readline.py
Usage: Extract few lines from an extremely big file.
txt2sql.py
Usage: Store .txt to mysql
get_relations_by_keyword.py
Usage: Literaly
Freebase has marked relations between 2 entities in dataset like FB15K, which you can download from THU OpenKE.
Mid2name mapping can be downloaded from CSDN.
The RDF triples in Freebase look like:
<m.01jzhl> <people.person.education> <m.0n1k91v> .
<m.01jzhl> <people.person.profession> <m.02h664x> .
The mid2name.txt
is extremely big that Sublime Text would open it for several minutes.
- I store the
mid2name.txt
into mysql usingtxt2sql.py
. - Searching from FB15K for relations intuitionaly via
get_relations_by_keyword,py
Terminaly the demo looks like:
Keyword to search: Wuhan
<Wuhan> administrative_division <Hubei>
<Wuhan> time_zones <Time in China>
<Wuhan> containedby <China>
<Wuhan> containedby <Hubei>
<Wuhan> country <China>
Keyword to search: Steve Jobs
Steve Jobs /m/06y3r
<Steve Jobs> condition <Pancreatic cancer>
<Steve Jobs> organizations_founded <Apple Inc.>
<Steve Jobs> award_winner <John Lasseter>
<Steve Jobs> institution <Reed College>
<Steve Jobs> ethnicity <German American>
<Steve Jobs> profession <Inventor (patent)>
<Steve Jobs> organization <Apple Inc.>
<Steve Jobs> ethnicity <Caucasian race>
<Steve Jobs> profession <Designer>
<Steve Jobs> gender <Male>
<Steve Jobs> organization <Pixar>
<Steve Jobs> place_of_birth <San Francisco>
<Steve Jobs> currency <United States dollar>
<Steve Jobs> nationality <United States>
<Steve Jobs> company <Hewlett-Packard>
<Steve Jobs> currency <United States dollar>
<Steve Jobs> religion <Lutheranism>
<Steve Jobs> place_of_death <Palo Alto, California>
<Steve Jobs> list <Time 100>
<Steve Jobs> company <Apple Inc.>
<Steve Jobs> cause_of_death <Pancreatic cancer>
<Steve Jobs> religion <Atheism>
<Steve Jobs> profession <Businessperson>
<Steve Jobs> type_of_union <Marriage>
<Steve Jobs> religion <Buddhism>
<Steve Jobs> type_of_union <Domestic partnership>
<Steve Jobs> profession <Entrepreneur>
Enjoy your study on Freebase!