🧠 wasserstoff-AiInternTask

A smart AI-powered Personal Email Assistant built as part of the Wasserstoff AI Internship Task.
It fetches unread emails, summarizes them using LLMs, detects calendar intent, sends Slack notifications, and drafts replies — showcasing AI integration and automation skills.

---

🚀 Features

• ✅ Fetch unread emails via Gmail API
• 🧠 Summarize email content using Hugging Face’s BART model
• 🗓️ Detect scheduling intent and create Google Calendar events
• 🔁 Thread-aware context gathering for smart replies
• 📨 Auto-generate reply drafts with option to confirm
• 🔔 Send Slack notifications for each processed email
• 💾 Store and update email data in SQLite using SQLAlchemy
• 🧰 Modular and extensible project structure

---

🗂️ Project Structure

AiInternTask/
├── app.py
├── check.py
├── credentials.json
├── emails.db
├── requirements.txt
├── .env
│
├── database/
│   ├── __init__.py
│   ├── db_connector.py
│   └── models.py
│
├── email_integration/
│   ├── __init__.py
│   ├── credentials.json
│   ├── email_fetcher.py
│   ├── email_parser.py
│   ├── email_sender.py
│   ├── threading_utils.py
│   ├── token.json
│   └── utils.py
│
├── email_integration/utils/
│   ├── __init__.py
│   ├── reply_utils.py
│   └── summarizer.py
│
└── tool_integrations/
    ├── __init__.py
    ├── calendar_integration.py
    ├── credential.json
    ├── search_utils.py
    ├── slack_integration.py
    └── web_search.py

---

⚙️ Setup Instructions

1. Clone the Repository

git clone https://github.com/aditya-gupta/wasserstoff/AiInternTask.git
cd AiInternTask

2. Create Virtual Environment & Install Dependencies

python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt

3. Setup Gmail API Credentials

• Visit Google Cloud Console
• Enable Gmail & Calendar APIs
• Create OAuth client ID and download the credentials.json
• Place it in the credentials/ directory

4. Setup Slack Webhook

• Go to Slack Incoming Webhooks
• Create a new webhook and copy the URL
• Add it to your .env file as SLACK_WEBHOOK_URL=your_webhook_url

5. Run the App

python app.py

---

🧠 Architecture Overview

⟲ Flow:

1. Emails are fetched via Gmail API
2. Each email is parsed and stored in SQLite
3. Email body is passed to the BART LLM for summarization
4. Context-aware reply is generated
5. If scheduling intent is detected, a calendar event is created
6. A Slack notification is sent for every processed email

🗂️ Components

• Gmail API – Email fetching
• Hugging Face BART – Summarization & reply generation
• Google Calendar API – Schedule meeting slots
• Slack API – Notify updates
• SQLite – Store email history and summaries

---

📽️ Video Walkthrough

A short demo video is included demonstrating:

• Running the app
• Fetching and summarizing emails
• Detecting scheduling intent
• Sending Slack notifications and replies

---

Windows pip install -r requirements.txt

### 3. Setup Gmail API Credentials

- Visit [Google Cloud Console](https://console.cloud.google.com/)
- Enable **Gmail** & **Calendar** APIs
- Create OAuth client ID and download the `credentials.json`
- Place it in the `email_integration/` directory

### 4. Setup Slack Webhook

- Go to [Slack Incoming Webhooks](https://api.slack.com/messaging/webhooks)
- Create a new webhook and copy the URL
- Add it to your `.env` file or directly in `slack_integration.py` (for testing only)

### 5. Run the App

```bash
python app.py

---

🧠 Architecture Overview

⟲ Flow:

1. Emails are fetched via Gmail API
2. Each email is parsed and stored in SQLite
3. Email body is passed to the BART LLM for summarization
4. Context-aware reply is generated
5. If scheduling intent is detected, a calendar event is created
6. A Slack notification is sent for every processed email

🗍️ Components

• Gmail API – Email fetching
• Hugging Face BART – Summarization & reply generation
• Google Calendar API – Schedule meeting slots
• Slack API – Notify updates
• SQLite – Store email history and summaries

---

📟️ Video Walkthrough

A short demo video is included (or to be provided via link) demonstrating:

• Running the app
• Fetching and summarizing emails
• Detecting scheduling intent
• Sending Slack notifications and replies

---

This project was created as part of an internship assessment task and is free for non-commercial educational use.

---