Python RDF知识库查询
本文知识库查询使用SPARQL；
SPARQL (SPARQL Protocol and RDF Query Language)，是为RDF开发的一种查询语言和数据获取协议，它是为W3C所开发的RDF数据模型所定义，但是可以用于任何可以用RDF来表示的信息资源。
其语法格式如下:
SELECT ?Predicate ?Object where {http://baike.com/resource/华南师大?Predicate ?Object}
但是其返回的是三元组的形式,并不适合直接阅读,需要通过预处理。这里将《Python 将json格式文件转存为RDF格式文件》保存的RDF文件进行测试，代码如下：
# -*- coding: utf-8 -*-
import rdflib
import re
g1 = rdflib.Graph()
g1.parse("data.rdf", format="xml")
while 1:
print "please input the three words that you want,Subject,Predicate,Object;ues ',' to split them,ues '' if you don't know:"
try:
Subject, Predicate, Object = input()
data = ['', '', '']
if Subject != '' and Predicate == '' and Object == '': # only search Subject
q = "select ?Predicate ?Object where { + \
Subject + "> ?Predicate ?Object}"
data[0] = Subject # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[1] = txt[0][9:-1]
data[2] = txt[1][9:-1]
print data[0], data[1], data[2]
elif Subject == '' and Predicate != '' and Object == '': # only search Predicate
q = "select ?Subject ?Object where {?Subject + \
Predicate + "> ?Object}"
data[1] = Predicate # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[0] = txt[0][9:-1]
data[2] = txt[1][9:-1]
print data[0], data[1], data[2]
elif Subject == '' and Predicate == '' and Object != '': # only search Object
q = "select ?Subject ?Predicate where {?Subject ?Predicate + Object + ">}"
data[2] = Object # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[0] = txt[0][9:-1]
data[1] = txt[1][9:-1]
print data[0], data[1], data[2]
elif Subject != '' and Predicate != '' and Object == '': # search Subject and Predicate
q = "select ?Object where { + Subject + ">\
+ Predicate + "> ?Object}"
data[0] = Subject # .encode('gbk')
data[1] = Predicate # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[2] = txt[0][9:-1]
print data[0], data[1], data[2]
elif Subject != '' and Predicate == '' and Object != '': # search Subject and Object
q = "select ?Predicate where { + Subject + "> ?Predicate \
+ Object + ">}"
data[0] = Subject # .encode('gbk')
data[2] = Object # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[1] = txt[0][9:-1]
print data[0], data[1], data[2]
elif Subject == '' and Predicate != '' and Object != '': # search Predicate and Object
q = "select ?Subject where { ?Subject + Predicate + ">\
+ Object + ">}"
data[1] = Predicate # .encode('gbk')
data[2] = Object # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[0] = txt[0][9:-1]
print data[0], data[1], data[2]
elif Subject != '' and Predicate != '' and Object != '': # search all
q = "select ?Predicate ?Object where { + \
Subject + "> ?Predicate ?Object}"
data[0] = Subject # .encode('gbk')
x1 = g1.query(q)
if len(list(x1)) > 20:
leng = 20
else:
leng = len(list(x1))
for i in range(leng):
s = str(list(x1)[i]).decode('unicode_escape')
txt = re.findall(r'resource/.*?\'', s)
data[1] = txt[0][9:-1]
data[2] = txt[1][9:-1]
print data[0], data[1], data[2]
else:
break
except:
pass
考虑到当数据量很大时，把所有匹配的结果显示出来太花时间，就只显示查询到的前20条
测试结果：
输入是三元组的任意一个或者任意两个或者全部，以逗号分割，当输入格式错误时pass
测试数据及程序：点击下载