Skip to content

Hyrin-mansoor/changai

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Changai - ERPNext Natural Language Interface

A modular AI pipeline that converts natural language questions into accurate ERPNext (Frappe) queries, executes them, and returns results in human-friendly language. Built to simplify ERP access for business users with no technical knowledge.


Overview

This system allows users to interact with ERPNext using natural questions such as:

“Which suppliers from India are transport providers?”

The pipeline :

  • Understands the user intent
  • Predicts the relevant ERP doctype and fields
  • Generates and executes a valid Frappe query
  • Converts the response into a clear, user-friendly sentence using an LLM

Response:

The following suppliers from India provide transport services: IndoTrans Logistics, QuickMove Carriers.


Pipeline Architecture

graph TD
  A[User Question] --> B[Doctype Detection: RoBERTa]
  B --> C[Top Field Detection: SBERT]
  C --> D[Query Generation: FLAN-S3]
  D --> E[Frappe Query Execution]
  E --> F[LLM Natural Language Output]

Loading

Contributions

Contributions are welcome!

You can help by:

  • Improving datasets or adding new question styles
  • Fixing issues or bugs in the pipeline
  • Enhancing query generation or accuracy

To contribute, please open an issue or submit a pull request.

License

This project is licensed under the MIT License.
#mit

About

AI for Frappe ERPNext

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 44.6%
  • JavaScript 32.7%
  • CSS 22.7%