# Task Card Processor

This project extracts a structured task card from a raw user text using a single LLM call and a Pydantic parser.

## Installation

```bash
# Create a virtual environment (recommended)
python -m venv .venv
source .venv/bin/activate   # On Windows use .venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt
```

## Configuration

The script uses OpenAI API.  Set the following environment variables before running:

- `OPENAI_API_KEY` – your OpenAI API key.
- `OPENAI_API_BASE` (optional) – override the default OpenAI base URL if you use a local LLM.

## Usage

You can pipe raw text into the script or provide it as a command‑line argument.

```bash
# From stdin
cat raw_text.txt | python agent.py

# As an argument
python agent.py "Generate a task card for a math homework assignment."
```

The script prints a validated JSON object followed by a short summary.

## Example Output

```json
--- Task Card JSON ---
{
  "title": "Math Homework",
  "subject": "Algebra",
  "deadline_hint": "Next Friday",
  "deliverable_type": "Problem set",
  "grading_hints": "Include solutions and explanations"
}

--- Summary ---
Title: Math Homework
Subject: Algebra
Deadline: Next Friday
Deliverable: Problem set
Grading hints: Include solutions and explanations
```

## License

MIT License.