"""
This module provides classes to represent and visualize LFG parse trees.
The LFGParseTree class provides a visualization of the tree structure of the sentence,
as well as functional annotations in the lexical items.
"""
import networkx as nx
import matplotlib.pyplot as plt
from typing import *
[docs]class LFGParseTreeNode:
"""
A class representing a node in a LFG (Lexical Functional Grammar) parse tree.
Each node has a label, which can be either a lexical or functional category, and a list of child nodes.
It also has a token attribute, which is the string representation of the token the node represents.
Parameters:
- label (str): The label of the node, typically a non-terminal symbol or a terminal symbol.
- token (str): The token the node represents, if any.
- functional_labels (Dict[str, str]): functional labels of the lexical item in the XLFG standard format, where keys are the functional labels and values are the corresponding label values
- children (List[LFGParseTreeNode]): The children of the node.
"""
def __init__(self, label: str, token: str, functional_labels: dict = None, children=None):
self.label = label
self.token = token
self.functional_labels = functional_labels if functional_labels is not None else {}
self.children = children or []
self.e_structure = None
def __repr__(self):
"""
Return a string representation of the node.
Returns:
- str: A string representation of the node.
"""
return f"LFGParseTreeNode(label={self.label}, token={self.token}, functional_labels={self.functional_labels}, children={self.children})"
def add_child(self, child):
self.children.append(child)
[docs] def add_functional_label(self, label: str, value: str):
"""
Add a new functional label
:param label: (str) functional label
:param value: (str) corresponding label value
"""
self.functional_labels[label] = value
[docs] def get_functional_label(self, label: str):
"""
get the value of the functional label specified
:param label: (str) functional label
:return: (str) corresponding label value
"""
return self.functional_labels.get(label)
[docs] def remove_functional_label(self, label: str):
"""
remove the specified functional label from the functional_labels
:param label: (str) functional label
"""
self.functional_labels.pop(label, None)
[docs] def get_all_functional_labels(self):
"""
get all functional labels
:return: dict of functional labels in the XLFG standard format
"""
return self.functional_labels
[docs] def calculate_e_structure(self):
"""
Calculate the e-structure representation of the parse tree node
"""
e_structure = {}
if self.label in e_structure:
e_structure[self.label] = (self.f_structure, [])
for child in self.children:
child.calculate_e_structure()
e_structure[self.label][1].append(child.e_structure)
self.e_structure = e_structure
[docs] def is_leaf(self):
"""
Determine if the node is a leaf node (i.e., has no children).
Returns:
- bool: True if the node is a leaf node, False otherwise.
"""
return not self.children
[docs]class LFGParseTreeNodeF(LFGParseTreeNode):
"""
Class representing a node in a LFG parse tree. This class is a child class of LFGParseTreeNode and adds functionality for storing and manipulating an F-structure.
"""
[docs] def add_to_f_structure(self, attribute: str, value: str):
"""
Add a new attribute and value to the f_structure of the node.
"""
self.f_structure.add(attribute, value)
[docs] def set_in_f_structure(self, attribute: str, value: str):
"""
Set the value of an attribute in the f_structure of the node.
"""
self.f_structure.set(attribute, value)
[docs] def get_from_f_structure(self, attribute: str):
"""
Get the value of an attribute in the f_structure of the node.
"""
return self.f_structure.get(attribute)
[docs] def remove_from_f_structure(self, attribute: str):
"""
Remove an attribute from the f_structure of the node.
"""
self.f_structure.remove(attribute)
[docs] def get_all_f_structure(self):
"""
Returns all attributes and values in the f_structure of the node.
"""
return self.f_structure.get_all()
[docs] def has_in_f_structure(self, attribute: str):
"""
Check if the attribute is present in the f_structure of the node.
"""
return self.f_structure.has(attribute)
[docs] def display_f_structure(self):
"""
Display the f_structure of the node.
"""
return self.f_structure.display()
[docs] def set_path_stm(self, path_stm):
"""
Set the path_stm of the node.
