Docker: Building and Running

1. General Instructions

docker build --platform linux/amd64 -t my-streamlit-app .

docker run --env-file .env -p 8501:8501 my-streamlit-app

sudo docker build -t my-streamlit-app .

sudo docker run --env-file .env -p 8501:8501 my-streamlit-app

Clone the Repository:

git clone https://github.com/yourusername/yourrepo.git
cd yourrepo

Install Dependencies: It’s recommended to use a virtual environment:
```
pip install -r requirements.txt
```
Required Python packages include:
- streamlit
- PyPDF2
- langchain
- gTTS
- OpenAI Python bindings
- SQLAlchemy
- SentenceTransformers
- FAISS
- AssemblyAI Python bindings
- yt-dlp
- BeautifulSoup4, urllib3 for web scraping
Modify the .env file
Run the App: Start the Streamlit application:
```
streamlit run application.py
```

Here’s an improved, well-organized version of your Docker notes:

By default, your Streamlit app will run at:

Simply open this link in your browser to view the app.

Home Page:
From the homepage, you can navigate to sign up, log in, or explore About/Contact pages.
Sign Up / Login:
- Professor: After logging in, you’re taken to a dashboard where you can create courses, upload files, set YouTube links, and view insights.
- Student: After logging in, you see available courses, can open them, ask questions, generate flashcards/MCQs, summarize materials, or convert them into podcasts.
Course Management (Professor Only):
- Create a new course.
- Upload PDFs/TXT files as course materials.
- Add a YouTube video link, which will automatically process transcripts (if configured).
Student Interactions:
- View Course Details: See what materials are available.
- Ask Questions: The system uses vector embedding and GPT to answer.
- Generate Flashcards/MCQs: Click the respective buttons and receive AI-generated study aids.
- Summarize Content: Quickly get a summary of all course materials.
- Podcast Generation: Upload PDFs and generate a script/audio podcast to study on the go.
Data Insights (Professor Only):
- View charts and a report of all asked questions (topics distribution, word clouds).
- Improve course content based on student query analysis.

Models:
You can switch GPT models (like gpt-4 or gpt-3.5-turbo) by editing the code in application.py or podcast.py.
Styling:
Custom CSS is injected at runtime. To adjust styling, modify application1.py or the inline CSS sections in the code.
Database:
The default SQLite database can be swapped with another SQL database by updating DATABASE_URL in database.py.
Embeddings & Vector Store: By default, it uses the all-MiniLM-L6-v2 SentenceTransformer model. You can change the embedding model in application.py or other embedding scripts.