edu.stanford.nlp.process

Class TSVUtils

java.lang.Object

edu.stanford.nlp.process.TSVUtils

public class TSVUtils
extends java.lang.Object

A set of utilities for parsing TSV files into CoreMaps

Author:

Gabor Angeli

Method Summary

Modifier and Type	Method and Description
static java.util.List<java.lang.String>	parseArray(java.lang.String array) Parse an SQL array.
static SemanticGraph	parseJsonTree(java.lang.String jsonString, java.util.List<CoreLabel> tokens) Parse a JSON formatted tree into a SemanticGraph.
static Annotation	parseSentence(java.util.Optional<java.lang.String> docid, java.util.Optional<java.lang.String> sentenceIndex, java.lang.String gloss, java.lang.String dependencies, java.lang.String maltDependencies, java.lang.String words, java.lang.String lemmas, java.lang.String posTags, java.lang.String nerTags, java.util.Optional<java.lang.String> sentenceid) Create an Annotation object (with a single sentence) from the given specification, as Postgres would output them
static SemanticGraph	parseTree(java.lang.String conll, java.util.List<CoreLabel> tokens) Parse a CoNLL formatted tree into a SemanticGraph.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

parseArray

public static java.util.List<java.lang.String> parseArray(java.lang.String array)

Parse an SQL array. This code allows fixing "doubly escaped" quotes, but would fail if you actually wanted \\\\ to become two backslashes in a string (see the tests). It also doesn't support single quote strings in SQL, since our output/tests write single quotes without escaping or quoting, but, really you should for SQL, as I understand things.... And it has special support for double double quoting an entire string ... but this disables having empty strings. I think that output comes from incorrectly not undoubling quotes in a String at an earlier stage, and so it should be fixed earlier.

Parameters:

array - The array to parse.

Returns:

The parsed array, as a list.

parseTree

public static SemanticGraph parseTree(java.lang.String conll,
                                      java.util.List<CoreLabel> tokens)

Parse a CoNLL formatted tree into a SemanticGraph.

Parameters:

conll - The CoNLL tree to parse.

tokens - The tokens of the sentence, to form the backing labels of the tree.

Returns:

A semantic graph of the sentence, according to the given tree.

parseJsonTree

public static SemanticGraph parseJsonTree(java.lang.String jsonString,
                                          java.util.List<CoreLabel> tokens)

Parse a JSON formatted tree into a SemanticGraph.

Parameters:

jsonString - The JSON string tree to parse, e.g: "[{\"\"dependent\"\": 7, \"\"dep\"\": \"\"root\"\", \"\"governorgloss\"\": \"\"root\"\", \"\"governor\"\": 0, \"\"dependentgloss\"\": \"\"sport\"\"}, {\"\"dependent\"\": 1, \"\"dep\"\": \"\"nsubj\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"chess\"\"}, {\"\"dependent\"\": 2, \"\"dep\"\": \"\"cop\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"is\"\"}, {\"\"dependent\"\": 3, \"\"dep\"\": \"\"neg\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"not\"\"}, {\"\"dependent\"\": 4, \"\"dep\"\": \"\"det\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"a\"\"}, {\"\"dependent\"\": 5, \"\"dep\"\": \"\"advmod\"\", \"\"governorgloss\"\": \"\"physical\"\", \"\"governor\"\": 6, \"\"dependentgloss\"\": \"\"predominantly\"\"}, {\"\"dependent\"\": 6, \"\"dep\"\": \"\"amod\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"physical\"\"}, {\"\"dependent\"\": 9, \"\"dep\"\": \"\"advmod\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"yet\"\"}, {\"\"dependent\"\": 10, \"\"dep\"\": \"\"nsubj\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"neither\"\"}, {\"\"dependent\"\": 11, \"\"dep\"\": \"\"cop\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"are\"\"}, {\"\"dependent\"\": 12, \"\"dep\"\": \"\"parataxis\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"shooting\"\"}, {\"\"dependent\"\": 13, \"\"dep\"\": \"\"cc\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"and\"\"}, {\"\"dependent\"\": 14, \"\"dep\"\": \"\"parataxis\"\", \"\"governorgloss\"\": \"\"sport\"\", \"\"governor\"\": 7, \"\"dependentgloss\"\": \"\"curling\"\"}, {\"\"dependent\"\": 14, \"\"dep\"\": \"\"conj:and\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"curling\"\"}, {\"\"dependent\"\": 16, \"\"dep\"\": \"\"nsubjpass\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"which\"\"}, {\"\"dependent\"\": 18, \"\"dep\"\": \"\"case\"\", \"\"governorgloss\"\": \"\"fact\"\", \"\"governor\"\": 19, \"\"dependentgloss\"\": \"\"in\"\"}, {\"\"dependent\"\": 19, \"\"dep\"\": \"\"nmod:in\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"fact\"\"}, {\"\"dependent\"\": 21, \"\"dep\"\": \"\"aux\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"has\"\"}, {\"\"dependent\"\": 22, \"\"dep\"\": \"\"auxpass\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"been\"\"}, {\"\"dependent\"\": 23, \"\"dep\"\": \"\"dep\"\", \"\"governorgloss\"\": \"\"shooting\"\", \"\"governor\"\": 12, \"\"dependentgloss\"\": \"\"nicknamed\"\"}, {\"\"dependent\"\": 25, \"\"dep\"\": \"\"dobj\"\", \"\"governorgloss\"\": \"\"nicknamed\"\", \"\"governor\"\": 23, \"\"dependentgloss\"\": \"\"chess\"\"}, {\"\"dependent\"\": 26, \"\"dep\"\": \"\"case\"\", \"\"governorgloss\"\": \"\"ice\"\", \"\"governor\"\": 27, \"\"dependentgloss\"\": \"\"on\"\"}, {\"\"dependent\"\": 27, \"\"dep\"\": \"\"nmod:on\"\", \"\"governorgloss\"\": \"\"chess\"\", \"\"governor\"\": 25, \"\"dependentgloss\"\": \"\"ice\"\"}, {\"\"dependent\"\": 29, \"\"dep\"\": \"\"amod\"\", \"\"governorgloss\"\": \"\"chess\"\", \"\"governor\"\": 25, \"\"dependentgloss\"\": \"\"5\"\"}]");

tokens - The tokens of the sentence, to form the backing labels of the tree.

Returns:

A semantic graph of the sentence, according to the given tree.

parseSentence

public static Annotation parseSentence(java.util.Optional<java.lang.String> docid,
                                       java.util.Optional<java.lang.String> sentenceIndex,
                                       java.lang.String gloss,
                                       java.lang.String dependencies,
                                       java.lang.String maltDependencies,
                                       java.lang.String words,
                                       java.lang.String lemmas,
                                       java.lang.String posTags,
                                       java.lang.String nerTags,
                                       java.util.Optional<java.lang.String> sentenceid)

Create an Annotation object (with a single sentence) from the given specification, as Postgres would output them