"""
self.path_stm = path_stm
[docs] def get_path_stm(self):
"""
Get the path_stm of the node.
"""
return self.path_stm
[docs]class LFGParseTreeNodeE(LFGParseTreeNode):
"""
Class representing a node in a LFG parse tree. This class is a child class of LFGParseTreeNode and adds functionality for storing and manipulating an E-structure.
"""
[docs] def init(self, label: str, token: str, children=None):
"""
:param label: label of the parse tree node
:type label: str
:param token: token of the parse tree node
:type token: str
:param children: children of the parse tree node, None by default
:type children: List[LFGParseTreeNodeE]
"""
super().__init__(label, token, children)
self.syntactic_category = None
self.tense = None
self.aspect = None
self.semantic_role = None
self.lexical_head = None
self.logical_form = None
self.mood = None
[docs] def set_syntactic_category(self, category: str):
"""
Set the syntactic category for this parse tree node
:param category: the syntactic category
:type category: str
"""
self.syntactic_category = category
[docs] def get_syntactic_category(self):
"""
Get the syntactic category for this parse tree node
:return: the syntactic category
:rtype: str
"""
return self.syntactic_category
[docs] def set_tense(self, tense: str):
"""
Set the tense for this parse tree node
:param tense: the tense
:type tense: str
"""
self.tense = tense
[docs] def get_tense(self):
"""
Get the tense for this parse tree node
:return: the tense
:rtype: str
"""
return self.tense
[docs] def set_aspect(self, aspect: str):
"""
Set the aspect for this parse tree node
:param aspect: the aspect
:type aspect: str
"""
self.aspect = aspect
[docs] def get_aspect(self):
"""
Get the aspect for this parse tree node
:return: the aspect
:rtype: str
"""
return self.aspect
[docs] def set_mood(self, mood: str):
"""
Set the mood for this parse tree node
:param mood: the mood
:type mood: str
"""
self.mood = mood
[docs] def get_mood(self):
"""
Get the mood for this parse tree node
:return: the mood
:rtype: str
"""
return self.mood
[docs] def set_semantic_role(self, role: str):
"""
Set the semantic role for this parse tree node
:param role: the semantic role
:type role: str
"""
self.semantic_role = role
[docs] def get_semantic_role(self):
"""
Get the semantic role for this parse tree node
"""
return self.semantic_role
[docs] def set_lexical_head(self, head: str):
"""
Set the lexical head for this parse tree node
:param role: the lexical head
:type role: str
"""
self.lexical_head = head
[docs] def get_lexical_head(self):
"""
Get the lexical head for this parse tree node
"""
return self.lexical_head
[docs]class LFGParseTree:
"""
The LFGParseTree class represents a parse tree generated using Lexical Functional Grammar (LFG) parsing.
It has a root node, which is an instance of the LFGParseTreeNode class, and a sentence string.
The class has several methods such as set_sentence, which allows setting the sentence string for the parse tree, is_leaf, which checks if the current node is a leaf node, to_string, which converts the parse tree to a string representation, f_structure_matrix, which returns the f-structure represented as an attribute-value matrix, to_latex, which writes the parse tree to a LaTeX file, to_networkx, which returns a directed graph (networkx DiGraph) representation of the parse tree and draw, which draws the parse tree using a specified layout and color map.
The class also has helper methods that are used internally by the other methods to perform specific tasks such as _to_c_structure_latex_helper which is used by to_latex method to recursively convert the c-structure parse tree to LaTeX format.
"""
def __init__(self, root: LFGParseTreeNode):
self.root = root
self.sentence = ""
[docs] def set_sentence(self, sentence: str):
"""
Set the sentence string for the parse tree.
Parameters:
sentence (str): The sentence string.
"""
self.sentence = sentence
[docs] def is_leaf(self):
"""
Check if the current node is a leaf node.
Returns:
bool: True if the current node is a leaf, False otherwise.
"""
return not self.children
[docs] def to_string(self) -> str:
"""
Convert the parse tree to a string representation.
Returns:
str: The string representation of the parse tree.
