https://github.com/hediske/chefai
AI Recipe Assistant to help users create a suitable recipe . This project is part of a Hackathon
https://github.com/hediske/chefai
Last synced: about 1 month ago
JSON representation
AI Recipe Assistant to help users create a suitable recipe . This project is part of a Hackathon
- Host: GitHub
- URL: https://github.com/hediske/chefai
- Owner: hediske
- Created: 2024-12-16T12:12:10.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2025-08-09T04:51:16.000Z (2 months ago)
- Last Synced: 2025-08-09T06:22:30.873Z (2 months ago)
- Language: Python
- Homepage:
- Size: 534 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
Awesome Lists containing this project
README
# 🍽️ AI Recipe Assistant with Chat History & RAG
An intelligent recipe assistant built with **Streamlit**, **LangChain**, and **ChromaDB**, capable of:
- Answering cooking-related queries
- Recalling past conversations with chat history
- Searching a recipe database using Retrieval-Augmented Generation (RAG)
- Streaming AI responses in real-time
---
## 🚀 Features
- **Chat History**: Your previous conversation is remembered during a session
- **RAG Integration**: Fetches relevant recipes from the vector database
- **Streaming Responses**: See AI answers in real-time
- **Persistent Storage**: Recipes stored locally in ChromaDB
---
## 🛠️ Tech Stack
- **Python 3.10+**
- [Streamlit](https://streamlit.io/)
- [LangChain](https://www.langchain.com/)
- [ChromaDB](https://www.trychroma.com/)
- [OLlama](https://www.ollama.com)
---
## 📂 Project Structure
```
Prepare_Hack/
├── data/
│ ├── recipes_raw_nosource_ar.json
│ ├── recipes_raw_nosource_epi.json
│ ├── recipes_raw_nosource_fn.json
│ └── test.json
├── RAG/
│ ├── .env
│ ├── docker-compose.yml
│ ├── requirements.txt
│ └── src/
│ ├── app.py # Streamlit app entry point
│ ├── config.py # Environment/config management
│ ├── RAG_pipeline.py # RAG pipeline logic
│ ├── lib/
│ │ ├── chroma_database.py
│ │ ├── chroma_store.py
│ │ └── ... # Other utility modules
│ ├── chroma_data/
│ │ └── chroma.sqlite3 # Local ChromaDB storage
│ ├── chroma_db/
│ ├── upload/
│ │ └── queue/
│ │ └── utils.py # Data ingestion helpers
│ └── test.py
└── script.py
```
---
## 🖼️ How it Works (Architecture Diagram)
```mermaid
flowchart TD
User[User - Web Browser]
Streamlit[Streamlit UI - app.py]
RAGPipeline[RAG Pipeline - RAG_pipeline.py]
Retriever[ChromaDB Retriever]
ChromaDB[ChromaDB Vector DB]
LLM[LLM - LangChain/OLlama]
History[Chat History]
DataIngest[Recipe Ingestion - upload/queue/utils.py]
User -- "Query" --> Streamlit
Streamlit -- "User Input" --> RAGPipeline
RAGPipeline -- "Retrieve Recipes" --> Retriever
Retriever -- "Query" --> ChromaDB
ChromaDB -- "Relevant Recipes" --> Retriever
Retriever -- "Recipes" --> RAGPipeline
RAGPipeline -- "Prompt + Context" --> LLM
LLM -- "AI Response" --> Streamlit
Streamlit -- "Show Response" --> User
RAGPipeline -- "Update" --> History
DataIngest -- "Add Recipes" --> ChromaDB
```
---
## 📦 Installation
1. **Clone the repository**
```bash
git clone https://github.com/hediske/ChefAI.git
cd Prepare_Hack/RAG
```
2. **Create a virtual environment**
```bash
python -m venv venv
source venv/bin/activate # macOS/Linux
venv\Scripts\activate # Windows
```
3. **Install dependencies**
```bash
pip install -r requirements.txt
```
4. **Set environment variables**
Create a `.env` file in `Prepare_Hack/RAG/`:
```env
'CHROMADB_DIRECTORY' = "YOUR_DIRECTORY"
'CHROMADB_CLOUD_HOST' = "YOUR_CLOUD_HOST"
```
---
## 📤 Upload Recipes to the Database
You can upload your own recipes to ChromaDB:
1. Place your recipe JSON files in `Prepare_Hack/data/` (see sample files like `test.json`).
2. Use the ingestion script to process and upload recipes:
For the type it can be : `recipe` if you want a recipe with the same
```bash
python -m upload.app --file --type
```
Or, if you have a full folder to embed you can do it in batch:
```bash
python -m upload.app --folder --type
```
This will process your recipes and store the embeddings in `src/chroma_data/chroma.sqlite3`.
For Our case we used data from Internet to create the recipe suggestions :
[Link to The dataset](https://eightportions.com/datasets/Recipes/)
---
## ▶️ Running the App
```bash
streamlit run src/app.py
```
Open your browser and go to:
```
http://localhost:8501
```
---
## 💡 Example Queries
- "Show me vegetarian pasta recipes"
- "I want a dessert with chocolate"
- "What can I cook with chicken and lemon?"
---
## ✨ Why This Project is Special
- **Combines memory and retrieval**: You can have a natural conversation and still pull from a recipe database.
- **Modular design**: Easy to swap datasets, LLMs, or prompts.
- **Fast**: Uses local ChromaDB for quick recipe retrieval.
- **Scalable**: Can be expanded to multiple cuisines, nutrition info, or personalized recommendations.
---
## 📜 License
MIT License © 2025 Hedi <3