Manual association of entities
- AlexI (Unlicensed)
- andrew johnston (Unlicensed)
Overview
This toolkit element enables the generation of one of more types of association between existing entities based on the document or content metadata. The expressions default to replacement strings, or $SCRIPT(...) can be used to return a string using javascript.
Format
{
"display": string,
"associations": [
{
"iterateOver":string, // OPTIONAL: If specified as a list of entity types, steps over entities with matching types (again, lock step or combinatorially)
// Can also specify a metadata field (nesting supported using dot notation), in which case they are looped over to generate calls with _value/_iterator/_index
"entity1":string, // OPTIONAL: String/script: In "iterateOver"/type cases, the disambiguated name of the entity type; otherwise using entity1_index is preferred.
"entity1_index":string, // OPTIONAL: String/script: should return the 'disambiguated_name/type' string, must resolve to an entity or is discarded
"entity2":string, // OPTIONAL: String/script: In "iterateOver"/type cases, the disambiguated name of the entity type; otherwise using entity1_index is preferred.
"entity2_index":string, // OPTIONAL: String/script: should return the 'disambiguated_name/type' string, must resolve to an entity or is discarded
"verb":string, // MANDATORY: String/script
"verb_category":string, // MANDATORY: String/script
"assoc_type":string, // MANDATORY: Must specify/return one of "Fact", "Event", "Summary" and be overridden (eg converted to summary if there is only 1 index)
// (if left blank, the Structured Analysis Handler will auto-generate this field reasonably accurately based on the contents)
"time_start":string, // OPTIONAL: String/script: Must specify/return a time in ISO date format ("yyyy-MM-dd'T'HH:mm:ss") or Javascript time format
"time_end":string, // OPTIONAL: String/script: Must specify/return a time in ISO date format ("yyyy-MM-dd'T'HH:mm:ss") or Javascript time format
"geo_index":string, // OPTIONAL: String/script: The entity index corresponding to the "geotag" below (or the Type in "iterateOver" cases)
"geotag": { // OPTIONAL: Format is identical to the docGeo format specified above
"lat":string, "lon":string,
"city":string, "stateProvince":string, "country":string, "countryCode":string
},
//(note the ontology_type for associations is always "point" - use geo_index to specify larger areas)
"creationCriteriaScript":string, // // OPTIONAL: script: If populated, runs a user script function and if return value is false doesn't create the object
}
]
}
Description
Associations can be "something that happens or is regarded as happening; an occurrence, especially one of some importance", "the outcome, issue, or result of anything", or "something that occurs in a certain place during a particular interval of time" (Definitions found here: http://dictionary.reference.com/browse/event). Within Infinit.e events are typically a combination of entities assembled in the form of Noun - Verb - Noun, e.g. "a car crashed into a building", "the plane flew to San Diego". In addition to the Noun - Verb - Noun form events can include geographic information (i.e., where an event happened) as well as a start and/or end time for an event.
| Field | Description | Data Type |
|---|---|---|
iterateOver | ||
entity | A free form text field containing information about the event "subject", i.e. an entity's disambiguous name.
| |
entity1_index | If present this is the "index" field of the entity matching the entity1 disambiguous name above. | |
verb | A free form text field describing the event "verb" | |
verb_category | Also a free form text field describing the event "verb", but intended to group related verbs together (eg "travel" for verbs: "flew", "drove") |
IN PROGRESS
iterateOver
iterateOver can be used in more advanced cases.
Iterate Over a Metadata Field:
Specify a metadata field, in order to generate associations.
eg. "iterateOver": "json.twitter_entities.user_mentions"
Metadata field nesting is also supported using dot notation, in which case they are looped over to generate calls with _value/_iterator/_index
This is the same as iterating over a metadata array to obtain entities using Manual entities.
For more information concerning metadata fields and support for nested fields, see section Manual entities.
