Skip to content

bag4355/HKT

Repository files navigation

ACCUTE: Accounting Compliance Cognitive Understanding Task Engine

ACCUTE is an AI-powered web application that assists accounting professionals in analyzing corporate disclosure documents and identifying similar financial reporting cases. It leverages Upstage API to process Korean business reports in PDF format and returns concise summaries and case comparisons to support compliance and decision-making tasks.

Features

  • PDF parsing and financial issue extraction
  • Intelligent summarization of key points using Upstage API
  • Similar case retrieval based on semantic similarity
  • Clean, responsive UI built with Tailwind CSS
  • Easily extensible for future AI agent integration

Project Structure

├── agi_pipeline.py         # Core logic using Upstage API for summarization & retrieval
├── app.py                  # Backend Flask app for API endpoints
├── index.html              # Frontend interface with file upload and response rendering
├── crawled_data.xlsx       # Sample case database for similarity search

Installation

  1. Clone this repository:
git clone https://github.com/your-username/accute.git
cd accute
  1. Create and activate a virtual environment:
python3 -m venv venv
source venv/bin/activate
  1. Install dependencies:
pip install -r requirements.txt
  1. Run the application:
python app.py

Usage

  1. Open index.html in your browser.
  2. Enter your query in Korean (e.g., "유사한 회계 보고서 찾아줘").
  3. Upload a business report in PDF format.
  4. Click the 실행하기 button.
  5. View the AI-generated summary and similar cases side by side.

Upstage API Integration

The agi_pipeline.py file uses the Upstage API to perform the following tasks:

  • Summarize PDF contents
  • Retrieve similar financial reporting cases based on vector similarity

These interactions are clearly marked within the code (see comments such as # --- Upstage Summary API Call ---).

Testing & Evaluation

You can test the app by uploading the sample business report:

  • [가이아코퍼레이션]사업보고서(2025.03.31).pdf

The app will process the document and return:

  • A structured summary
  • Relevant similar case(s) from the internal case database (crawled_data.xlsx)

Dependencies

  • Flask
  • pandas
  • PyMuPDF
  • openai / Upstage API SDK (you must set your API key)

License

This project is developed as part of an academic coursework submission and is currently not open-sourced under a specific license.


Developed by: Brave Team of Five
Affiliation: [Your School or Course Name]

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors