Awesome

xmllpegparser

xmllpegparser is a fast XML parser who uses LPeg library.

1. Installation
2. Test
3. xmllpegparser API
   1. Parsing
   2. Entity
   3. Parser
   4. Utility
   5. Document structure (default parser)
   6. Parser structure
   7. Visitor structure
   8. Default parser limitations
4. Licence

Installation

luarocks install --local https://raw.githubusercontent.com/jonathanpoelen/lua-xmllpegparser/master/xmllpegparser-2.2-0.rockspec

# or in your local directory lua-xmllpegparser

luarocks make --local xmllpegparser-2.2-0.rockspec

Test

Run ./example.lua.

./example.lua xmlfile [replaceentities]

replaceentities = anything, only to enable replacement of entities.

xmllpegparser API

Parsing

• parse(xmlstring[, visitorOrsubEntities[, visitorInitArgs...]]):
  Returns a tuple document table, (string error or nil) (see visitor.finish).
  If subEntities is true, the entities are replaced and a tentity member is added to the document table.
• parseFile(filename[, visitorOrsubEntities[, visitorInitArgs...]]):
  Returns a tuple document table, error file or error document.

Entity

• defaultEntitiyTable():
  Returns the default entity table ({ quot='"', ... }).
• createEntityTable(docEntities[, resultEntities]):
  Creates an entity table from the document entity table. Return resultEntities.
• mkReplaceEntities(entityTable_or_func):
  Returns an LPeg expression that can replace entities
• replaceEntities(s, entityTable_or_func):
  Returns a string.

Parsers

• parser(visitor[, safeVisitor: bool]):
  Returns a parser. If all visitor functions return nil (excepted accuattr, init and finish), then safeVisitor may be true and the parser will optimize the visitor's calls.
• lazyParser(visitorCreator):
  Returns a parser.
  parser(visitorCreator()) is used on the first call of myparser.parse(...).
• mkVisitor(evalEntities: bool, defaultEntities: table | function | nil, withoutPosition):
  If not defaultEntities and evalEntities then defaultEntities = defaultEntityTable().
  If withoutPosition, then pos parameter does not exist for the visitor functions except for finish.
• treeParser:
  The default parser used by parse(str, false)
• treeParserWithReplacedEntities:
  The default parser used by parse(str, true)
• treeParserWithoutPos:
  Parser without pos parameter
• treeParserWithoutPosWithReplacedEntities:
  Parser without pos parameter

Global parser options

• enableWithoutPosParser([bool]):
  Enable default parser with treeParserWithoutPos* version.
  enableParserWithoutPos(false) is same to setDefaultParsers().
  Returns the previous parsers.
• setDefaultParsers(parser, parserWithReplacedEntities | bool | nil):
  If parserWithReplacedEntities == true, then parserWithReplacedEntities = p.
  nil or false value restore the default parser.
  Returns the previous parsers.

Utility

• toString(doc: table, indentationText: nil | string, params: nil | table):\
  • indentationText corresponds to the text used at each indentation level. If nil, there is no formatting.
  • params is table with
    • shortEmptyElements: bool = true: empty tag are self-closed or not.
    • stableAttributes: bool | function = true: If true, attribute are sorted by name. If a function, it takes the attribute table and should return an iterator function that gives the attribute name and its value.
    • inlineTextLengthMax: number = 9999999: a node that contains only one text is formatted on one line. When the text exceeds this value, it is indented.
    • escape: table: table of function(string):string
      • attr: text in double quote
      • text: text node
      • cdata: text between <![CDATA[ and ]]>
      • comment: text between <!-- and -->
• escapeFunctions(escapeAmp: bool = false):
  Utility function for params.escape parameter of toString
  • escapeAmp: escape & char in text and attribute
• escapeComment(string):string: replace -- with —
• escapeAttribute(string):string: replace < with < and " with "
• escapeAttributeAndAmp(string):string: like escapeAttribute + replace & with &
• escapeCDATA(string):string: replace ]]> with ]]>]]><![CDATA[
• escapeText(string):string: replace < with <
• escapeTextAndAmp(string):string replace < with < and & with &

Document structure (default parser)

-- pos member = index of string
document = {
  children = {
    { pos=number, parent=table or nil, text=string[, cdata=true] } or
    { pos=number, parent=table or nil, tag=string, attrs={ { name=string, value=string }, ... }, children={ ... } },
    ...
  },
  bad = { children={ ... } } -- when a closed node has no match
  preprocessor = { { pos=number, tag=string, attrs={ { name=string, value=string }, ... } },
  doctype = { pos=number, name=string, ident=string or nil, pubident=string or nil, dtd=string or nil }, -- if there is a doctype
  error = string, -- if error
  lastpos = number, -- last known position of parse()
  entities = { { pos=number, name=string, value=string }, ... },
  tentities = { name=value, ... } -- only if subEntities = true
}

Parser structure

{
  parse = function(xmlstring, visitorInitArgs...) ... end,
  parseFile = function(filename, visitorInitArgs...) ... end,
  __call = function(xmlstring, visitorInitArgs...) ... end,
}

Visitor structure

Each member is optionnal.

{
  withPos = bool -- indicates if pos parameter exists in function parameter (except `finish`)
  init = function(...), -- called before parsing, returns the position of the beginning of match or nil
  finish = function(err, pos, xmlstring), -- called after parsing, returns (doc, err) or nil
  proc = function(pos, name, attrs), -- for `<?...?>`
  entity = function(pos, name, value),
  doctype = function(pos, name, ident, pubident, dtd), -- called after all entity()
  accuattr = function(table, name, value), -- `table` is an accumulator that will be transmitted to tag.attrs. Set to `false` for disable this function.
                                           -- If `nil` and `tag` is `not nil`, a default accumalator is used.
                                           -- If `false`, the accumulator is disabled.
                                           -- (`tag(pos, name, accuattr(accuattr({}, attr1, value1), attr2, value2)`)
  tag = function(pos, name, attrs), -- for a new tag (`<a>` or `<a/>`)
  open = function(), -- only for a open node (`<a>` not `<a/>`), called after `tag`.
  close = function(name),
  text = function(pos, text),
  cdata = function(pos, text), -- or `text` if nil
  comment = function(str)
}

Default parser limitations

• Non-validating
• No DTD support
• Ignore processing instructions

Licence

MIT license