\/(.+)>\s<http:\/\/dbpedia.org\/ontology\/.*#(.+)>\s<http:\/\/dbpedia.org\/resource\/(.*)>.*/){ #print URLDecode($1).' '.URLDecode($2).' '.unescape(URLDecode($3)); #@#print CSV '"'.URLDecode($1).'","'.URLDecode($2).'","'.URLDecode($3).'"'; doInsert($1,$2,$3); } ################################################################################### # # standard triple with date, remove trailing declaration. # ################################################################################### elsif($_ =~ /<http:\/\/dbpedia\.org\/resource\/(.+)>\s<http:\/\/dbpedia.org\/ontology\/.*#(.+)>(.+)\^\^<http:\/\/www.w3.org\/2001\/XMLSchema#date>\s\./){ #print URLDecode($1).' '.URLDecode($2).' '.unescape(URLDecode($3)); #@#print CSV '"'.URLDecode($1).'","'.URLDecode($2).'","'.URLDecode($3).'"'; doInsert($1,$2,$3); } ################################################################################### # # "Differently" presented data # ################################################################################### elsif($_ =~ /<http:\/\/dbpedia\.org\/resource\/(.+)>\s<http:\/\/dbpedia.org\/ontology\/.*#(.+)>(.+)\s\./){ #print URLDecode($1).' '.URLDecode($2).' '.unescape(URLDecode($3)); #@#print CSV '"'.URLDecode($1).'","'.URLDecode($2).'","'.URLDecode($3).'"'; doInsert($1,$2,$3); } else{ $lost++; } #@#print CSV "\n"; } close(IB); #@#close(CSV); $dbh->disconnect; #let the world know how many we dropped print $lost; |
Like this post? Move it on along with:
Email |
delicious |
Digg |
Tweet |
Reddit |
Newsvine |
Furl |
Google |
Stumble |
HaoHao
| Trackback: |
Scroll to post title
Science Needs Your Brain!
Sam, Cathy and I have been developing an algorithm created in early 2008 by Olena and Cathy to a point where it not only finds equivalent Wikipedia articles for existing Cyc concepts, but also discovers new concepts and adds them to the knowledge base. Discovery of new concepts is achieved by using a combination of Cyc’s inference engine, automated semantic disambiguation methods developed at the University of Waikato Digital Libraries Lab, Wikipedia’s category structure and infobox information, and parsing the first sentences of Wikipedia articles.
We need volunteers to help us evaluate our data. The evaluation will only take about 25 minutes of your time, is straightforward, and quite interesting!
The aim of the evaluation is to determine whether our algorithm is smart as you are: whether it is as good as you are at saying which two concepts are related, and in what way.
Completing an evaluation will not only help us tremendously, it’ll also put you in the draw to win a US$25 Amazon.com voucher!
What is Cyc?
Cyc is an artificial intelligence project that attempts to assemble a comprehensive ontology of common-sense knowledge, with the goal of enabling computers to perform human-like reasoning.
To learn more, read this see Cyc’s Wikipedia Entry, or visit Cycorp.
UPDATE: The evaluation is finished! Results have been made available here.
Like this post? Move it on along with:
Email |
delicious |
Digg |
Tweet |
Reddit |
Newsvine |
Furl |
Google |
Stumble |
HaoHao
| Trackback: |
Scroll to post title

























Recent Comments
Great Wall
LinkedIn
Thu, 14 Jan 2010 12:54:50 +0000
Tag Cloud
LinkedIn
Thu, 14 Jan 2010 12:54:50 +0000
The Best Photographer In Shanghai
When we pick them up :( Don't know when that will be yet!
Chose another 49 today
Wed, 13 Jan 2010 14:13:21 +0000
The Best Photographer In Shanghai
When do we get to see them!!!!!
Tue, 12 Jan 2010 19:08:51 +0000
Installing Jdownloader In Ubuntu
Thank You!!!!
It works like a charm!!!!!!!!!!!!
Tue, 05 Jan 2010 13:02:35 +0000
Arrived In Shanghai
Done, look left!
Sun, 03 Jan 2010 15:37:03 +0000