Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 1 | import handler |
| 2 | |
| 3 | """An XML Reader is the SAX 2 name for an XML parser. XML Parsers |
| 4 | should be based on this code. """ |
| 5 | # ===== XMLREADER ===== |
| 6 | |
| 7 | class XMLReader: |
| 8 | def __init__(self): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 9 | self._cont_handler = handler.ContentHandler() |
| 10 | #self._dtd_handler = handler.DTDHandler() |
| 11 | #self._ent_handler = handler.EntityResolver() |
| 12 | self._err_handler = handler.ErrorHandler() |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 13 | |
| 14 | def parse(self, source): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 15 | "Parse an XML document from a system identifier or an InputSource." |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 16 | raise NotImplementedError("This method must be implemented!") |
| 17 | |
| 18 | def getContentHandler(self): |
| 19 | "Returns the current ContentHandler." |
| 20 | return self._cont_handler |
| 21 | |
| 22 | def setContentHandler(self, handler): |
| 23 | "Registers a new object to receive document content events." |
| 24 | self._cont_handler = handler |
| 25 | |
| 26 | def getDTDHandler(self): |
| 27 | "Returns the current DTD handler." |
| 28 | return self._dtd_handler |
| 29 | |
| 30 | def setDTDHandler(self, handler): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 31 | "Register an object to receive basic DTD-related events." |
| 32 | self._dtd_handler = handler |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 33 | |
| 34 | def getEntityResolver(self): |
| 35 | "Returns the current EntityResolver." |
| 36 | return self._ent_handler |
| 37 | |
| 38 | def setEntityResolver(self, resolver): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 39 | "Register an object to resolve external entities." |
| 40 | self._ent_handler = resolver |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 41 | |
| 42 | def getErrorHandler(self): |
| 43 | "Returns the current ErrorHandler." |
| 44 | return self._err_handler |
| 45 | |
| 46 | def setErrorHandler(self, handler): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 47 | "Register an object to receive error-message events." |
| 48 | self._err_handler = handler |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 49 | |
| 50 | def setLocale(self, locale): |
| 51 | """Allow an application to set the locale for errors and warnings. |
| 52 | |
Thomas Wouters | 7e47402 | 2000-07-16 12:04:32 +0000 | [diff] [blame] | 53 | SAX parsers are not required to provide localization for errors |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 54 | and warnings; if they cannot support the requested locale, |
| 55 | however, they must throw a SAX exception. Applications may |
| 56 | request a locale change in the middle of a parse.""" |
| 57 | raise SAXNotSupportedException("Locale support not implemented") |
| 58 | |
| 59 | def getFeature(self, name): |
| 60 | "Looks up and returns the state of a SAX2 feature." |
| 61 | raise SAXNotRecognizedException("Feature '%s' not recognized" % name) |
| 62 | |
| 63 | def setFeature(self, name, state): |
| 64 | "Sets the state of a SAX2 feature." |
| 65 | raise SAXNotRecognizedException("Feature '%s' not recognized" % name) |
| 66 | |
| 67 | def getProperty(self, name): |
| 68 | "Looks up and returns the value of a SAX2 property." |
| 69 | raise SAXNotRecognizedException("Property '%s' not recognized" % name) |
| 70 | |
| 71 | def setProperty(self, name, value): |
| 72 | "Sets the value of a SAX2 property." |
| 73 | raise SAXNotRecognizedException("Property '%s' not recognized" % name) |
| 74 | |
| 75 | |
| 76 | class IncrementalParser(XMLReader): |
| 77 | """This interface adds three extra methods to the XMLReader |
| 78 | interface that allow XML parsers to support incremental |
| 79 | parsing. Support for this interface is optional, since not all |
| 80 | underlying XML parsers support this functionality. |
| 81 | |
| 82 | When the parser is instantiated it is ready to begin accepting |
| 83 | data from the feed method immediately. After parsing has been |
| 84 | finished with a call to close the reset method must be called to |
| 85 | make the parser ready to accept new data, either from feed or |
| 86 | using the parse method. |
| 87 | |
| 88 | Note that these methods must _not_ be called during parsing, that |
| 89 | is, after parse has been called and before it returns. |
| 90 | |
| 91 | By default, the class also implements the parse method of the XMLReader |
| 92 | interface using the feed, close and reset methods of the |
| 93 | IncrementalParser interface as a convenience to SAX 2.0 driver |
| 94 | writers.""" |
| 95 | def __init__(self, bufsize=2**16 ): |
| 96 | self._bufsize=bufsize |
| 97 | XMLReader.__init__( self ) |
| 98 | |
| 99 | def parse(self, source): |