See detailed example below
Iterate Over a Single Entity Type:
A less common scenario is when you want to iterate over a single entity type. You can do this by using "dummy."
eg "entity1/dummy" or "entity2,dummy"
See detailed example below
Multiplicative:
Create one association for every combination of entities of specified types. Also referred to as combinatorial association.
eg. "iterateOver": "entity1/entity2/geo_index",
See detailed example below
Associative:
Create one association for every pair (/set) of entities of specified types. Also referred to as lock-step or additive association.
eg. "iterateOver": "entity1,entity2"
See detailed example below
creationCriteriaScripts
Association fields generated from the entity loop are placed in "_iterator". For example, for "iterateOver": "entity1/entity2/geo_index", an _iterator object with the following fields is available in the Javascript: "_iterator.entity1_index", "_iterator.entity2_index", "_iterator.geo_index".
These fields can be usefully used together with "creationCriteriaScript" scriptlets to filter out unwanted associations, eg when looping over entity1 and entity2 with the same entity type, the following script would ensure the association didn't involve the same entity:
"creationCriteriaScript": "$SCRIPT( return _iterator.entity1_index != _iterator.entity2_index; )", "iterateOver": "entity1/entity2", "entity1": "EmailAddress", "entity2": "EmailAddress", //etc
The creationCriteriaScript runs before the association is generated (so can be safely used to remove items that would return errors).
See detailed example below
Examples
Basic Association
In the basic association example, the code specifies an entity from the document metadata. The association specifies the relationship between the type of offense and the date and time at which it was reported.
{
"associations": [
{
"entity1": "$metadata.offense,$metadata.method",
"geotag": {},
"time_start": "$metadata.reportdatetime",
"verb": "reported",
"verb_category": "crime"
}
]
}
Sample Output:
The resulting output displays the details of the date and time, and the specific type of offense reported.
"associations": [
{
"entity1": "theft,2",
"entity1_index": "theft,2/criminalactivity",
"verb": "reported",
"verb_category": "crime",
"time_start": "2011-01-29T00:00:00",
"geotag": {
"lat": 38.9099278028729,
"lon": -77.0436067765966
},
"assoc_type": "Summary"
}
iterateOver
iterateOver a Metadata Field
The following code example is used to process email communication between two parties. The association between the two entities is formed by iterating over the metadata object: "email_meta.Message-To"
{
"associations": [
{
"assoc_type": "Event",
"entity1": "$SCRIPT( return _doc.metadata._FILE_METADATA_[0].metadata.Author[0];)",
"entity2": "$SCRIPT(return _value;)",
"iterateOver": "email_meta.Message-To",
"time_start": "$SCRIPT( return _doc.publishedDate;)",
"verb": "emailed",
"verb_category": "emailed/communicated"
}
]
}
Metadata:
The email metadata "Message-To" used to generate the associations is included in the source block below.
],
"email_meta": [
[
{
"Creation-Date": [
"2001-07-09T18:33:32Z"
],
"Message-To": [
"will.smith@enron.com"
],
"Content-Type": [
"message/rfc822"
],
"subject": [
"RE: Testing Preschedule workspace"
],
"date": [
"2001-07-09T18:33:32Z"
],
"Author": [
"cara.semperger@enron.com"
],
"Message-From": [
"cara.semperger@enron.com"
]
}
]
]
Output:
The output displays the association between the sender and receiver in the email correspondence.
"associations": [
{
"entity1": "cara.semperger@enron.com",
"entity1_index": "cara.semperger@enron.com/account",
"verb": "emailed",
"verb_category": "emailed/communicated",
"entity2": "will.smith@enron.com",
"entity2_index": "will.smith@enron.com/account",
"time_start": "2001-07-09T14:33:32",
"assoc_type": "Event"
}
],
iterateOver a single entity type
In some cases you will only want to iterate over a single entity type, rather than two or three, which is more common. For these cases you can use "dummy" for iterateOver.
In the following source, iterateOver is setup to only iterate over the entity type: "VictimType."
