Overview 🔮

Email system with spam detection, spelling check, AES encryption.

Screen.Recording.2024-02-26.at.22.59.07.online-video-cutter.com.mp4

Functionality 🛠️

• Automatic spam detection
• Auto refresh model after (10) new emails, or manually after marking an email as spam
• Check spelling when writing an email
• AES encryption
• Send, receive, delete, mark as spam
• Status: received ... (read, unread)

How it works 🧠

• The spam detection uses data (only) generated by prompt from the OpenAI API - GPT Turbo 3.5
  • Make the email into UTF-8
  • Parse eventual HTML with BeautifulSoup
  • Remove links, mentions, hashtags, emails, numbers, punctuation, and stopwords
  • Stem the words using: PorterStemmer
  • Make Term Frequency-Inverse Document Frequency (TF-IDF) matrix
  • Train the model (limit 500 first features), and took LinearSVC as the best model
  • Save the model into file using joblib
• The spelling check uses the Levenshtein distance to find the most similar word in a French (lite) dictionary

How to use ℹ️

• Create a virtual environment

python3 -m venv venv

• Activate the virtual environment

source venv/bin/activate

• Install the requirements

pip install -r requirements.txt

• Download the necessary data

python3 -m nltk.downloader stopwords
python3 -m nltk.downloader punkt
python3 -m nltk.downloader wordnet
python3 -m nltk.downloader omw

• Run the program

python3 manage.py runserver