Article
· May 28 2m read

Multilingual GenerativeAI Pattern Match WorkBench Utility

Audience

Those curious in exploring new GenerativeAI usecases.

Developers and analysts looking for a quick way to tame the Pattern Match operator.

In both ObjectScript and SQL this has a quite visually dense format.

 if booking?2A1"-"1(1"CARD",1"RAD")1.5N
SELECT DISTINCT PatientRef
FROM APPOINTMENT.BOOKING
WHERE Active='Y' AND
LocationCode %PATTERN '2A1"-"1(1"CARD",1"RAD")1.5N'

Challenge

Use generative AI to assist create and modify pattern match code from different natural language input.

English Description French Description Spanish Description
module A
  one of String "CARD"
   or
  one of String "RAD"
the main pattern is as follows:
  two of alphabetic characters
  followed by one of Character minus
  followed by one of module A
  followed by one to five of numeric characters
module A
  un de Corde "CARD"
   ou
  un de Corde "RAD"
le motif principal est le suivant:
  deux de caractères alphabétique
  suite par un de caractère moins
  suite par un de module A
  suite par un à cinq de caractères numériques
módulo A
  uno de Cadena "CARD"
   o
  uno de Cadena "RAD"
el patrón principal es la siguiente:
  dos de caracteres alfabéticos
  seguido de uno de carácter menos
  seguido de uno de módulo A
  seguido de uno para cinco de caracteres numéricos

 

Link to online demo

Pattern Match WorkBench ( Hosted on HuggingFace  )

Existing Features

  • A multilingual display
  • Self Documented application. Shows where patterns are used in the IRIS product.
  • Transform English, French and Spanish description input, into new pattern match code.
  • Describe an existing pattern match code in plain English, French or Spanish
  • Generates lists of matching and non-matching sample values
  • Validates a table batch of matching values to an existing pattern match expression in code.
  • Extracts patterns from Object Script and SQL. ie: Just paste a line of code and the demo will detect the pattern 

Pending further model training

  • Generate suggested generalized new patterns from a given list of matching values
  • Shortcuts for common information types. For example "email", "web address"

Technologies used

  • Python for Serverside logic, model training and dataset creation
  • Gradio - web page framework
  • Huggingface - for Training models and application Deployment

Appreciate any feedback.

Thank you

Discussion (0)1
Log in or sign up to continue