-
Notifications
You must be signed in to change notification settings - Fork 24
JSON LD
Keith Alcock edited this page Sep 28, 2018
·
75 revisions
The revised JSON-LD output format for Eidos is documented in the two tables below. An example follows to illustrate some of the notation, which just mimics JSON. [] indicates an array, {} indicates a map, and @id(Word) expands to something like "@id" : "_:Word_1".
Note: Recent changes are indicated with bold text. While many values are optional, they are written in italics where it is particularly important to keep that in mind. The example further below does not yet match the tables.
| Name | Property | Type | Description |
|---|---|---|---|
| Corpus | @type | "Corpus" | A corpus is typed. |
| documents | [Document] | It has a list of documents | |
| extractions | [Extraction] | and a set of mixed extractions. | |
| Document | @type | "Document" | A document is typed |
| @id | IRI | and provided an ID. | |
| title | string | It has a title, | |
| text | string | some text, | |
| location | string | a human-interpretable indication of where the text was found, | |
| sentences | [Sentence] | and a list of sentences. | |
| Sentence | @type | "Sentence" | A sentence is typed |
| @id | IRI | and provided an ID. | |
| text | string | It has a text, | |
| words | [Word] | a list of words, | |
| dependencies | [Dependency] | a set of universal enhanced dependencies, | |
| timexes | [TimeExpression] | and a list of time expressions. | |
| Word | @type | "Word" | A word is typed |
| @id | IRI | and provided an ID. | |
| text | string | It has a text, | |
| tag | string | a tag from the Penn Treebank tag set, | |
| entity | string | an entity type, | |
| startOffset | integer | an inclusive, 0-based index of the first letter of the word in the text, | |
| endOffset | integer | an exclusive, 0-based index of the last letter of the word, | |
| lemma | string | a lemma, | |
| chunk | string | a chunk, | |
| norm | string | and a norm, including "B-TIME" or "I-TIME" | |
| Dependency | @type | "Dependency" | A dependency is typed. |
| source | {@id(Word)} | It has a source ID referring to a Word, | |
| destination | {@id(Word)} | a destination ID referring to a Word, | |
| relation | string | and a relation. | |
| Extraction | @type | "Extraction" | An extraction is typed |
| @id | IRI | and provided an ID. | |
| type | string | a description of type from the list below, | |
| subtype | string | a description of subtype from the list below, | |
| labels | [string] | It has a list of labels, | |
| text | [string] | a text, | |
| rule | [string] | a rule, | |
| canonicalName | string | a canonical name, | |
| groundings | [Groundings] | a list of groundings(es), | |
| provenance | [Provenance] | a set of provenance values, | |
| trigger | Trigger | a trigger, | |
| states | [State] | and a set of states. | |
| arguments | [Argument] | and a list of arguments. | |
| Groundings | @type | "Groundings" | A groundings is typed. |
| name | string | It has a name such as "un", "wdi", or "fao" | |
| values | [Grounding] | and a list of grounding values. | |
| Grounding | @type | "Grounding" | A grounding is typed. |
| ontologyConcept | string | It has an ontology concept | |
| value | string | and a matching value. | |
| Provenance | @type | "Provenance" | A "provenance" is typed. |
| document | {@id(Document)} | It has a document ID referring to a Document, | |
| documentCharPositions | Interval | an interval for characters within the document, | |
| sentence | {@id(Sentence)} | a sentence ID referring to a Sentence, | |
| sentenceWordPositions | Interval | and an interval for words within the sentence. | |
| Interval | @type | "Interval" | An interval is typed. |
| start | integer | For sentenceWordPositions, this is an inclusive, 1-based index of the first word of the interval in the sentence. For documentCharPositions, it is an inclusive, 0-based index of the first character of the interval in the document | |
| end | integer | For sentenceWordPositions, this is an inclusive, 1-based index of the last word of the interval in the sentence. For documentCharPositions, it is an inclusive, 0-based index of the last character of the interval in the document. | |
| State | @type | "State" | A state is typed. |
| type | string | It has an Eidos type such as INC, DEC, QUANT, TIMEX, or HEDGE, or a BBN type such as TENSE, POLARITY, and so on. This list is non-exhaustive. See details below. | |
| text | string | a text, | |
| value | value_type | a value of type detailed below, | |
| provenance | [Provenance] | a set of provenance values, currently unused, | |
| modifiers | [Modifier] | and a set of modifiers. | |
| Modifier | @type | "Modifier" | A modifier is typed. |
| text | string | It has a text, | |
| provenance | [Provenance] | a set of provenance values, currently unused, | |
| intercept | double | an intercept, | |
| mu | double | a mu, | |
| sigma | double | and a sigma. | |
| Trigger | @type | "Trigger" | A trigger is typed. |
| text | string | It has a text | |
| text | string | Head word of the trigger. | |
| provenance | [Provenance] | and a set of provenance values. | |
| Argument | @type | "Argument" | An argument is typed. |
| type | string | It has a description of type from the table below, | |
| value | {@id(Extraction)} | and an ID referring to the extraction. | |
| TimeExpression | @type | "TimeExpression" | A time expression is typed |
| @id | IRI | and provided with an ID. | |
| startOffset | integer | It has a starting, zero-based character offset into the sentence text, | |
| endOffset | integer | and an ending, exclusive character offset, | |
| text | string | along with the actual text determined by the offsets, | |
| intervals | [TimeInterval] | and a list of time intervals. | |
| TimeInterval | @type | "TimeInterval" | A time interval is typed |
| @id | IRI | and provided with an ID. | |
| start | string | It starts (in ISO-8601 format), | |
| end | string | it ends (in ISO-8601 format), | |
| duration | integer | and it lasts for a number of seconds. |
| Extraction type | Extraction subtype | Argument type | Notes |
|---|---|---|---|
| "concept" | This is just something being there. | ||
| "entity" | none | ||
| "event" | actor | ||
| place | |||
| time | |||
| "relation" | Link between concepts | ||
| "causation" | This is a directed relation. | ||
| "source" | This source may appear multiple times. | ||
| "destination" | This destination may appear multiple times. | ||
| "precondition" | The "source" concept must have occurred/begun for the "destination" concept to happen. This is a directed relation. | ||
| "source" | This source may appear multiple times. | ||
| "destination" | This destination may appear multiple times. | ||
| "catalyst" | The "source" concept increases the intensity of the "destination" concept, but does not cause the "destination" concept. This is a directed relation. | ||
| "source" | This source may appear multiple times. | ||
| "destination" | This destination may appear multiple times. | ||
| "mitigation" | The "source" concept decreases the intensity of the "destination" concept, but does not cause the "destination" concept. This is a directed relation. | ||
| "source" | This source may appear multiple times. | ||
| "destination" | This destination may appear multiple times. | ||
| "prevention" | The "source" concept will stop or prevent the 'destination' concept. This is a directed relation. | ||
| "source" | This source may appear multiple times. | ||
| "destination" | This destination may appear multiple times. | ||
| "temporallyPrecedes" | The "source" concept occurs before the 'destination' concept. This is a directed relation. | ||
| "source" | This source may appear multiple times. | ||
| "destination" | This destination may appear multiple times. | ||
| "correlation" | This is an undirected relation. | ||
| "argument" | The argument must appear multiple times. | ||
| "unification" | This is a directed relation. | ||
| "group" | The argument must appear only once. | ||
| "member" | The argument must appear only once. | ||
| "coreference" | This is a directed relation. | ||
| "anchor" | The argument must appear only once. | ||
| "reference" | The argument must appear only once. |
| State type | Value type | Notes |
|---|---|---|
| INC | Indicates that the Concept has been increased. | |
| DEC | Indicates that the Concept has been increased. | |
| QUANT | The gradable adjective associated with the Concept. | |
| TIMEX | {@id(TimeExpression)} | This is an ID referring to a time expression. |
| HEDGE | Indicates that the Extraction has been hedged. | |
| polarity | Negative when it is explicitly indicated that the Event did not occur. All other Events are Positive. | |
| modality | Asserted when the author or speaker makes reference to it as though it were a real occurrence. All other Events are Other. | |
| genericity | Specific if it is a singular occurrence at a particular place and time, or a finite set of such occurrences. All other Events are Generic. |
This example shows valid JSON-LD syntax and links between elements, even though the linguistic analysis is fabricated.
