Welcome to CoreNLP XML Library’s documentation!¶

This library is designed to add a data model over Stanford CoreNLP’s basic XML output.

The Document class is designed to provide lazy-loaded access to information from syntax, coreference, and dependency parse structures within the XML.

Installing the Library¶

It’s as easy as

pip install corenlp_xml

What You Can Do With This Library¶

Some code examples:

from corenlp_xml import Document

doc = Document(xml_string)

# The first sentence
s1 = doc.sentences[0]

# Noun phrases for the first sentence
s1_nps = s1.phrase_strings("np")

# Text of semantic head of first sentence
s1_head = s1.semantic_head.text

# Find all representative coreferences matching noun phrases in sentence 1
s1_corefs = [coref for coref in doc.coreferences
             if coref.representative and coref.sentence == s1]

Contents:

The Document Model¶

Sub-module for handling document-level stuff

class corenlp_xml.document.Document(xml_string)[source]¶

This class abstracts a Stanford CoreNLP Document

coreferences[source]¶

Returns a list of Coreference classes

Getter:	Returns a list of coreferences
Type:	list of corenlp_xml.coreference.Coreference

get_sentence_by_id(id)[source]¶

Gets sentence by ID

Parameters:	id (int) – the ID of the sentence, as defined in the XML
Returns:	a sentence
Return type:	corenlp_xml.document.Sentence

sentences[source]¶

Returns the ordered dict of sentences as a list.

Getter:	returns list of sentences, in order
Type:	list of corenlp_xml.document.Sentence

sentiment[source]¶

Returns average sentiment of document. Must have sentiment enabled in XML output.

Getter:	returns average sentiment of the document
Type:	float

class corenlp_xml.document.Sentence(element)[source]¶

This abstracts a sentence

basic_dependencies[source]¶

Accesses basic dependencies from the XML output

Getter:	Returns the dependency graph for basic dependencies
Type:	corenlp_xml.dependencies.DependencyGraph

collapsed_ccprocessed_dependencies[source]¶

Accesses collapsed, CC-processed dependencies

Getter:	Returns the dependency graph for collapsed and cc processed dependencies
Type:	corenlp_xml.dependencies.DependencyGraph

collapsed_dependencies[source]¶

Accessess collapsed dependencies for this sentence

Getter:	Returns the dependency graph for collapsed dependencies
Type:	corenlp_xml.dependencies.DependencyGraph

get_token_by_id(id)[source]¶

Accesses token by the XML ID

Parameters:	id (int) – The XML ID of the token
Returns:	The token
Return type:	corenlp_xml.document.Token

id[source]¶

Returns:	the ID attribute of the sentence
Return type:	int

parse[source]¶

Accesses the parse tree based on the S-expression parse string in the XML

Getter:	Returns the NLTK parse tree
Type:	nltk.Tree

parse_string[source]¶

Accesses the S-Expression parse string stored on the XML document

Getter:	Returns the parse string
Type:	str

phrase_strings(phrase_type)[source]¶

Returns strings corresponding all phrases matching a given phrase type

Parameters:	phrase_type (str) – POS such as “NP”, “VP”, “det”, etc.
Returns:	a list of strings representing those phrases

semantic_head[source]¶

Returns the semantic head of the sentence – AKA the dependent of the root node of the dependency parse

Returns:	the mention related to the semantic head
Return type:	corenlp_xml.coreference.Mention

sentiment[source]¶

The sentiment of this sentence

Getter:	Returns the sentiment value of this sentence
Type:	int

subtrees_for_phrase(phrase_type)[source]¶

Returns subtrees corresponding all phrases matching a given phrase type

Parameters:	phrase_type (str) – POS such as “NP”, “VP”, “det”, etc.
Returns:	a list of NLTK.Tree.Subtree instances
Return type:	list of NLTK.Tree.Subtree

