Introduction

The Western Australian Centre for Road Safety Research (WACRSR) is initiating a project to automate the manual process of scanning and analyzing road safety literature. Currently, this process is labor-intensive, taking up valuable time that could be better spent on critical research tasks. To address this, the project aims to develop a system that leverages advanced large language models (LLMs) to streamline literature scanning and evaluation.

Group Members

Project Structure

./
├── api/
│   ├── app/
│   │   ├── controllers/
│   │   ├── models/
│   │   ├── services/
│   │   ├── utils/
│   │   └── __init__.py
│   ├── logs/
│   ├── config.py
│   ├── Dockerfile.<env>
│   ├── README.md
│   └── requirements.txt
├── secrets/
│   ├── els_api_key.txt
│   ├── els_token.txt
│   └── llm_api_key.txt
├── ui/
│   ├── app/
│   │   ├── about/
│   │   ├── assets/
│   │   ├── components/
│   │   ├── context/
│   │   ├── home/
│   │   ├── results/
│   │   ├── search/
│   │   ├── page.tsx
│   │   ├── layout.tsx
│   │   └── ...
│   ├── public/
│   ├── .eslintrc.json
│   ├── .gitignore
│   ├── Dockerfile.<env>
│   ├── next-env.d.ts
│   ├── next.config.mjs
│   ├── package-lock.json
│   ├── package.json
│   ├── README.md
│   └── tsconfig.json
├── .gitignore
├── docker-compose.<env>.yml
└── README.md

Launch Papermate Application

0: Install Docker Desktop

Go to Docker website to download and install the application. Once Installed, make sure the docker engine is running.

1. Secrets

Contact developers to acquire the necessary three secret files: API key for Gemini, API key for Elsevier and Institutional token. After getting the files, create a folder named secrets/ under root folder:

├── secrets/
│   ├── els_api_key.txt
│   ├── els_token.txt
│   └── llm_api_key.txt

1.5. (Optional) Create launch.json

If you would like to use vscode debugger function, add the following content to cofiguration file launch.json:

{
    "version": "0.2.0",
    "configurations": [
        {
            "name": "Python: Remote Attach",
            "type": "debugpy",
            "request": "attach",
            "connect": {
                "host": "127.0.0.1",
                "port": 5678
            },
            "pathMappings": [
                {
                    "localRoot": "${workspaceFolder}/api",
                    "remoteRoot": "/app"
                }
            ],
            "justMyCode": true
        }
    ]
}

2. Run docker-compose

Open the terminal that is under root foler, choose an environment:

• For normal usage, enter docker compose up --build
• For debugger with launch.json, enter docker compose -f docker-compose.debug.yml
• For running test environment, enter docker compose -f docker-compose.test.yml up --build

3. Access Papermate

Go to the url of Papermate http://localhost:3000/, the app should be up and running now.

User Manual

0. User Interface:

Before you dive into the platform, familiarize yourself with its layout and functionalities. The webpage is divided into four main sections: Home, Search, Result, and About.

[1] Home

The Home page provides a brief introduction to the platform and its key functionalities.

[2] Search (Literature Paper Search) picture will be provided after ui done

The Search page is where you can query the database for relevant academic papers.

[3] Result picture will be provided after ui done

The Result page displays the search results only after you click Proceed on the Search page.

[4] About

The About page provides details about our development team and links to the GitHub repository for project updates and contributions.

1. Operation Steps:

In the search page:

• Enter a Query (compulsory): Input keywords or phrases into the query field to search for specific topics or papers.

• Enter Date Range (optional): Specify the time frame for the search by selecting "From Date" and "To Date."

• Enter Title and Author (optional): Provide specific titles or authors to narrow down the search results.

• Click Upload Publication CSV (optional): Upload a CSV file containing additional search parameters.

• Click Gemini 1.5 Pro Option (optional): Users can enable or disable the Gemini Pro model for enhanced search capabilities.

• Click Search Button: Click the "Search" button to initiate the query.

• Click Proceed Button: After reviewing the results count, click the "Proceed" button to view detailed results on the Result page.

In the Result page

• After you click Proceed on the Search page (weather you have clicked search button or not):

• View Results: Check the list of academic papers returned by the search, which includes details like DOI, title, abstract, author, publication, publish date, URL, relevance, and synopsis.

• Click Columns: You can customize which columns to display, including additional categories like "Study Type" using the Show Instruction Box.

• Click Filter, Density or Export: Utilize options to filter items, change page density, and export results as CSV files for further analysis.

• Click Pagination: Navigate through the results using pagination controls, allowing users to view additional items if the results exceed the displayed limit.

• Adjust Column Widths: Users can customize the appearance of the results table by adjusting column widths.

• Click Show Instruction Box:

• Add categories to your search results (e.g., type "add Study Type" to create a new column).

• Hover to see the time and click to see the detailed instructions that you typed.

2. FAQ:

[1] Image Conflict:

If you encounter issue during building images, you can clear the Docker cache by using the following command:

docker system prune –a 

[2] Search and Proceed Functions:

Please be patient, as the search or proceed functions may take some time to complete.