{
"assoc_type": "Event",
"entity1": "$SCRIPT( return _doc.metadata.eventtype[0]; )",
"entity2": "VictimType",
"iterateOver": "entity2/dummy",
"verb_category": "suffered"
},
The code acts on the following entities in this example
"entities": [
{
"actual_name": "Targeted, Paramilitary/Private Security, Adult, Tribal from Pakistan",
"dimension": "Who",
"disambiguated_name": "Targeted, Paramilitary/Private Security, Adult, Tribal from Pakistan",
"doccount": 0,
"frequency": 3,
"index": "targeted, paramilitary/private security, adult, tribal from pakistan/victimtype",
"relevance": 0,
"totalfrequency": -1,
"type": "VictimType"
},
{
"actual_name": "Targeted, Non Official Public Figure, Adult, Tribal from Pakistan",
"dimension": "Who",
"disambiguated_name": "Targeted, Non Official Public Figure, Adult, Tribal from Pakistan",
"doccount": 0,
"frequency": 1,
"index": "targeted, non official public figure, adult, tribal from pakistan/victimtype",
"relevance": 0,
"totalfrequency": -1,
"type": "VictimType"
},
{
"actual_name": "Targeted, Civilian, Adult, Tribal from Pakistan",
"dimension": "Who",
"disambiguated_name": "Targeted, Civilian, Adult, Tribal from Pakistan",
"doccount": 0,
"frequency": 2,
"index": "targeted, civilian, adult, tribal from pakistan/victimtype",
"relevance": 0,
"totalfrequency": -1,
"type": "VictimType"
},
Output:
In the example output you can see that associations are only created based on entities of entity type: "VictimType." Three associations are created, for each of the Victim Type entities specified above.
{
"associations": [
{
"assoc_type": "Event",
"entity1": "Armed Attack",
"entity1_index": "armed attack/eventtype",
"entity2": "targeted, paramilitary/private security, adult, tribal from pakistan",
"entity2_index": "targeted, paramilitary/private security, adult, tribal from pakistan/victimtype",
"verb": "suffered",
"verb_category": "suffered"
},
{
"assoc_type": "Event",
"entity1": "Armed Attack",
"entity1_index": "armed attack/eventtype",
"entity2": "targeted, non official public figure, adult, tribal from pakistan",
"entity2_index": "targeted, non official public figure, adult, tribal from pakistan/victimtype",
"verb": "suffered",
"verb_category": "suffered"
},
{
"assoc_type": "Event",
"entity1": "Armed Attack",
"entity1_index": "armed attack/eventtype",
"entity2": "targeted, civilian, adult, tribal from pakistan",
"entity2_index": "targeted, civilian, adult, tribal from pakistan/victimtype",
"verb": "suffered",
"verb_category": "suffered"
},
Multiplicative
Multiplicative association are associations that are created by "multiplying" a combination of entities, locations, and times together to determine the number of associations to extract from the source data.
In the example, a perpetrator (Sunni Islamic Extremist) attacked multiple types of victims (an adult and a child) in Sri Lanka.
The association specification uses the multiplicative format to create events using the following math to determine the total number of associations: Entity1 (Person Perpetrator) * Entity2 (Victim Type) * Geo_index (Location) = Total Number of Associations.
{
"creationCriteriaScript": "$FUNC( isOrganizationSpecified(); )",
"entity1": "Organization",
"entity2": "VictimType",
"iterateOver": "entity1/entity2",
"time_start": "$SCRIPT( return _doc.metadata.incidentdate[0]; )",
"verb": "attacked",
"verb_category": "assault/attack"
},
On the following set of entities:
"entities": [
{
"actual_name": "Taliban",
"dimension": "Who",
"disambiguated_name": "Taliban",
"doccount": 0,
"frequency": 1,
"index": "taliban/organization",
"relevance": 0,
"totalfrequency": -1,
"type": "Organization"
},
{
"actual_name": "Targeted, Civilian, Adult from Afghanistan",
"dimension": "Who",
"disambiguated_name": "Targeted, Civilian, Adult from Afghanistan",
"doccount": 0,
"frequency": 1,
"index": "targeted, civilian, adult from afghanistan/victimtype",
"relevance": 0,
"totalfrequency": -1,
"type": "VictimType"
},
Sample Output:
{
"associations": [
{
"assoc_type": "Event",
"entity1": "taliban",
"entity1_index": "taliban/organization",
"entity2": "targeted, civilian, adult from afghanistan",
"entity2_index": "targeted, civilian, adult from afghanistan/victimtype",
"time_start": "2009-07-01T00:00:00",
"verb": "attacked",
"verb_category": "assault/attack"
}
Associative
Additive associations cover the less common case where (eg) 2 entity types have the same number of elements and are ordered "in lock step". For example:
"entities": [
{
"index": "alex/person",
"index": "craig/person",
"index": "baltimore/city",
"index": "washington dc/city",
}
]
In this case the additive association specification:
{
"iterateOver": "entity1,entity2", // note "," instead of "/"
"entity1": "Person",
"entity2": "City",
"verb_category": "lives in",
}
Would generate the 2 associations depicted in the sample output below
],
"associations": [
{
"entity1": "Alex/person",
"verb": "lives in",
"verb_category": "residence",
"entity2": "Baltimore/city",
"assoc_type": "Fact"
},
{
"entity1": "Craig/person",
"verb": "lives in",
"verb_category": "residence",
"entity2": "Washington dc/city",
"assoc_type": "Fact"
}
],
creationCriteriaScript
The associations code block below creates associations based on a twitter user(actor).
