PACLI

A powerful, AI-driven personal assistant for your calendar, scheduling, and event management. Built with LangChain, Groq LLM, and a flexible tool system, PACLI helps you manage your schedule, add and edit events, and answer natural language queries about your calendar—all from the command line.

Features

• Natural Language Scheduling: Add events using plain English (e.g., "Schedule OA next Tuesday at 5pm").
• Flexible Event Retrieval: Ask for events in any time range ("next week", "next 2 weeks", "next 5 days", etc.).
• Event Editing: Edit or update events by name and date.
• Event Search: Find events by name and date, even with fuzzy matching.
• Persistent Storage: All events are stored in local JSON files for reliability and privacy.
• Autonomous Tool Calling: The agent automatically selects and calls the right tool for your query.
• Customizable Prompts: Easily update assistant instructions via the YAML prompt file.
• Colorful CLI Output: Enjoy a clear, readable interface with rich formatting.
• Codeforces Contest Integration: Query upcoming Codeforces contests, schedule them to your calendar, and get contest details for any date or time range using natural language (e.g., "Show Codeforces contests next week").
• Automated Next-Day Event Email: Get a beautiful HTML email of your next day's events sent automatically via Gmail.

Automated Next-Day Event Email

1. Schedule send_scheduled_mail.py with Windows Task Scheduler to send the email at your preferred time.
2. The email uses HTML formatting for a clear, modern look.
3. After sending, the events file is deleted for privacy.

Setup:

• Add your Gmail credentials and recipient email to .env (use an App Password if 2FA is enabled).
• Make sure python-dotenv is installed (pip install python-dotenv).
• Use the real Python executable path in Task Scheduler (not the Windows Store launcher).

Note:

• Your confidential data stays local; no event info is sent to external servers except Gmail.
• Docker support is currently not added for email automation or the calendar/frontend.
• To use the calendar UI (frontend), run it locally and fetch event data directly from the backend. You can schedule and interact with the frontend and email/calendar automation using Windows Task Scheduler or by running the scripts manually. See below for instructions.

Usage

1. Installation

1. Clone the repository:

   git clone https://github.com/saqlain2204/PACLI.git
   cd PACLI

2. Install dependencies:

   pip install -r requirements.txt

3. Set up your environment: Create .env and add your Groq API key and Gmail credentials:

   GROQ_API_KEY=your_groq_api_key_here
   GMAIL_ADDRESS=your_gmail_address@gmail.com
   GMAIL_APP_PASSWORD=your_gmail_app_password
   RECIPIENT_EMAIL=recipient_email@gmail.com

2. Running the Assistant

Run with Python (local)

python main.py

You'll see:

How can i help you today? >

Type your query, such as:

• Schedule OA next Wednesday at 10am
• What events do I have next week?
• Edit Interviews on 09-08-2025, set time to 9:30 AM
• Find OA on 05-08-2025

3. Event Data

• Events are stored in events/event_data.json (excluded from git by default).
• The assistant supports date formats like DD-MM-YYYY and natural language ("next Friday").

4. Customizing Prompts

• Edit prompts/personal_assistant.yaml to change the assistant's instructions or logic.
• No code changes needed—just update the YAML file.

5. Adding/Editing Tools

• Tools are defined in the tools/ folder.
• Add new tools or update existing ones to extend functionality.

Project Structure

Personal Assistant/
├── agents/
│   └── pa_agent.py
├── tools/
│   ├── event_scheduler.py
│   ├── get_events.py
│   ├── edit_event_tool.py
│   ├── find_event_tool.py
│   └── ...
├── prompts/
│   └── personal_assistant.yaml
├── events/
│   └── event_data.json
├── main.py
├── send_scheduled_mail.py
├── README.md
└── requirements.txt

Advanced Usage

• The assistant supports flexible queries and autonomous tool selection.
• You can extend the agent with new tools, prompts, or UI (e.g., Streamlit).

License

MIT

Author

Mohammed Saqlain

---

Enjoy PACLI, your AI-powered personal assistant! If you have questions or want to contribute, open an issue or pull request.