{
"@context" : {
"Argument" : "https://w3id.org/wm/cag/Argument",
"Corpus" : "https://w3id.org/wm/cag/Corpus",
"Dependency" : "https://w3id.org/wm/cag/Dependency",
"Document" : "https://w3id.org/wm/cag/Document",
"Extraction" : "https://w3id.org/wm/cag/Extraction",
"Interval" : "https://w3id.org/wm/cag/Interval",
"Modifier" : "https://w3id.org/wm/cag/Modifier",
"Provenance" : "https://w3id.org/wm/cag/Provenance",
"Sentence" : "https://w3id.org/wm/cag/Sentence",
"State" : "https://w3id.org/wm/cag/State",
"Trigger" : "https://w3id.org/wm/cag/Trigger",
"Word" : "https://w3id.org/wm/cag/Word"
},
"@type" : "Corpus",
"documents" : [ {
"@type" : "Document",
"@id" : "_:Document_1",
"title" : "Example Document",
"sentences" : [ {
"@type" : "Sentence",
"@id" : "_:Sentence_1",
"text" : "Hello, world!",
"words" : [ {
"@type" : "Word",
"@id" : "_:Word_1",
"text" : "Hello",
"tag" : "UH",
"entity" : "O",
"startOffset": 0,
"endOffset" : 5,
"lemma" : "hello",
"chunk" : "B-ADVP"
}, {
"@type" : "Word",
"@id" : "_:Word_2",
"text" : ",",
"tag" : ",",
"entity" : "O",
"startOffset" : 5,
"endOffset" : 6,
"lemma" : ",",
"chunk" : "O"
}, {
"@type" : "Word",
"@id" : "_:Word_3",
"text" : "world",
"tag" : "NN",
"entity" : "O",
"startOffset" : 7,
"endOffset" : 12,
"lemma" : "world",
"chunk" : "B-NP"
}, {
"@type" : "Word",
"@id" : "_:Word_4",
"text" : "!",
"tag" : ".",
"entity" : "O",
"startOffset" : 12,
"endOffset" : 13,
"lemma" : "!",
"chunk" : "O"
} ],
"dependencies" : [ {
"@type" : "Dependency",
"source" : {
"@id" : "_:Word_3"
},
"destination" : {
"@id" : "_:Word_1"
},
"relation" : "discourse"
}, {
"@type" : "Dependency",
"source" : {
"@id" : "_:Word_3"
},
"destination" : {
"@id" : "_:Word_2"
},
"relation" : "punct"
} ]
} ]
} ],
"extractions" : [ {
"@type" : "Extraction",
"@id" : "_:Extraction_1",
"type" : "concept",
"subtype" : "entity",
"labels" : [ "NounPhrase", "Entity" ],
"text" : "world",
"rule" : "simple-np",
"canonicalName" : "world",
"grounding" : [ {
"@type" : "Grounding",
"ontologyConcept" : "/entities/human/livelihood",
"value" : 0.47524851930210044
}, {
"@type" : "Grounding",
"ontologyConcept" : "/entities/human/financial/economic/economy",
"value" : 0.4713680118187502
} ],
"provenance" : [ {
"@type" : "Provenance",
"document" : {
"@id" : "_:Document_1"
},
"sentence" : {
"@id" : "_:Sentence_1"
},
"positions" : {
"@type" : "Interval",
"start" : 3,
"end" : 3
}
} ],
"states" : [ {
"@type" : "State",
"type" : "INC",
"text" : "Hello",
"modifiers" : [ {
"@type" : "Modifier",
"text" : "world",
"intercept" : 0.6154,
"mu" : 1.034E-5,
"sigma" : -0.001123
} ]
} ]
}, {
"@type" : "Extraction",
"@id" : "_:Extraction_2",
"type" : "relation",
"subtype" : "causation",
"labels" : [ "Causal", "DirectedRelation", "EntityLinker", "Event" ],
"text" : "Hello",
"rule" : "dueToSyntax1-Causal",
"canonicalName" : "hello",
"provenance" : [ {
"@type" : "Provenance",
"document" : {
"@id" : "_:Document_1"
},
"sentence" : {
"@id" : "_:Sentence_1"
},
"positions" : {
"@type" : "Interval",
"start" : 1,
"end" : 1
}
} ],
"trigger" : {
"@type" : "Trigger",
"text" : "world",
"provenance" : [ {
"@type" : "Provenance",
"document" : {
"@id" : "_:Document_1"
},
"sentence" : {
"@id" : "_:Sentence_1"
},
"positions" : {
"@type" : "Interval",
"start" :3,
"end" : 3
}
} ]
},
"arguments" : [ {
"@type" : "Argument",
"type" : "source",
"value" : {
"@id" : "_:Extraction_1"
}
}, {
"@type" : "Argument",
"type" : "destination",
"value" : {
"@id" : "_:Extraction_3"
}
} ]
}, {
"@type" : "Extraction",
"@id" : "_:Extraction_3",
"type" : "relation",
"subtype" : "correlation",
"labels" : [ "Correlation", "UndirectedRelation", "EntityLinker", "Event" ],
"text" : "world",
"rule" : "dueToSyntax1-Causal",
"canonicalName" : "world",
"provenance" : [ {
"@type" : "Provenance",
"document" : {
"@id" : "_:Document_1"
},
"sentence" : {
"@id" : "_:Sentence_1"
},
"positions" : {
"@type" : "Interval",
"start" : 3,
"end" : 3
}
} ],
"trigger" : {
"@type" : "Trigger",
"text" : "Hello",
"provenance" : [ {
"@type" : "Provenance",
"document" : {
"@id" : "_:Document_1"
},
"sentence" : {
"@id" : "_:Sentence_1"
},
"positions" : {
"@type" : "Interval",
"start" : 1,
"end" : 1
}
} ]
},
"arguments" : [ {
"@type" : "Argument",
"type" : "argument",
"value" : {
"@id" : "_:Extraction_2"
}
}, {
"@type" : "Argument",
"type" : "argument",
"value" : {
"@id" : "_:Extraction_1"
}
} ]
} ]
}