What is Nourish?

Nourish is an AI solution to food management industries in optimizing inventory management by predicting Projected Par Levels (PPL for short). Nourish's goal is to improve efficiency, reduce waste, and ensure that there is enough food to reach those in need. Nourish's flagship feature is allowing users to upload csv files which a seperate AI on a server will predict categories and PPL levels based on upload data.

Usage (User)

A CSV file uploaded must atleast have the columns with names and values exactly as follows:

• A column with the header "Name" where each row has the name of each food.
• A column with the header "Units" where each row has the amount of units orders at that time.
• A column with the header "Forecast" where each row has the forecasted unit orders.

Nourish also comes with an LLM that users can ask for help in understanding how to interact with the project.

If the file is correctly formatted and processed, the user may download a new CSV file with categories and PPL (The amount to order) filled in.

Usage (Host)

NourishAI may be used by anyone. Here are the steps to setup the AI:

• Clone/Download the repository
• Run pip install -r requirements.txt in a terminal/console to install required dependencies
• Run app.py in src/backend/app/app.py
• Host the index.html in src/frontend/index.html

Users may now upload files which will be stored and automatically processed in the Data Folder and become downloadable. LLM messages will also be logged.

How Nourish Works

After a CSV File that is correctly formatted is uploaded, the backend AI will now process the data:

• A column named "Categories" which will be filled in with the category of each food predicted using AI.
• The Unit Ratio is the ratio between the amount of each item and the amount of that item's category. The unit ratio will be predicted which is the item's units divided by its total category units.
• PPL is essentially the amount of units the should be ordered. A new column name "PPL" is created by multiplying the unit ratio by the forecast for each item.
• The user can now download a new CSV file with both "Categories" and "PPL" filled in.