| 2 | | A working draft for the CQP flavor for CLARIN Federated Content Search (FCS). |
| 3 | | == EBNF == |
| 4 | | {{{#!comment |
| 5 | | Please keep the EBNF nicely formatted. Thanks! |
| 6 | | }}} |
| 7 | | {{{ |
| 8 | | [1] query ::= main-query within-part? |
| 9 | | |
| 10 | | [2] main-query ::= simple-query |
| 11 | | | "(" main-query ")" /* grouping */ |
| 12 | | | main-query "|" main-query /* or */ |
| 13 | | | main-query main-query /* sequence */ |
| 14 | | | main-query quantifier /* quatification */ |
| 15 | | |
| 16 | | [2.11] main-query ::= simple-query |
| 17 | | | simple-query "|" main-query /* or */ |
| 18 | | | simple-query main-query /* sequence */ |
| 19 | | | simple-query quantifier /* quatification */ |
| 20 | | |
| 21 | | [3] simple-query ::= implicit-query |
| 22 | | | segment-query |
| 23 | | |
| 24 | | [3.11] simple-query ::= '(' main_query ')' /* grouping */ |
| 25 | | | implicit-query |
| 26 | | | segment-query |
| 27 | | |
| 28 | | [4] implicit-query ::= flagged-regexp |
| 29 | | |
| 30 | | [5] segment-query ::= "[" expression? "]" |
| 31 | | |
| 32 | | [6] within-part ::= simple-within-part |
| 33 | | |
| 34 | | [7] simple-within-part ::= "within" simple-within-scope |
| 35 | | |
| 36 | | [8] simple-within-scope ::= "sentence" |
| 37 | | | "s" |
| 38 | | | "utterance" |
| 39 | | | "u" |
| 40 | | | "paragraph" |
| 41 | | | "p" |
| 42 | | | "turn" |
| 43 | | | "t" |
| 44 | | | "text" |
| 45 | | | "session" |
| 46 | | |
| 47 | | [11] expression ::= basic-expression |
| 48 | | | expression "|" expression /* or */ |
| 49 | | | expression "&" expression /* and */ |
| 50 | | | "(" expression ")" /* grouping */ |
| 51 | | | "!" expression /* not */ |
| 52 | | |
| 53 | | [11.11] expression ::= basic-expression |
| 54 | | | expression "|" expression /* or */ |
| 55 | | | expression "&" expression /* and */ |
| 56 | | |
| 57 | | |
| 58 | | [12] basic-expression ::= attribute operator flagged-regexp |
| 59 | | |
| 60 | | [12.11] basic-expression ::= '(' expression ')' /* grouping */ |
| 61 | | | "!" expression /* not */ |
| 62 | | | attribute operator flagged-regexp |
| 63 | | |
| 64 | | [13] operator ::= "=" /* equals */ |
| 65 | | | "!=" /* non-equals */ |
| 66 | | |
| 67 | | [14] quantifier ::= "+" /* one-or-more */ |
| 68 | | | "*" /* zero-or-more */ |
| 69 | | | "?" /* zero-or-one */ |
| 70 | | | "{" integer "}" /* exactly n-times */ |
| 71 | | | "{" integer? "," integer "}" /* at most */ |
| 72 | | | "{" integer "," integer? "}" /* min-max */ |
| 73 | | |
| 74 | | [15] flagged-regexp ::= regexp |
| 75 | | | regexp "/" regexp-flag+ |
| 76 | | |
| 77 | | [16] regexp-flag ::= "i" /* case-insensitive; Poliqarp/Perl compat */ |
| 78 | | | "I" /* case-sensitive; Poliqarp compat */ |
| 79 | | | "c" /* case-insensitive, CQP compat */ |
| 80 | | | "C" /* case-sensitive */ |
| 81 | | | "l" /* literal matching, CQP compat*/ |
| 82 | | | "d" /* diacritic agnostic matching, CQP compat */ |
| 83 | | |
| 84 | | [17] regexp ::= quoted-string |
| 85 | | |
| 86 | | [18] attribute ::= simple-attribute |
| 87 | | | qualified-attribute |
| 88 | | |
| 89 | | [19] simple-attribute ::= identifier |
| 90 | | |
| 91 | | [20] qualified-attribute ::= identifier ":" identifier |
| 92 | | |
| 93 | | [21] identifier ::= identifier-char identifier-char* |
| 94 | | |
| 95 | | [21.11] identifier ::= identifier-first-char identifier-char* |
| 96 | | |
| 97 | | [21.12] identifier-first-char ::= [a-zA-Z] |
| 98 | | |
| 99 | | [22] identifier-char ::= [a-zA-Z0-9\-] |
| 100 | | |
| 101 | | [24] integer ::= [0-9]+ |
| 102 | | |
| 103 | | [26] quoted-string ::= "'" (char | ws)* "'" /* single-quotes */ |
| 104 | | | """ (char | ws)* """ /* double-quotes */ |
| 105 | | |
| 106 | | [27] char ::= <any unicode codepoint excluding whitespace codepoints> |
| 107 | | | "\" escaped-char |
| 108 | | |
| 109 | | [28] ws ::= <any whitespace codepoint> |
| 110 | | |
| 111 | | [29] escaped-char ::= "\" /* backslash (\) */ |
| 112 | | | "'" /* single quote (') */ |
| 113 | | | """ /* double quote (") */ |
| 114 | | | "n" /* generic newline, i.e "\n", "\r", etc */ |
