Natural Language Processing
A.V.A.T.A.R incorporates a modest but functional Natural Language Processing (NLP).
This step is automatic and internal to A.V.A.TA.R. It is called before checking the intention file of all plugins.
Example
If the sentence is :
Make a 1 hour appointment with the dentist tomorrow at 2:00 p.m. in Paris
A.V.A.T.A.R will analyze the sentence and create a relationship scenario, then suggest :
- Action:
make - Object:
appointment - Duration:
1 hour - Item:
dentist - When:
June 25th 2024 2:00 pm - Location:
paris
If the intention is verified, the data object accessible in the plugin script will receive the relationship scenario that NLP has built, which you can then use to, for example, create an appointment in your agenda.
Set an appointment in phone's calendar like: ${Object} ${Item} in ${Location} on ${When} of ${Duration}
Objects
Below are the objects created by NLP that you can use in the data object parameter of the action method of the script file.
data.tokens - Array
Array of words of the sentence.
data.tokens: [
'make', 'a',
'1', 'hour',
'appointment', 'with',
'the', 'dentist',
'tomorrow', 'at',
'2:00', 'pm',
'in', 'paris'
]
data.tags - Array
Array of word types of the sentence.
tags: [
'Imperative', 'Determiner',
'Date', 'Singular',
'Noun', 'Preposition',
'Determiner', 'Singular',
'Date', 'Time',
'Time', 'Time',
'Preposition', 'City'
]
You can consult the complete list of possible tags here
Warning
As tags are the types of each word, the tokens and tags arrays must have identical lengths.
If this is not the case, the intention is not resolved.
This is usually due to a compound word, which must then be correctly written into the collection of terms to correct the problem.
data.relations - Object
The relationship scenario built by NLP. This scenario can then be used to facilitate development.
relations: {
action: {
verb: {
tense: 'verb',
negative: false
},
tag: 'verb',
text: 'make'
},
duration: {
tag: 'duration',
text: '1 hour'
},
object: {
tag: 'noun',
text: 'appointment'
},
item: {
tag: 'noun',
text: 'dentist'
},
when: {
info: {
start: '2024-06-25T14:00:00.000+02:00',
timezone: 'Europe/Paris'
},
tag: 'date',
text: 'June 25th 2024 2:00 pm'
},
location: {
tag: 'city',
text: 'paris'
}
}