| 100 | self.prepareParser(source) |
| 101 | #FIXME: do some type checking: could be already stream, URL or |
| 102 | # filename |
| 103 | inf=open( source ) |
| 104 | buffer = inf.read(self._bufsize) |
| 105 | while buffer != "": |
| 106 | self.feed(buffer) |
| 107 | buffer = inf.read(self._bufsize) |
| 108 | self.close() |
| 109 | self.reset() |
| 110 | |
| 111 | def feed(self, data): |
| 112 | """This method gives the raw XML data in the data parameter to |
| 113 | the parser and makes it parse the data, emitting the |
| 114 | corresponding events. It is allowed for XML constructs to be |
| 115 | split across several calls to feed. |
| 116 | |
| 117 | feed may raise SAXException.""" |
| 118 | raise NotImplementedError("This method must be implemented!") |
| 119 | def prepareParser(self, source): |
| 120 | """This method is called by the parse implementation to allow |
| 121 | the SAX 2.0 driver to prepare itself for parsing.""" |
| 122 | raise NotImplementedError("prepareParser must be overridden!") |
| 123 | |
| 124 | def close(self): |
| 125 | """This method is called when the entire XML document has been |
| 126 | passed to the parser through the feed method, to notify the |
| 127 | parser that there are no more data. This allows the parser to |
| 128 | do the final checks on the document and empty the internal |
| 129 | data buffer. |
| 130 | |
| 131 | The parser will not be ready to parse another document until |
| 132 | the reset method has been called. |
| 133 | |
| 134 | close may raise SAXException.""" |
| 135 | raise NotImplementedError("This method must be implemented!") |
| 136 | |
| 137 | def reset(self): |
| 138 | """This method is called after close has been called to reset |
| 139 | the parser so that it is ready to parse new documents. The |
| 140 | results of calling parse or feed after close without calling |
| 141 | reset are undefined.""" |
| 142 | raise NotImplementedError("This method must be implemented!") |
| 143 | |
| 144 | # ===== LOCATOR ===== |
| 145 | class Locator: |
| 146 | """Interface for associating a SAX event with a document |
| 147 | location. A locator object will return valid results only during |
| 148 | calls to DocumentHandler methods; at any other time, the |
| 149 | results are unpredictable.""" |
| 150 | |
| 151 | def getColumnNumber(self): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 152 | "Return the column number where the current event ends." |
| 153 | return -1 |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 154 | |
| 155 | def getLineNumber(self): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 156 | "Return the line number where the current event ends." |
| 157 | return -1 |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 158 | |
| 159 | def getPublicId(self): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 160 | "Return the public identifier for the current event." |
| 161 | return None |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 162 | |
| 163 | def getSystemId(self): |
Skip Montanaro | f9059eb | 2000-07-06 03:01:40 +0000 | [diff] [blame] | 164 | "Return the system identifier for the current event." |
| 165 | return None |
Fred Drake | 45cd9de | 2000-06-29 19:34:54 +0000 | [diff] [blame] | 166 | |
| 167 | # --- AttributesImpl |
| 168 | class AttributesImpl: |
| 169 | def __init__(self, attrs, rawnames): |
| 170 | self._attrs = attrs |
| 171 | self._rawnames = rawnames |
| 172 | |
| 173 | def getLength(self): |
| 174 | return len(self._attrs) |
| 175 | |
| 176 | def getType(self, name): |
| 177 | return "CDATA" |
| 178 | |
| 179 | def getValue(self, name): |
| 180 | return self._attrs[name] |
| 181 | |
| 182 | def getValueByQName(self, name): |
| 183 | return self._attrs[self._rawnames[name]] |
| 184 | |
| 185 | def getNameByQName(self, name): |
| 186 | return self._rawnames[name] |
| 187 | |
| 188 | def getNames(self): |
| 189 | return self._attrs.keys() |
| 190 | |
| 191 | def getQNames(self): |
| 192 | return self._rawnames.keys() |
| 193 | |
| 194 | def __len__(self): |
| 195 | return len(self._attrs) |
| 196 | |
| 197 | def __getitem__(self, name): |
| 198 | return self._attrs[name] |
| 199 | |
| 200 | def keys(self): |
| 201 | return self._attrs.keys() |
| 202 | |
| 203 | def has_key(self, name): |
| 204 | return self._attrs.has_key(name) |
| 205 | |
| 206 | def get(self, name, alternative=None): |
| 207 | return self._attrs.get(name, alternative) |
| 208 | |
| 209 | def copy(self): |
| 210 | return self.__class__(self._attrs, self._rawnames) |
| 211 | |
| 212 | def items(self): |
| 213 | return self._attrs.items() |
| 214 | |
| 215 | def values(self): |
| 216 | return self._attrs.values() |
| 217 | |
| 218 | def _test(): |
| 219 | XMLReader() |
| 220 | IncrementalParser() |
| 221 | Locator() |
| 222 | AttributesImpl() |
| 223 | |
| 224 | if __name__=="__main__": |
| 225 | _test() |