| 115 | | | "t" /* character tabulation (U+0009) */ |
| 116 | | | "x" hex hex /* Unicode codepoint with hex value hh */ |
| 117 | | | "u" hex hex hex hex /* Unicode codepoint with hex value hhhh */ |
| 118 | | | "U" hex hex hex hex hex hex hex hex /* Unicode codepoint with hex value hhhhhhhh */ |
| 119 | | |
| 120 | | [30] hex ::= [0-9a-fA-F] |
| 121 | | }}} |
| 122 | | == Notes == |
| 123 | | * based on Poliqarp with inspiration from others |
| 124 | | * "attribute": the annotation layer to be used, e.g. "word", "lemma", "pos" or qualified "pos:stts" the supported values for this construct are beyond the grammar and need to be defined in supplementary documents |
| 125 | | * "simple-within-scope": possible values for scope |
| 126 | | * "sentence", "s", "utterance", "u": denote a matching scope of something like a sentence or utterance. provides compatibility with FCS 1.0 ("Generic Hits", "Each hit SHOULD be presented within the context of a complete sentence.") |
| 127 | | * "paragraph" | "p" | "turn" | "t": denote the next larger unit, e.g. something like a paragraph |
| 128 | | * "article" | "session": something like a whole document |
| 129 | | * {{{[27]}}} and {{{[28]}}} "any $SOMETING codepoint" are a pain to get easily done in at least ANTLR and JavaCC. Especially in combination with {{{[29]}}} :/ |
| 130 | | * regex are not defined/guarded by this grammar :/ |
| 131 | | * non-continuous rule numbers are currently intended; we've already removed some. Rules will be renumbered, when grammar is fixed. |
| 132 | | * Integrated Peter B's suggestion {{{[2v2]}}} and {{{[3v2]}}} together with {{{[11v2]}}} and {{{[12v2]}}} for resolving structural ambiguity eventhough antlr handles this perfectly fine. |
| 133 | | * Changed "identifier" {{{[21]}}} to only be allowed to start with a letter e.g. not digits and - (hyphen) to more resemble XML names. |
| 134 | | |
| 135 | | == Discussion == |
| 136 | | Peter Beinema, MPI, proposes some minor changes to the grammar: |
| 137 | | * main query {{{[2]}}} / simple query {{{[3]}}}: above definition generate structural ambiguity. Not a problem for ANTLR (which selects the right-recursive solution), but some other parser generators generate all solutions - which are exponential wrt the number of main queries. I propose to use the alternative rules given below. |
| 138 | | * above rule {{{[2]}}} can generate infinite array of quantifiers: {{{"word" +*{23}{,17}{2,}?+}}} would be legal. |
| 139 | | * rule {{{[5]}}}: option marker '?' makes "[]" a valid query. Propose to remove question mark. |
| 140 | | * expression {{{11}}} / basic expression {{{[12]}}}: structural ambiguity. See proposed alternative below. |
| 141 | | * attached file contains an antlr4 grammar, including comments on how to use it on mac / unix / linux platform |
| 142 | | == ENBF == |
| 143 | | {{{ |
| 144 | | [2 v2] /* in combination with [3 v2] no more left recursion or ambiguity, max 1 quantifier per simple-query */ |
| 145 | | main-query ::= |
| 146 | | simple-query |
| 147 | | | simple-query '|' main-query /* 'or' */ |
| 148 | | | simple-query main-query /* sequence */ |
| 149 | | | simple-query quantifier /* quantification */ |
| 150 | | |
| 151 | | [3 v2] |
| 152 | | simple_query ::= |
| 153 | | '(' main_query ')' /* embedding moved from main-query to simple-query |
| 154 | | | implicit-query |
| 155 | | | segment-query |
| 156 | | |
| 157 | | [5 v2] /* expression no longer optional */ |
| 158 | | segment-query : |
| 159 | | '[' expression ']' |
| 160 | | |
| 161 | | [11 v2] /* in combination with [12 v2] no longer left recursive or ambiguous |
| 162 | | expression : |
| 163 | | basic-expression |
| 164 | | | basic-expression '|' expression // or |
| 165 | | | basic-expression '&' expression // and |
| 166 | | |
| 167 | | [12 v2] |
| 168 | | basic-expression : |
| 169 | | '(' expression ')' // grouping |
| 170 | | | '!' expression // not |
| 171 | | | attribute operator flagged-regexp |
| 172 | | }}} |
| 173 | | |
| | 2 | The FCS-QL is integrated in the CLARIN-FCS Core 2.0 specification. See section [https://trac.clarin.eu/wiki/FCS/Specification-Draft#FCS-QL]. |
