Parse XPath Expressions

问题:

I am trying to create a 'AET' (Abstract Expression Tree) for XPath (as I am writing a WYSIWYG XSL editor). I have been hitting my head against the wall with the XPath BNF for the past three to four hours.

I have thought of another solution. I thought I could write a class that implements IXPathNavigable, which returns a XPathNavigator of my own when CreateNavigator is called. This XPathNavigator would always succeed on any method calls, and would keep track of those calls - e.g. we moved to the customers node and then the customer node. I could then use this information (hopefully) to create the 'AET' (so we would have customers/customer in a object model now).

Only question is: how on earth do I run a IXPathNavigable through an XPathExpression?

I know this is excessively lazy. But has anyone else gone through the effort and written a XPath expression parser? I haven't yet POC'd my possible solution, because I can't test it (because I can't run the XPathExpression against a IXPathNavigable), so I don't even know if my solution will even work.

回答1:

There is an antlr xpath grammar here. Since its license permits, I copied the whole grammar here to avoid link rot in the future.

grammar xpath;

/*
XPath 1.0 grammar. Should conform to the official spec at
http://www.w3.org/TR/1999/REC-xpath-19991116. The grammar
rules have been kept as close as possible to those in the
spec, but some adjustmewnts were unavoidable. These were
mainly removing left recursion (spec seems to be based on
LR), and to deal with the double nature of the '*' token
(node wildcard and multiplication operator). See also
section 3.7 in the spec. These rule changes should make
no difference to the strings accepted by the grammar.
Written by Jan-Willem van den Broek
Version 1.0
Do with this code as you will.
*/
/*
    Ported to Antlr4 by Tom Everett <[e-mail]>
*/


main  :  expr
  ;

locationPath 
  :  relativeLocationPath
  |  absoluteLocationPathNoroot
  ;

absoluteLocationPathNoroot
  :  '/' relativeLocationPath
  |  '//' relativeLocationPath
  ;

relativeLocationPath
  :  step (('/'|'//') step)*
  ;

step  :  axisSpecifier nodeTest predicate*
  |  abbreviatedStep
  ;

axisSpecifier
  :  AxisName '::'
  |  '@'?
  ;

nodeTest:  nameTest
  |  NodeType '(' ')'
  |  'processing-instruction' '(' Literal ')'
  ;

predicate
  :  '[' expr ']'
  ;

abbreviatedStep
  :  '.'
  |  '..'
  ;

expr  :  orExpr
  ;

primaryExpr
  :  variableReference
  |  '(' expr ')'
  |  Literal
  |  Number  
  |  functionCall
  ;

functionCall
  :  functionName '(' ( expr ( ',' expr )* )? ')'
  ;

unionExprNoRoot
  :  pathExprNoRoot ('|' unionExprNoRoot)?
  |  '/' '|' unionExprNoRoot
  ;

pathExprNoRoot
  :  locationPath
  |  filterExpr (('/'|'//') relativeLocationPath)?
  ;

filterExpr
  :  primaryExpr predicate*
  ;

orExpr  :  andExpr ('or' andExpr)*
  ;

andExpr  :  equalityExpr ('and' equalityExpr)*
  ;

equalityExpr
  :  relationalExpr (('='|'!=') relationalExpr)*
  ;

relationalExpr
  :  additiveExpr (('<'|'>'|'<='|'>=') additiveExpr)*
  ;

additiveExpr
  :  multiplicativeExpr (('+'|'-') multiplicativeExpr)*
  ;

multiplicativeExpr
  :  unaryExprNoRoot (('*'|'div'|'mod') multiplicativeExpr)?
  |  '/' (('div'|'mod') multiplicativeExpr)?
  ;

unaryExprNoRoot
  :  '-'* unionExprNoRoot
  ;

qName  :  nCName (':' nCName)?
  ;

functionName
  :  qName  // Does not match nodeType, as per spec.
  ;

variableReference
  :  '$' qName
  ;

nameTest:  '*'
  |  nCName ':' '*'
  |  qName
  ;

nCName  :  NCName
  |  AxisName
  ;

NodeType:  'comment'
  |  'text'
  |  'processing-instruction'
  |  'node'
  ;

Number  :  Digits ('.' Digits?)?
  |  '.' Digits
  ;

fragment
Digits  :  ('0'..'9')+
  ;

AxisName:  'ancestor'
  |  'ancestor-or-self'
  |  'attribute'
  |  'child'
  |  'descendant'
  |  'descendant-or-self'
  |  'following'
  |  'following-sibling'
  |  'namespace'
  |  'parent'
  |  'preceding'
  |  'preceding-sibling'
  |  'self'
  ;


  PATHSEP 
       :'/';
  ABRPATH   
       : '//';
  LPAR   
       : '(';
  RPAR   
       : ')';
  LBRAC   
       :  '[';
  RBRAC   
       :  ']';
  MINUS   
       :  '-';
  PLUS   
       :  '+';
  DOT   
       :  '.';
  MUL   
       : '*';
  DOTDOT   
       :  '..';
  AT   
       : '@';
  COMMA  
       : ',';
  PIPE   
       :  '|';
  LESS   
       :  '<';
  MORE_ 
       :  '>';
  LE   
       :  '<=';
  GE   
       :  '>=';
  COLON   
       :  ':';
  CC   
       :  '::';
  APOS   
       :  '\'';
  QUOT   
       :  '\"';

Literal  :  '"' ~'"'* '"'
  |  '\'' ~'\''* '\''
  ;

Whitespace
  :  (' '|'\t'|'\n'|'\r')+ ->skip
  ;

NCName  :  NCNameStartChar NCNameChar*
  ;

fragment
NCNameStartChar
  :  'A'..'Z'
  |   '_'
  |  'a'..'z'
  |  '\u00C0'..'\u00D6'
  |  '\u00D8'..'\u00F6'
  |  '\u00F8'..'\u02FF'
  |  '\u0370'..'\u037D'
  |  '\u037F'..'\u1FFF'
  |  '\u200C'..'\u200D'
  |  '\u2070'..'\u218F'
  |  '\u2C00'..'\u2FEF'
  |  '\u3001'..'\uD7FF'
  |  '\uF900'..'\uFDCF'
  |  '\uFDF0'..'\uFFFD'
// Unfortunately, java escapes can't handle this conveniently,
// as they're limited to 4 hex digits. TODO.
//  |  '\U010000'..'\U0EFFFF'
  ;

fragment
NCNameChar
  :  NCNameStartChar | '-' | '.' | '0'..'9'
  |  '\u00B7' | '\u0300'..'\u036F'
  |  '\u203F'..'\u2040'
  ;

回答2:

I have both written an XPath parser and an implementation of IXPathNavigable (I used to be a developer for XMLPrime). Neither is easy; and I suspect that the IXPathNavigable is not going to be the cheap win you hope, as there is quite a lot of subtlety in the interactions between different methods - I suspect a full blown XPath parser will be simpler (and more reliable).

To answer your question though:

var results xpathNavigable.CreateNavigator().Evaluate("/my/xpath[expression]").

You'd probably need to enumerate over results to cause the node to be navigated.

If you always returned true then all you'd know about the following XPath is that it looks for bar children of foo: foo[not(bar)]/other/elements

If you always return a fixed number of nodes then you'd never know about most of this XPath a[100]/b/c/

Essentially, this won't work.