"""
if self.is_leaf():
return self.token
child_strings = [child.to_string() for child in self.children]
return f"({self.label} {' '.join(child_strings)})"
[docs] def f_structure_matrix(self):
"""
Returns the f-structure represented as an attribute-value matrix.
"""
# Initialize the matrix
matrix = []
# Recursive function to traverse the parse tree
def traverse(node):
# Get the functional labels of the current node
labels = node.get_all_functional_labels()
# If the node has children, recursively traverse them
if node.children:
for child in node.children:
traverse(child)
# If the node is a leaf, add its functional labels to the matrix
else:
row = []
for key, value in labels.items():
row.append((key, value))
matrix.append(row)
# Start traversing the tree
traverse(self.root)
# Return the matrix
return matrix
[docs] def to_latex(self, filepath: str):
"""
Write the parse tree to a LaTeX file.
Parameters:
filepath (str): The path to the LaTeX file.
"""
f_structure_latex = self._to_latex_helper(self.root)
c_structure_latex = self._to_c_structure_latex_helper(self.root)
caption = "C-structure tree and F-structure representation of the parse tree"
label = "parse_tree"
latex_code = f"""
\\documentclass{{article}}
\\usepackage{{tikz-qtree}}
\\usepackage{{tabularx}}
\\usepackage[active,tightpage]{{preview}}
\\usepackage{{graphicx}}
\\PreviewEnvironment{{center}}
\\begin{{document}}
\\begin{{figure}}[h]
\\centering
\\begin{{tabularx}}{{1.5\\textwidth}}{{X l}}
{c_structure_latex} & {f_structure_latex}\\\\
\\end{{tabularx}}
\\caption{{{caption}}}
\\label{{{label}}}
\\end{{figure}}
\\end{{document}}
"""
with open(filepath, "w") as f:
f.write(latex_code)
[docs] def _to_c_structure_latex_helper(self, node):
"""
Helper method that recursively converts the c-structure parse tree to latex format.
"""
# children = " ".join([self._to_c_structure_latex
[docs] def to_networkx(self):
"""
Returns a directed graph (networkx DiGraph) representation of the parse tree.
"""
graph = nx.DiGraph()
stack = [(self.root, None)]
while stack:
node, parent = stack.pop()
graph.add_edge(parent, node.label)
for child in node.children:
stack.append((child, node.label))
return graph
[docs] def draw(self, layout: str = 'spring', color_map: Dict[str, str] = None):
"""
Draws the parse tree using a specified layout and color map.
If a color map is not provided, the default is white for all nodes.
Example:
tree = LFGParseTree(root_node)
color_map = {'NP': 'red', 'VP': 'green', 'PP': 'blue'}
tree.draw(layout='spring', color_map=color_map)
"""
G = nx.Graph()
queue = [(self.root, None)]
while queue:
node, parent = queue.pop(0)
G.add_node(node.label)
if parent:
G.add_edge(parent.label, node.label)
for child in node.children:
queue.append((child, node))
if color_map:
color_values = [color_map.get(node.label, 'white') for node in G.nodes()]
nx.draw(G, pos=nx.drawing.nx_agraph.graphviz_layout(G, prog=layout), node_color=color_values)
else:
nx.draw(G, pos=nx.drawing.nx_agraph.graphviz_layout(G, prog=layout))
plt.show()
[docs] def get_annotations(self):
"""
Returns a dictionary of annotations for the parse tree. The keys are the node labels and the values are the annotations.
"""
annotations = {}
stack = [(self.root, None)]
while stack:
node, parent = stack.pop()
for label, value in node.get_all_functional_labels().items():
annotations[(parent, node.label)] = f"{label}={value}"
for child in node.children:
stack.append((child, node.label))
return annotations
[docs]class FStructure:
"""
This class is used to create and manipulate f-structures, which are data structures used in formal semantics to represent the meaning of a sentence.
The class contains several methods for adding, setting, getting, and removing attributes and their values, as well as functional labels.
It also has methods for adding, removing, and checking constraints on the attributes and for adding, getting, and generating sentences from f-structures.
