author | ymh <ymh.work@gmail.com> |
Mon, 12 Jun 2017 14:53:59 +0200 | |
changeset 537 | d2e6ee099125 |
parent 387 | 7fba86fa8604 |
permissions | -rw-r--r-- |
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
1 |
<?php |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
2 |
namespace CorpusParole\Libraries\Sparql; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
3 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
4 |
use CorpusParole\Libraries\Utils; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
5 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
6 |
class SparqlQueryAnalyser { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
7 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
8 |
const SPARQL_PREFIX_BASE_REGEXP = '(((?:prefix\s+([\p{L}-\d]+)\s*\:)|base)\s*\<((?:(?:https?|ftp)://)(?:\S+(?::\S*)?@)?(?:(?!10(?:\.\d{1,3}){3})(?!127(?:\.\d{1,3}){3})(?!169\.254(?:\.\d{1,3}){2})(?!192\.168(?:\.\d{1,3}){2})(?!172\.(?:1[6-9]|2\d|3[0-1])(?:\.\d{1,3}){2})(?:[1-9]\d?|1\d\d|2[01]\d|22[0-3])(?:\.(?:1?\d{1,2}|2[0-4]\d|25[0-5])){2}(?:\.(?:[1-9]\d?|1\d\d|2[0-4]\d|25[0-4]))|(?:(?:[a-z\x{00a1}-\x{ffff}0-9]+-?)*[a-z\x{00a1}-\x{ffff}0-9]+)(?:\.(?:[a-z\x{00a1}-\x{ffff}0-9]+-?)*[a-z\x{00a1}-\x{ffff}0-9]+)*(?:\.(?:[a-z\x{00a1}-\x{ffff}]{2,})))(?::\d{2,5})?(?:/[^\s]*)?)\>)\s*'; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
9 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
10 |
const SPARQL_SELECT_QUERY_REGEXP = '^(?:\s*(?:'.self::SPARQL_PREFIX_BASE_REGEXP.')*select)'; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
11 |
const SPARQL_ASK_QUERY_REGEXP = '^(?:\s*(?:'.self::SPARQL_PREFIX_BASE_REGEXP.')*ask)'; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
12 |
const SPARQL_GRAPH_QUERY_REGEXP = '^(?:\s*(?:'.self::SPARQL_PREFIX_BASE_REGEXP.')*(?:(?:construct)|(?:describe)))'; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
13 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
14 |
const SPARQL_LIMIT_OFFSET_QUERY_REGEXP = '(?:(?:(limit\s+(\d+))|(offset\s+(\d+)))\s*)+\s*$'; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
15 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
16 |
const UNKNOWN_QUERY = 0; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
17 |
const SELECT_QUERY = 1; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
18 |
const GRAPH_QUERY = 2; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
19 |
const ASK_QUERY = 2; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
20 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
21 |
private $query; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
22 |
private $queryType = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
23 |
private $rawPrefixes = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
24 |
private $prefixes = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
25 |
private $limit = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
26 |
private $offset = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
27 |
private $rawQuery = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
28 |
private $countVar = false; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
29 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
30 |
public function __construct($query) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
31 |
$this->query = $query; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
32 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
33 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
34 |
public function getQueryType() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
35 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
36 |
if($this->queryType === false) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
37 |
if(preg_match("%".self::SPARQL_SELECT_QUERY_REGEXP."%iu", $this->query) === 1) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
38 |
$this->queryType = self::SELECT_QUERY; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
39 |
} elseif(preg_match("%".self::SPARQL_GRAPH_QUERY_REGEXP."%iu", $this->query) === 1) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
40 |
$this->queryType = self::GRAPH_QUERY; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
41 |
} elseif(preg_match("%".self::SPARQL_ASK_QUERY_REGEXP."%iu", $this->query) === 1) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
42 |
$this->queryType = self::ASK_QUERY; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
43 |
} else { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
44 |
$this->queryType = self::UNKNOWN_QUERY; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
45 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
46 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
47 |
return $this->queryType; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
48 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
49 |
|
387 | 50 |
private function extractPrefixLimit() { |
51 |
$this->prefixes = []; |
|
52 |
$this->rawPrefixes = []; |
|
53 |
$res = preg_replace_callback("%".self::SPARQL_PREFIX_BASE_REGEXP."%iu", function($m) { |
|
54 |
$this->rawPrefixes[] = trim($m[0]); |
|
55 |
$this->prefixes[$m[3]?$m[3]:""] = $m[4]; |
|
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
56 |
return ""; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
57 |
}, $this->query); |
387 | 58 |
$res = preg_replace_callback("%".self::SPARQL_LIMIT_OFFSET_QUERY_REGEXP."%iu", function($m) { |
59 |
for($i=0;$i<(count($m)-1)/2;$i++) { |
|
60 |
if(Utils::startsWith(strtolower($m[2*$i+1]), "limit")) { |
|
61 |
$this->limit = intval($m[$i*2+2]); |
|
62 |
} elseif (Utils::startsWith(strtolower($m[2*$i+1]), "offset")) { |
|
63 |
$this->offset = intval($m[$i*2+2]); |
|
64 |
} |
|
65 |
} |
|
66 |
}, $res); |
|
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
67 |
|
387 | 68 |
$this->rawQuery = trim($res); |
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
69 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
70 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
71 |
public function getRawPrefixes() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
72 |
if($this->rawPrefixes === false) { |
387 | 73 |
$this->extractPrefixLimit(); |
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
74 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
75 |
return $this->rawPrefixes; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
76 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
77 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
78 |
public function getPrefixes() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
79 |
if($this->prefixes === false) { |
387 | 80 |
$this->extractPrefixLimit(); |
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
81 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
82 |
return $this->prefixes; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
83 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
84 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
85 |
public function getRawQuery() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
86 |
if($this->rawQuery === false) { |
387 | 87 |
$this->extractPrefixLimit(); |
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
88 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
89 |
return $this->rawQuery; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
90 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
91 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
92 |
public function getCountVar() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
93 |
if($this->countVar === false) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
94 |
$this->countVar = "?count_cp_".hash('md5', $this->query); |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
95 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
96 |
return $this->countVar; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
97 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
98 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
99 |
public function getCountQuery() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
100 |
return implode(" ", $this->getRawPrefixes())." select (count(*) as ".$this->getCountVar().") { ".$this->getRawQuery()." }"; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
101 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
102 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
103 |
private function setLimitOffset() { |
387 | 104 |
$this->extractPrefixLimit(); |
386
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
105 |
if($this->limit === false) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
106 |
$this->limit = null; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
107 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
108 |
if($this->offset === false) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
109 |
$this->offset = null; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
110 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
111 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
112 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
113 |
public function getLimit() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
114 |
if($this->limit === false) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
115 |
$this->setLimitOffset(); |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
116 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
117 |
return $this->limit; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
118 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
119 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
120 |
public function getOffset() { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
121 |
if($this->offset === false) { |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
122 |
$this->setLimitOffset(); |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
123 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
124 |
return $this->offset; |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
125 |
} |
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
126 |
|
c731ab9b934d
implement first version of sparql client interface
ymh <ymh.work@gmail.com>
parents:
diff
changeset
|
127 |
} |