Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
100.00% |
54 / 54 |
|
100.00% |
3 / 3 |
CRAP | |
100.00% |
1 / 1 |
| N3Quoter | |
100.00% |
54 / 54 |
|
100.00% |
3 / 3 |
6 | |
100.00% |
1 / 1 |
| setEscapeUnicode | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
2 | |||
| escapeIRI | |
100.00% |
12 / 12 |
|
100.00% |
1 / 1 |
1 | |||
| escapeLiteral | |
100.00% |
41 / 41 |
|
100.00% |
1 / 1 |
3 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace Wikimedia\Purtle; |
| 4 | |
| 5 | /** |
| 6 | * Helper class for quoting literals and URIs in N3 output. |
| 7 | * Optionally supports shorthand and prefix resolution. |
| 8 | * |
| 9 | * @license GPL-2.0-or-later |
| 10 | * @author Daniel Kinzler |
| 11 | */ |
| 12 | class N3Quoter { |
| 13 | |
| 14 | /** |
| 15 | * @var UnicodeEscaper |
| 16 | */ |
| 17 | private $escaper = null; |
| 18 | |
| 19 | /** |
| 20 | * @param bool $escapeUnicode |
| 21 | */ |
| 22 | public function setEscapeUnicode( $escapeUnicode ) { |
| 23 | $this->escaper = $escapeUnicode ? new UnicodeEscaper() : null; |
| 24 | } |
| 25 | |
| 26 | /** |
| 27 | * @param string $iri |
| 28 | * |
| 29 | * @return string |
| 30 | */ |
| 31 | public function escapeIRI( $iri ) { |
| 32 | // FIXME: apply unicode escaping?! |
| 33 | return strtr( $iri, [ |
| 34 | ' ' => '%20', |
| 35 | '"' => '%22', |
| 36 | '<' => '%3C', |
| 37 | '>' => '%3E', |
| 38 | '\\' => '%5C', |
| 39 | '`' => '%60', |
| 40 | '^' => '%5E', |
| 41 | '|' => '%7C', |
| 42 | '{' => '%7B', |
| 43 | '}' => '%7D', |
| 44 | ] ); |
| 45 | } |
| 46 | |
| 47 | /** |
| 48 | * @param string $s |
| 49 | * |
| 50 | * @return string |
| 51 | */ |
| 52 | public function escapeLiteral( $s ) { |
| 53 | // Performance: If the entire string is just (a safe subset) of ASCII, let it through. |
| 54 | // Ok are space (31), ! (32), # (35) - [ (91) and ] (93) to ~ (126), excludes " (34) and \ (92). |
| 55 | if ( preg_match( '/^[ !#-[\]-~]*\z/', $s ) ) { |
| 56 | return $s; |
| 57 | } |
| 58 | |
| 59 | // String escapes. Note that the N3 spec is more restrictive than the Turtle and TR |
| 60 | // specifications, see <https://www.w3.org/TeamSubmission/n3/#escaping> |
| 61 | // and <https://www.w3.org/TR/turtle/#string> |
| 62 | // and <https://www.w3.org/TR/n-triples/#grammar-production-literal>. |
| 63 | // Allowed escapes according to the N3 spec are: |
| 64 | // ECHAR ::= '\' [tbnrf"'\] |
| 65 | // The single quote however does not require escaping when used in double quotes. |
| 66 | $escaped = strtr( $s, [ |
| 67 | "\x00" => '\u0000', |
| 68 | "\x01" => '\u0001', |
| 69 | "\x02" => '\u0002', |
| 70 | "\x03" => '\u0003', |
| 71 | "\x04" => '\u0004', |
| 72 | "\x05" => '\u0005', |
| 73 | "\x06" => '\u0006', |
| 74 | "\x07" => '\u0007', |
| 75 | "\x08" => '\b', |
| 76 | "\x09" => '\t', |
| 77 | "\x0A" => '\n', |
| 78 | "\x0B" => '\u000B', |
| 79 | "\x0C" => '\f', |
| 80 | "\x0D" => '\r', |
| 81 | "\x0E" => '\u000E', |
| 82 | "\x0F" => '\u000F', |
| 83 | "\x10" => '\u0010', |
| 84 | "\x11" => '\u0011', |
| 85 | "\x12" => '\u0012', |
| 86 | "\x13" => '\u0013', |
| 87 | "\x14" => '\u0014', |
| 88 | "\x15" => '\u0015', |
| 89 | "\x16" => '\u0016', |
| 90 | "\x17" => '\u0017', |
| 91 | "\x18" => '\u0018', |
| 92 | "\x19" => '\u0019', |
| 93 | "\x1A" => '\u001A', |
| 94 | "\x1B" => '\u001B', |
| 95 | "\x1C" => '\u001C', |
| 96 | "\x1D" => '\u001D', |
| 97 | "\x1E" => '\u001E', |
| 98 | "\x1F" => '\u001F', |
| 99 | '"' => '\"', |
| 100 | '\\' => '\\\\', |
| 101 | ] ); |
| 102 | |
| 103 | if ( $this->escaper !== null ) { |
| 104 | $escaped = $this->escaper->escapeString( $escaped ); |
| 105 | } |
| 106 | |
| 107 | return $escaped; |
| 108 | } |
| 109 | |
| 110 | } |