The creationCriteriaScript establishes conditions for the creation of associations. For objects returned that do not match, associations are not created.
},
{
"associations": [
{
"assoc_type": "Event",
"creationCriteriaScript": "$SCRIPT( return (null != _doc.metadata.json[0].object.actor); )",
"entity1_index": "$SCRIPT( return _doc.metadata.json[0].actor.preferredUsername + '/twitterhandle';)",
"entity2_index": "$SCRIPT( return _doc.metadata.json[0].object.actor.preferredUsername + '/twitterhandle';)",
"verb": "retweets",
"verb_category": "retweets"
},
{
"assoc_type": "Event",
"creationCriteriaScript": "$SCRIPT( return (null != _doc.metadata.json[0].object.actor) && (null != _doc.metadata.json[0].object.actor.location); )",
"entity1_index": "$SCRIPT( return _doc.metadata.json[0].object.actor.preferredUsername + '/twitterhandle';)",
"entity2_index": "$SCRIPT( return _doc.metadata.json[0].object.actor.location.displayName+ '/location';)",
"verb": "twitter_location",
"verb_category": "twitter_location"
},
{
"assoc_type": "Event",
"entity1_index": "$SCRIPT( return _doc.metadata.json[0].actor.preferredUsername + '/twitterhandle';)",
"entity2_index": "$SCRIPT( return _iterator.text + '/hashtag'; )",
"iterateOver": "json.twitter_entities.hashtags",
"verb": "tweets_about",
"verb_category": "tweets_about"
},
{
"assoc_type": "Event",
"entity1_index": "$SCRIPT( return _doc.metadata.json[0].actor.preferredUsername + '/twitterhandle';)",
"entity2_index": "$SCRIPT( return _iterator.screen_name + '/twitterhandle'; )",
"iterateOver": "json.twitter_entities.user_mentions",
"verb": "tweets_to",
"verb_category": "tweets_to"
},
{
"assoc_type": "Event",
"entity1_index": "$SCRIPT( return _doc.metadata.json[0].actor.preferredUsername + '/twitterhandle';)",
"entity2_index": "$SCRIPT( return _iterator.expanded_url + '/url'; )",
"iterateOver": "json.gnip.urls",
"verb": "tweets_link",
"verb_category": "tweets_link"
}
]
},
Output:
The following associations are output, based on the entity types and the creationcriteriaScript.
{
"associations": [
{
"assoc_type": "Event",
"entity1": "focalcrm",
"entity1_index": "focalcrm/twitterhandle",
"entity2": "crm",
"entity2_index": "crm/hashtag",
"verb": "tweets_about",
"verb_category": "tweets_about"
},
{
"assoc_type": "Event",
"entity1": "focalcrm",
"entity1_index": "focalcrm/twitterhandle",
"entity2": "http://www.crmbuyer.com/rsstory/76578.html",
"entity2_index": "http://www.crmbuyer.com/rsstory/76578.html/url",
"verb": "tweets_link",
"verb_category": "tweets_link"
}
Footnotes:
Legacy documentation:
Legacy documentation:
- StructuredAnalysis object (under "Associations")