def extract_domain(top_domain_file, full_extract=True):
"""
Extract a dialogue domain from the XML specification
:param top_domain_file: the filename of the top XML file
:param full_extract: whether to extract the full domain or only the files
:return: the extracted dialogue domain
"""
# create a new, empty domain
domain = Domain()
# determine the root path and filename
fl = open(top_domain_file, 'r')
domain.set_source_file(Path(top_domain_file))
# extract the XML document
document = XMLUtils.get_xml_document(fl)
main_node = XMLUtils.get_main_node(document)
root_path = Path(top_domain_file).parent
for child in main_node:
domain = XMLDomainReader.extract_partial_domain(
child, domain, root_path, full_extract)
return domain
def extract_bayesian_network_from_string(full_string):
"""
Extracts the bayesian network from a XML string.
:param full_string: the string containing the initial state content
:return: the corresponding Bayesian network
"""
# extract the XML document
document = XMLUtils.get_xml_document(io.StringIO(full_string))
main_node = XMLUtils.get_main_node(document)
if main_node.tag == 'state':
return XMLStateReader.get_bayesian_network(main_node)
for child_node in main_node:
if child_node.tag == 'state':
return XMLStateReader.get_bayesian_network(child_node)
return BNetwork()
def extract_dialogue(data_file):
"""
Extracts the dialogue specified in the data file. The result is a list of
dialogue state (one for each turn).
:param data_file: the XML file containing the turns
:return: the list of dialogue state
"""
doc = XMLUtils.get_xml_document(data_file)
main_node = XMLUtils.get_main_node(doc)
f = open(data_file)
root_path = f.name
sample = []
for node in main_node:
node_name = node.keys()[0]
if "Turn" in node_name:
state = DialogueState(
XMLStateReader.get_bayesian_network(node))
sample.append(state)
if node_name == "systemTurn" and state.has_chance_node("a_m"):
assign = Assignment("a_m",
state.query_prob("a_m").get_best())
state.add_evidence(assign)
elif node_name == "wiazard":
assign = Assignment.create_from_string(
node.get_first_child().get_node_value().trim())
sample[-1].add_evidence(assign)
elif node_name == "import":
file_name = main_node.get_attributes().get_named_item(
"href").get_node_value()
points = XMLDialogueReader.extract_dialogue(root_path + "/" +
file_name)
sample.append(points)
return sample
def extract_bayesian_network(file, tag):
"""
Returns the initial state or parameters from the XML document, for the given
domain (where the variable types are already declared)
:param file: the file to process
:param tag: the XML tag to search for
:return: the specified Bayesian network
"""
# extract the XML document
document = XMLUtils.get_xml_document(file)
main_node = XMLUtils.get_main_node(document)
if main_node.tag == tag:
return XMLStateReader.get_bayesian_network(main_node)
for child_node in main_node:
if child_node.tag == tag:
return XMLStateReader.get_bayesian_network(child_node)
raise ValueError()
def extract_partial_domain(main_node, domain, root_path, full_extract):
"""
Extracts a partially specified domain from the XML node and add its content to
the dialogue domain.
:param main_node: main XML node
:param domain: dialogue domain
:param root_path: root path (necessary to handle references)
:param full_extract: whether to extract the full domain or only the files
:return: the augmented dialogue domain
"""
tag = main_node.tag
if tag == 'domain':
# extracting rule-based probabilistic model
for child in main_node:
domain = XMLDomainReader.extract_partial_domain(
child, domain, root_path, full_extract)
elif tag == 'import':
# extracting imported references
try:
file_name = main_node.attrib['href']
file_path = str(root_path) + os.sep + file_name
fl = Path(file_path)
domain.add_imported_files(fl)
sub_document = XMLUtils.get_xml_document(file_path)
domain = XMLDomainReader.extract_partial_domain(
XMLUtils.get_main_node(sub_document), domain, root_path,
full_extract)
except:
raise ValueError()
if not full_extract:
return domain
if tag == 'settings':
# extracting settings
settings = XMLUtils.extract_mapping(main_node)
domain.get_settings().fill_settings(settings)
if tag == 'function':
# extracting custom functions
# try:
domain_function_name = main_node.attrib['name'].strip()
module_name, actual_function_name = main_node.text.rsplit('.', 1)
mod = importlib.import_module(module_name)
func = getattr(mod, actual_function_name)
domain.get_settings().add_function(domain_function_name, func)
# except:
# raise ValueError()
if tag == 'initialstate':
# extracting initial state
state = XMLStateReader.get_bayesian_network(main_node)
domain.set_initial_state(DialogueState(state))
if tag == 'model':
# extracting rule-based probabilistic model
model = XMLDomainReader._create_model(main_node)
domain.add_model(model)
if tag == 'parameters':
# extracting parameters
parameters = XMLStateReader.get_bayesian_network(main_node)
domain.set_parameters(parameters)
if XMLUtils.has_content(main_node):
if main_node == '#text': # TODO: main_node -> main_node.tag ??
raise ValueError()
return domain