1 #

     2 # # minixsv, Release 0.3

     3 # file: minidomif.py

     4 #

     5 # XML interface class to minidom

     6 #

     7 # history:

     8 # 2004-09-09 rl   created

     9 # 2004-09-22 rl   XML interface classes completely re-designed

    10 # 2004-09-23 rl   added filename and line number support

    11 # 2004-09-29 rl   URL processing added

    12 # 2004-10-01 rl   URL processing improved

    13 # 2004-10-12 rl   XML text processing added

    14 #

    15 # Copyright (c) 2004 by Roland Leuthe.  All rights reserved.

    16 #

    17 # --------------------------------------------------------------------

    18 # The minixsv XML schema validator is

    19 #

    20 # Copyright (c) 2004 by Roland Leuthe

    21 #

    22 # By obtaining, using, and/or copying this software and/or its

    23 # associated documentation, you agree that you have read, understood,

    24 # and will comply with the following terms and conditions:

    25 #

    26 # Permission to use, copy, modify, and distribute this software and

    27 # its associated documentation for any purpose and without fee is

    28 # hereby granted, provided that the above copyright notice appears in

    29 # all copies, and that both that copyright notice and this permission

    30 # notice appear in supporting documentation, and that the name of

    31 # the author not be used in advertising or publicity

    32 # pertaining to distribution of the software without specific, written

    33 # prior permission.

    34 #

    35 # THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD

    36 # TO THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANT-

    37 # ABILITY AND FITNESS.  IN NO EVENT SHALL THE AUTHOR

    38 # BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY

    39 # DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,

    40 # WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS

    41 # ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE

    42 # OF THIS SOFTWARE.

    43 # --------------------------------------------------------------------

    44 

    45 import urllib

    46 from xmlifbase            import *

    47 from xml.parsers.expat    import ExpatError

    48 from xml.dom.expatbuilder import ExpatBuilderNS

    49 from xml.dom.minidom      import *

    50 

    51 

    52 class MiniDomInterface (XmlInterfaceBase):

    53     #####################################################

    54     # for description of the interface methods see xmlifbase.py

    55     #####################################################

    56 

    57     def __init__ (self, verbose):

    58         XmlInterfaceBase.__init__ (self, verbose)

    59         if self.verbose:

    60             print "Using minidom interface module..."

    61 

    62 

    63     def parse (self, file, baseUrl):

    64         url    = self._convertToUrl(file)

    65         absUrl = self._convertToAbsUrl (url, baseUrl)

    66         fp     = urllib.urlopen (absUrl)

    67         try:

    68             builder = ExtExpatBuilderNS(file, url, absUrl)

    69             tree = builder.parseFile(fp)

    70         finally:

    71             fp.close()

    72 

    73         return TreeWrapper(self, tree)

    74 

    75 

    76     def parseString (self, text):

    77         builder = ExtExpatBuilderNS("", "", "")

    78         tree = builder.parseString (text)

    79         return TreeWrapper (self, tree)

    80 

    81 

    82 #########################################################

    83 # Wrapper class for minidom.Document class

    84 

    85 class TreeWrapper (TreeWrapperBase):

    86 

    87     def getRootNode (self):

    88         return ElementWrapper (self.xmlIf, self, self.tree.documentElement)

    89 

    90     def insertSubtree (self, nextSiblingWrapper, file, baseUrl):

    91         subTreeWrapper = self.xmlIf.extParse (file, baseUrl)

    92         assert self.getRootNode().getLocalName() == subTreeWrapper.getRootNode().getLocalName(),\

    93                "%s: line %d: root node of include file %s does not match!" %(self.getRootNode().getUrl(), self.getRootNode().getStartLineNumber(), file)

    94 

    95         elementWrapperList = subTreeWrapper.getRootNode().getChildren()

    96         for elementWrapper in elementWrapperList:

    97             if nextSiblingWrapper != None:

    98                 self.tree.documentElement.insertBefore(elementWrapper.element, nextSiblingWrapper.element)

    99             else:

   100                 self.tree.documentElement.appendChild(elementWrapper.element)

   101 

   102 

   103 #########################################################

   104 # Wrapper class for minidom.Document class

   105 

   106 class ElementWrapper (ElementWrapperBase):

