Skip to content

Latest commit

 

History

History
64 lines (47 loc) · 3.27 KB

README.md

File metadata and controls

64 lines (47 loc) · 3.27 KB
Raw

Meal Master - AI-Powered Recipe Recommendations 👨‍🍳🤖

Meal Master is a simple meal suggestion application built with Streamlit, leveraging the power of Google's Gemini Pro LLM and a SQLite database.

Features ✨

  • Personalized Recommendations:
    • Suggests recipes based on user dietary preferences (e.g., American 🇺🇸, Italian 🇮🇹) and allergies (e.g., Dairy 🥛, Shellfish 🦐).
    • Considers past liked recipes ❤️ to enhance suggestions.
  • Ingredient-Based Suggestions:
    • Users can upload images 📸 of ingredients, and the app uses Gemini to extract ingredient information.
    • Recommends recipes that incorporate the detected ingredients 🍅🥕🧅.
  • Recipe Details: Provides a full recipe, including ingredients and instructions, in an easy-to-read markdown format.
  • Nutritional Information: Displays a bar chart 📊 visualizing the estimated nutritional value of the recommended recipe (calories, protein, fat, etc.).
  • Like and Save: Users can "like" 👍 recipes, which are then saved to their history for future reference.

Technologies Used 💻

  • Frontend: Streamlit
  • Language Model: Google Gemini Pro
  • Database: SQLite
  • Database ORM: SQLAlchemy
  • Visualization: Plotly

Setup and Installation 🚀

  1. Clone the Repository: git clone git@github.com:Red-54/MealMaster.git

  2. Install Dependencies: pip install -r requirements.txt

  3. Google Cloud API Key:

    • Obtain a Google Cloud API Key 🔑 and enable the Gemini API.
    • Create a secrets.toml file in your Streamlit app directory.
    • Add the API Key to the secrets.toml file:
    [secrets]
GOOGLE_API_KEY = "your_api_key_here"

  4. Run the App: streamlit run app.py

Usage 🧑‍💻

  1. Login/Sign-up: Create an account or log in.
  2. Preferences: Go to the "Preferences" section to set your dietary preferences and allergies.
  3. Suggestions:
    • Optionally, upload an image of ingredients.
    • Click the "Recommend" button to receive a personalized recipe suggestion.
    • View the recipe details, nutritional information, and "like" ❤️ it to save it to your history.

Limitations ⚠️

  • No Fine-tuning: The app currently relies on Gemini's pre-trained capabilities without any fine-tuning. This may lead to occasional inaccuracies in ingredient detection or recipe suggestions.
  • Simplified Recommendations: The recommendation system is rule-based and primarily focuses on ingredient matching and dietary restrictions. Collaborative filtering or more advanced AI algorithms are not yet implemented.
  • SQLite Database: SQLite is suitable for a small-scale application, but it may not be ideal for handling a large user base or complex data analysis in the future.

Future Enhancements ➡️

  • Improved Authentication 🔐: Enforce strong password policies and provide options for password recovery.
  • Database Migration: Consider migrating to a more scalable database (e.g., PostgreSQL) if the application grows significantly.
  • User History 📖: Allow users to see previously liked recipes or remove them.

Contributing 🤝

Contributions are welcome! Please open an issue or submit a pull request if you have ideas for improvements.