tokens[source]¶

The tokens related to this sentence

Getter:	Returns a a list of Token instances
Type:	corenlp_xml.document.TokenList

class corenlp_xml.document.Token(element)[source]¶

Wraps the token XML element

character_offset_begin[source]¶

Lazy-loads character offset begin node

Getter:	Returns the integer value of the beginning offset
Type:	int

character_offset_end[source]¶

Lazy-loads character offset end node

Getter:	Returns the integer value of the ending offset
Type:	int

id[source]¶

Lazy-loads ID

Getter:	Returns the ID of the token element
Type:	int

lemma[source]¶

Lazy-loads the lemma for this word

Getter:	Returns the plain string value of the word lemma
Type:	str

ner[source]¶

Lazy-loads the NER for this word

Getter:	Returns the plain string value of the NER tag for the word
Type:	str

pos[source]¶

Lazy-loads the part of speech tag for this word

Getter:	Returns the plain string value of the POS tag for the word
Type:	str

speaker[source]¶

Lazy-loads the speaker for this word

Getter:	Returns the plain string value of the speaker tag for the word
Type:	str

word[source]¶

Lazy-loads word value

Getter:	Returns the plain string value of the word
Type:	str

Interacting with Dependencies¶

This component is responsible for managing dependency parses

class corenlp_xml.dependencies.DependencyGraph(element)[source]¶

Dependency graph, models a dependency parse

get_node_by_idx(idx)[source]¶

Stores each distinct node in a dict

Parameters:	idx (int) – the “idx” value of the node
Returns:	the node instance for that index
Type:	corenlp_xml.dependencies.DependencyNode

links[source]¶

Accesses links within the graph

Returns:	a list of corenlp_xml.dependencies.DependencyLink instances
Type:	list of corenlp_xml.dependencies.DependencyLink

links_by_type(dep_type)[source]¶

Accesses links within the graph

Parameters:	dep_type (str) – the depency type
Returns:	a list of corenlp_xml.dependencies.DependencyLink instances
Type:	list of corenlp_xml.dependencies.DependencyLink

class corenlp_xml.dependencies.DependencyLink(graph, element)[source]¶

Represents a relationship between two nodes in a dependency graph

dependent[source]¶

Accesses the dependent node

Getter:	returns the Dependent node
Type:	corenlp_xml.dependencies.DependencyNode

governor[source]¶

Accesses the governor node

Getter:	Returns the Governor node
Type:	corenlp_xml.dependencies.DependencyNode

class corenlp_xml.dependencies.DependencyNode(graph, element)[source]¶

Represents a node in a dependency graph

dependent(dep_type, node)[source]¶

Registers a node as dependent on this node

Parameters:	dep_type (str) – The dependency type node (corenlp_xml.dependencies.DependencyNode) – The node to be registered as a dependent
Returns:	self, provides fluent interface
Return type:	corenlp_xml.dependencies.DependencyNode

dependents[source]¶

Gets dependent nodes

Getter:	returns a flat list of all governing nodes
Type:	list of corenlp_xml.dependencies.DependencyNode

dependents_by_type(dep_type)[source]¶

Gets the dependents of this node by a given dependency type

Parameters:	dep_type (str) – The dependency type
Returns:	dependents matching the provided type

governor(dep_type, node)[source]¶

Registers a node as governing this node

Parameters:	dep_type (str) – The dependency type node –
Returns:	self, provides fluent interface
Return type:	corenlp_xml.dependencies.DependencyNode

governors[source]¶

Gets governing nodes

Getter:	returns a flat list of all governing nodes
Type:	list of corenlp_xml.dependencies.DependencyNode

governors_by_type(dep_type)[source]¶

Gets the governors of this node filtered by a dependency type

Parameters:	dep_type (str) – The dependency type
Returns:	governors matching the provided type

classmethod load(graph, element)[source]¶

Instantiates the node in the graph if it’s not already stored in the graph

Parameters:	graph (corenlp_xml.dependencies.DependencyGraph) – The dependency graph this node is a member of element (lxml.ElementBase) – The lxml element wrapping the node

text = None¶: These properties are dicts of link type to node

Coreference Resolution¶

This library is responsible for handling coreference resolution parsing from the XML output

class corenlp_xml.coreference.Coreference(document, element)[source]¶

Reflects a grouping of mentions

mentions[source]¶

Returns mentions

Returns:	list of mentions
Return type:	list

representative[source]¶

Representative mention

Returns:	representative Mention
Return type:	corenlp_xml.coreference.Mention

class corenlp_xml.coreference.Mention(coref, element)[source]¶

Reflects a given mention

head[source]¶

The token serving as the “head” of the mention

Getter:	the token corresponding to the head
Type:	corenlp_xml.document.Token

representative[source]¶

Interprets and normalizes the “representative” attribute”

Getter:	determines whether the mention is representative
Type:	bool

sentence[source]¶

The sentence related to this mention

Getter:	returns the sentence this mention relates to
Type:	corenlp_xml.document.Sentence

siblings[source]¶

Accesses other mentions in this coref group

Getter:	the other mentions for this coref group
Type:	list of corenlp_xml.coreference.Mention