Should You Eat This?

An easy-to-use app that provides advice to help you make informed decisions about your food choices.

• Streamlit
• OpenAI API (GPT-4)
• Flask

Making healthy food choices can be challenging, especially when trying to achieve specific health goals like weight loss or maintaining a balanced diet. This project aims to simplify the decision-making process by providing users with personalized dietary advice based on their health information and the food items they plan to consume via taking and uploading a simple picture of what they plan to eat.

1. Clone the repository
2. Install the required dependencies: pip3 install -r requirements.txt
3. Set up your OpenAI API key as an environment variable via a .env file or via the command export OPENAI_API_KEY=your_api_key_here
4. Also add the base URL for the openai API usage via the .env file or via the command export OPENAI_ENDPOINT=your_base_url
5. Run the Streamlit app: streamlit run app.py
6. Additionally to expose the local API, run python3 server.py
7. Now you can add user inputs. The local API can be found and available via CLI: curl http://localhost:5000/get_health_details/<user_id>

Through this project, I gained valuable experience in integrating various technologies and APIs to create a practical application. Some key learnings include:

• Utilizing computer vision and natural language processing via large language models to analyze food images and extract relevant information.
• Developing a user-friendly interface with Streamlit to improve the overall experience.
• Implementing algorithms to calculate nutritional values and provide personalized dietary advice based on user inputs and visualizing this data.
• Handling API requests and responses, as well as incorporating error handling and user feedback mechanisms.

1. Integrating Multiple APIs: Combining the functionality of different APIs (e.g., OpenAI, computer vision APIs) required careful coordination and data handling to ensure a seamless user experience.

2. Local storage: As dietary recommendations and nutritional guidelines evolve, maintaining and updating the application's knowledge base becomes crucial for providing accurate and up-to-date advice via storage of user data.

3. Handling Edge Cases: Accounting for various edge cases, such as incomplete or invalid user inputs, required additional error handling and input validation mechanisms.

Despite these challenges, working on this project has been a learning experience, reinforcing the importance of attention to detail, thorough testing, and continuous iteration to create a robust and user-friendly application.

1. After I upload an image on streamlit, an error with Failed to get a response from the server. or a similar error is shown. Ans: Your openai key is either revoked or expired. Use a new key and add it as an environment variable either in streamlit env runtime or as a variable if running locally.

2. What is a base-URL for Openai? Ans: The project was built based on a different custom URL provided with credits for a class project work. For normal usage, please use the base api url from openai documentation. For 510 analysis, use base URL provided by the Instructor to successfully execute the commands.