Automated Cover Letter Generator

Automated pipeline for creating personalized cover letters:

• Data Collection: Clipping job postings from Notion (or another source).
• Processing & Generation: Using an LLM (via Outlines) to produce a structured cover letter.
• PDF Output: Generating final PDF cover letters using LaTeX templates.
• Orchestration: Optional Zapier integration for automation, or direct polling/webhook triggers.

Table of Contents

• Automated Cover Letter Generator
  • Table of Contents
  • Features
  • Project Structure
  • Requirements
    • Python Dependencies
  • Installation
    • Windows Setup
    • Unix Setup (Linux/Mac)
    • Docker Setup (Cross-Platform)
  • Usage
  • Environment Variables
  • Development
  • Roadmap
  • License
  • Questions or Contributions?

Features

• Notion Integration: Seamlessly clip job postings into your Notion database.
• LLM-driven Generation: Uses open-source LLMs (LLaMA, Mistral) or fallback to GPT-4/Claude.
• Structured Output: Enforced JSON schema via Outlines, ensuring reliable cover letter segments.
• Latex PDF Generation: Fills a TeX template, automatically produces professional PDFs.
• Dockerized: Simple deployment with docker-compose.

Project Structure

.
├─.venv/               # Local virtual environment (Windows or Unix)
├─.venv-3.13/          # Another local virtual environment for Python 3.13
├─assets/              # Images or static assets
├─config/              # Configuration files (YAML, TOML, etc.)
├─cover_letters/       # Generated PDFs or letter drafts
├─data/                # If you store CSV/JSON data or intermediate results
├─docs/                # Documentation (api, dev, user)
├─examples/            # Example scripts or usage demos
├─logs/                # Log files
├─requirements.lock    # Locked dependencies
├─requirements.txt     # Main Python dependencies
├─scripts/             # Setup or utility scripts
│  ├─setup/
│  └─tools/
├─src/
│  ├─api/
│  ├─core/
│  ├─server/
│  │  └─webhook_server.py  # A Flask or FastAPI server for receiving triggers
│  └─utils/
├─templates/
│  ├─awesome-cv.cls                   # CV style (from Awesome-CV)
│  ├─awesome_cv_cover_letter_template.tex
│  └─coverletter.tex                  # A simpler cover letter template
├─tests/
│  ├─fixtures/
│  ├─integration/
│  └─unit/
├─Dockerfile
├─docker-compose.yml
├─README.md                           # You're reading it now
└─...

Requirements

Below is the core list of dependencies you'll need outside of Python:

• Python 3.9+
  • Recommended to use Python 3.11 or 3.12, or you can try 3.13 if you're prepared to compile Rust code for certain packages.
• Rust toolchain (if needed for outlines-core compilation)
  • On Windows, Install Rust via rustup.
  • On Unix, install via your package manager or rustup.
• TeX distribution
  • On Windows: MiKTeX or TeX Live.
  • On Unix: sudo apt-get install texlive-latex-extra texlive-xetex (or distro equivalent).
• Docker (optional)
  • If you prefer containerized deployment.

Python Dependencies

All Python libraries are listed in requirements.txt. Key libraries include:

• Outlines for structured generation.
• PyTorch or Transformers (depending on your open-source LLM usage).
• Flask or FastAPI if you run a server.
• requests or any HTTP library for triggers/integrations.

Installation

Windows Setup

1. Install Python

   • Get the latest Python 3.x from python.org.
   • During install, check "Add Python to PATH."

2. (Optional) Install Rust

   • If outlines-core needs to compile from source (no pre-built wheel for your Python version).
   • Download from rustup.rs → run the .exe.

3. Install a TeX Distribution

   • MiKTeX or TeX Live for Windows.

4. Create & Activate Virtual Environment

# In your project folder:
python -m venv .venv
.\.venv\Scripts\activate

5. Install Python Dependencies

pip install --upgrade pip
pip install -r requirements.txt

6. Verify

# Example check
python -m outlines --version

Unix Setup (Linux/Mac)

1. Install Python
   • Use your package manager or python.org.
   • e.g., on Ubuntu:

sudo apt-get update
sudo apt-get install python3 python3-venv python3-pip

2. (Optional) Install Rust

   • If needed for building outlines-core.
   • e.g. curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

3. Install a TeX Distribution

   • e.g. sudo apt-get install texlive-latex-extra texlive-xetex.

4. Create & Activate Virtual Environment

python3 -m venv .venv
source .venv/bin/activate

5. Install Python Dependencies

pip install --upgrade pip
pip install -r requirements.txt

6. Verify

python -m outlines --version

Docker Setup (Cross-Platform)

1. Install Docker

   • Windows: Docker Desktop
   • macOS: Docker Desktop or brew cask install docker
   • Linux: Check your distro's repo or Docker Docs

2. Build Image

docker build -t cover-letter-app .

3. Run Container

docker run -p 8080:8080 cover-letter-app

4. (Optional) docker-compose
   • Use the provided docker-compose.yml if you have multiple services or want a simpler spin-up:

docker-compose up --build

• This approach ensures a consistent environment across all platforms.

Usage

1. Start Your Backend (Locally or via Docker)
   • If local (without Docker):

# Still in your .venv
python src/server/webhook_server.py

• If Docker:

docker-compose up

2. Add a Job Posting

   • For example, in Notion, create a new entry in your "Job Postings" database.
   • If using Zapier or a custom script, it triggers the endpoint (e.g. POST /generate-letter).

3. Cover Letter Generated

   • The LLM processes it via Outlines, ensures JSON format.
   • A LaTeX template is filled, compiled to PDF, and saved in ./cover_letters.

4. Review the PDF

   • Check in cover_letters/YourCoverLetter.pdf.

Environment Variables

You can use a .env file or system environment variables to store:

Variable Name	Description	Example
OPENAI_API_KEY	API key if using GPT-4 fallback.	sk-ABC123
NOTION_API_TOKEN	If using direct Notion API polling.	secret_...
FLASK_ENV	Set to development or production.	development

(Ensure .env is in your .gitignore to avoid committing secrets.)

Development

• Code Linting
  • Use a linter (flake8, black, ruff) for consistent style:

pip install black
black src/

• Unit Tests
  • Place unit tests under tests/unit/.
  • Example:

pytest tests/unit

• Integration Tests
  • Check pipeline flow from Notion input to PDF output.
  • Place them in tests/integration/.

Roadmap

• Add Kubernetes Deployment: Helm chart for scaling multiple LLM containers.
• Switch/Upgrade Models: Try LLaMA-2-13B, Mistral, or a fine-tuned version for better cover letters.
• Retrieval-Augmented Generation: Store resume or user data in a vector DB for highly contextual letters.
• User Interface: A small React/Flask front end for easier manual triggers.
• CI/CD: GitHub Actions to run tests, build Docker images automatically.

License

MIT License – Feel free to use or adapt this project for your own purposes. (Replace with your actual chosen license.)

Questions or Contributions?

• Open an issue on GitHub or create a Pull Request.
• Feel free to reach out with suggestions, feature requests, or bug reports!

Happy Generating!