Added the ability to parse specdata XML files to create Schemas.

[~jspiros/python-ebml.git] / ebml / schema / specs.py

diff --git a/ebml/schema/specs.py b/ebml/schema/specs.py
new file mode 100644 (file)
index 0000000..7320b6b
--- /dev/null
+++ b/ebml/schema/specs.py
@@ -0,0 +1,74 @@
+from xml.etree.ElementTree import parse as parse_xml
+from .base import INT, UINT, FLOAT, STRING, UNICODE, DATE, BINARY, CONTAINER, Element, Schema
+
+
+SPECDATA_TYPES = {
+       'integer': INT,
+       'uinteger': UINT,
+       'float': FLOAT,
+       'string': STRING,
+       'utf-8': UNICODE,
+       'date': DATE,
+       'binary': BINARY,
+       'master': CONTAINER
+}
+
+
+def parse_specdata(source, schema_name):
+       """
+       
+       Reads a schema specification from a file (e.g., specdata.xml) or file-like object, and returns a tuple containing:
+       
+               * a mapping of class names to Element subclasses
+               * a Schema subclass
+       
+       :arg source: the file or file-like object
+       :type source: str or file-like object
+       :arg schema_name: the name of the schema
+       :type schema_name: str
+       :returns: tuple
+       
+       """
+       
+       tree = parse_xml(source)
+       elements = {}
+       parent_elements = []
+       
+       for element_element in tree.getiterator('element'):
+               raw_attrs = element_element.attrib
+               
+               element_name = '%sElement' % raw_attrs.get('cppname', raw_attrs.get('name'))
+               element_level = int(raw_attrs['level'])
+               element_attrs = {
+                       '__module__': None,
+                       'class_id': int(raw_attrs['id'], 0),
+                       'class_name': raw_attrs['name'],
+                       'data_type': SPECDATA_TYPES[raw_attrs['type']]
+               }
+               
+               while parent_elements and element_level <= parent_elements[-1][0]:
+                       parent_elements.pop()
+               
+               if element_level == -1:
+                       element_attrs['class_global'] = True
+                       parent_elements = []
+               elif element_level == 0:
+                       element_attrs['class_root'] = True
+                       parent_elements = []
+               else:
+                       if raw_attrs.get('recursive', '0') == '1':
+                               element_attrs['class_parents'] = (parent_elements[-1][1], 'self')
+                       else:
+                               element_attrs['class_parents'] = (parent_elements[-1][1],)
+               
+               element = type(element_name, (Element,), element_attrs)
+               elements[element_name] = element
+               parent_elements.append((element_level, element))
+       
+       schema_attrs = {
+               '__module__': None,
+               'elements': tuple(elements.values())
+       }
+       schema = type(schema_name, (Schema,), schema_attrs)
+       
+       return elements, schema
\ No newline at end of file