The class also has a method for transferring the language by applying transfer rules to the f-structure.
The class also contains a list for storing other f-structures.
The class is a dict-like structure where attributes are the keys and values are the corresponding values.
"""
def __init__(self):
"""
Construct a new FStructure object.
"""
self.attributes = {}
self.constraints = []
self.f_structures = []
[docs] def add(self, attribute: str, value: str):
"""
Add a new attribute-value pair to the f-structure
:param attribute: (str) the attribute name
:param value: (str) the value of the attribute
"""
if attribute not in self.attributes:
self.attributes[attribute] = set()
self.attributes[attribute].add(value)
[docs] def set(self, attribute: str, value: str):
"""
Set the value of the specified attribute
:param attribute: (str) the attribute name
:param value: (str) the value of the attribute
"""
self.attributes[attribute] = {value}
[docs] def get(self, attribute: str):
"""
Get the value of the specified attribute
:param attribute: (str) the attribute name
:return: (str) the value of the attribute
"""
if attribute in self.attributes:
return self.attributes[attribute]
else:
return None
[docs] def remove(self, attribute: str):
"""
Remove the specified attribute from the f-structure
:param attribute: (str) the attribute name
"""
if attribute in self.attributes:
self.attributes.pop(attribute)
[docs] def get_all_attributes(self):
"""
get all attributes and their values
:return: dict of attributes and their values
"""
return self.attributes
[docs] def has_attribute(self, attribute: str):
"""
check if the f-structure has the specified attribute
:param attribute: (str) attribute name
:return: (bool) True if has the attribute, False otherwise
"""
return attribute in self.attributes
[docs] def add_functional_label(self, label: str, value: str):
"""
Add a new functional label
:param label: (str) functional label
:param value: (str) corresponding label value
"""
self.attributes[label] = value
[docs] def get_functional_label(self, label: str):
"""
get the value of the functional label specified
:param label: (str) functional label
:return: (str) corresponding label value
"""
return self.attributes.get(label)
[docs] def remove_functional_label(self, label: str):
"""
remove the specified functional label from the f-structure
:param label: (str) functional label
"""
self.attributes.pop(label, None)
[docs] def get_all_functional_labels(self):
"""
get all functional labels
:return: dict of functional labels in the XLFG standard format
"""
return self.attributes
[docs] def add_constraint(self, constraint: str):
"""
Add a new constraint to the f-structure
:param constraint: (str) the constraint to be added
"""
self.constraints.append(constraint)
[docs] def remove_constraint(self, constraint: str):
"""
Remove the specified constraint from the f-structure
:param constraint: (str) the constraint to be removed
"""
self.constraints.remove(constraint)
[docs] def check_constraints(self):
"""
Check if all the constraints are satisfied
:return: (bool) True if all constraints are satisfied, False otherwise
"""
for constraint in self.constraints:
if not constraint(self.attributes):
return False
return True
[docs] def add_f_structure(self, f_structure: dict):
"""
Add a new f-structure to the current f-structure
:param f_structure: (dict) the f-structure to be added
"""
self.f_structures.append(f_structure)
[docs] def get_f_structures(self):
"""
Get all f-structures
:return: (list) all f-structures
"""
return self.f_structures
[docs] def generate_sentence(self) -> str:
"""
Generate a sentence from the f-structure
:return: (str) the sentence
"""
sentence = ""
for attribute, value in self.attributes.items():
if attribute == "pred":
sentence += value + " "
elif attribute.startswith("arg"):
sentence += value + " "
return sentence.strip()
[docs] def transfer_language(self, transfer_rules: dict):
"""
Transfer the f-structure to a different language using the provided transfer rules.
:param transfer_rules: (Dict[str, str]) A dictionary where keys are the attributes in the original language,
and values are the corresponding attributes in the target language.
:return: (dict) the transfered attributes
"""
new_attributes = {}
for attribute, value in self.attributes.items():
if attribute in transfer_rules:
new_attribute = transfer_rules[attribute]
new_attributes[new_attribute] = value
return new_attributes