   107 

   108     def getTagName (self):

   109         return self.element.tagName

   110 

   111 

   112     def getLocalName (self):

   113         # base method replaced (optimized method)

   114         return self.element.localName

   115 

   116 

   117     def getNamespaceURI (self):

   118         # base method replaced (optimized method)

   119         return self.element.namespaceURI

   120 

   121 

   122     def getChildren (self, filterTag=None):

   123         return map(lambda element: ElementWrapper(self.xmlIf, self.treeWrapper, element), self._getElemNodes(filterTag))

   124 

   125 

   126     def getElementsByTagName (self, filterTag=None):

   127         if filterTag == None:

   128             filterTag = '*'

   129         return map(lambda element: ElementWrapper(self.xmlIf, self.treeWrapper, element), self.element.getElementsByTagName(filterTag))

   130 

   131 

   132     def removeChild (self, childElementWrapper):

   133         self.element.removeChild(childElementWrapper.element)

   134 

   135 

   136     def getAttributeDict (self):

   137         """Return dictionary of attributes"""

   138         attribDict = {}

   139         for attrName, attrValue in self.element.attributes.items():

   140             attribDict[attrName] = attrValue

   141         return attribDict

   142 

   143 

   144     def getAttribute (self, attributeName):

   145         if self.element.attributes.has_key (attributeName):

   146             return self.element.getAttribute (attributeName)

   147         else:

   148             return None

   149 

   150 

   151     def hasAttribute (self, attributeName):

   152         return self.element.attributes.has_key (attributeName)

   153 

   154 

   155     def setAttribute (self, attributeName, attributeValue):

   156         if not self.hasAttribute(attributeName):

   157             self.element.setAttributeNode (self.treeWrapper.tree.createAttribute (attributeName))

   158         self.element.setAttribute(attributeName, attributeValue)

   159 

   160 

   161     ###############################################################

   162     # Caution! Currently returns only the first text node!!

   163     def getElementValue (self):

   164         textNodes = self._getTextNodes ()

   165         if textNodes:

   166             return textNodes[0].data

   167         else:

   168             return ""

   169 

   170 

   171     ###############################################################

   172     # Caution! Currently only sets the first text node!!

   173     def setElementValue (self, value):

   174         textNodes = self._getTextNodes()

   175         if textNodes:

   176             textNodes[0].data = value

   177 

   178 

   179     ###############################################################

   180     # PRIVATE methods

   181     ###############################################################

   182 

   183     def _getTextNodes (self):

   184         """Return list of TEXT nodes."""

   185         return filter (lambda e: (e.nodeType == Node.TEXT_NODE),        # - only TEXT-NODES

   186                        self.element.childNodes)                         # from element's child nodes

   187 

   188 

   189     def _getElemNodes (self, filterTag):

   190         """Return list of ELEMENT nodes and check for tag name (if given)."""

   191         return filter (lambda e: (e.nodeType == Node.ELEMENT_NODE) and                          # - only ELEMENTs

   192                                  (not filterTag or filterTag == '*' or e.tagName == filterTag), # - if tagName given --> check tagName

   193                        self.element.childNodes)                                                 # from element's nodes

   194 

   195 

   196 

   197 ###################################################

   198 # Expat builder class derived from ExpatBuilderNS

   199 # extended to store related line numbers in the node object

   200 

   201 class ExtExpatBuilderNS (ExpatBuilderNS):

   202     def __init__ (self, filePath, url, absUrl):

   203         self.filePath = filePath

   204         self.url      = url

   205         self.absUrl   = absUrl

   206         ExpatBuilderNS.__init__(self)

   207 

   208     def start_element_handler(self, name, attributes):

   209         ExpatBuilderNS.start_element_handler(self, name, attributes)

   210         self.curNode.url      = self.url

   211         self.curNode.absUrl   = self.absUrl

   212         self.curNode.filePath = self.filePath

   213         self.curNode.startLineNumber = self.getParser().ErrorLineNumber

   214 

   215     def end_element_handler(self, name):

   216         self.curNode.endLineNumber = self.getParser().ErrorLineNumber

   217         ExpatBuilderNS.end_element_handler(self, name)

   218