GenAI-Week2-Multimodal Agent News Summarizer

2_openai/community_contributions/emmy_news_summariser/README.md

@@ -0,0 +1,157 @@
+# 🧠 Multimodal Agent News Summarizer
+
+An AI-powered system that **aggregates news**, **summarizes using GPT-4o-mini**, and **creates audio briefings** with **MiniMax TTS**.
+Built with the **official OpenAI Agents SDK** for real autonomous decision-making.
+
+---
+
+## 🚀 Features
+
+* 🗞️ **Multi-Source Aggregation** – Fetches and merges RSS feeds from multiple topics
+* 🧠 **AI Summarization** – GPT-4o-mini produces concise (≈300 words) audio-optimized briefs
+* 🔊 **Text-to-Speech** – MiniMax TTS converts summaries to high-quality MP3
+* 🤖 **Autonomous Agents** – Agents decide *when and how* to use tools
+* ⚡ **Async/Await** – Fully asynchronous for speed and scalability
+* 🎨 **Modern Gradio UI** – Simple blue-themed interface
+
+---
+
+## 🧩 Architecture
+
+```
+User → Orchestrator → Autonomous Agents
+                ↓
+   [1] News Aggregator → [2] Summarizer → [3] Audio Generator
+```
+
+Each agent independently chooses its tool:
+
+| 🧩 Agent            | 🧰 Tool                    | 🎯 Purpose                |
+| :------------------ | :------------------------- | :------------------------ |
+| **Aggregator**      | `aggregate_news(topic)`    | Fetch & merge articles    |
+| **Summarizer**      | `summarize_articles(json)` | Create engaging briefings |
+| **Audio Generator** | `synthesize_speech(text)`  | Generate MP3 audio        |
+
+---
+
+<details>
+<summary>⚙️ Installation & Setup</summary>
+
+### Prerequisites
+
+* Python 3.13 +
+* pip
+
+### 1️⃣ Install
+
+```bash
+git clone <repo>
+cd news_summariser
+python -m venv .venv && source .venv/bin/activate
+pip install -e .
+```
+
+### 2️⃣ Environment Variables
+
+Create a `.env` file in the root:
+
+```env
+OPENAI_API_KEY=your_openai_api_key
+MINIMAX_API_KEY=your_minimax_api_key
+```
+
+</details>
+
+---
+
+<details>
+<summary>▶️ Usage</summary>
+
+```bash
+python main.py
+```
+
+Then open the Gradio UI (default → [http://127.0.0.1:7860](http://127.0.0.1:7860)).
+
+**Steps**
+
+1. Pick a topic → Tech | World | Business | Politics | Sports
+2. Click **Submit** to generate a summary + audio briefing
+
+</details>
+
+---
+
+## 🏗️ Project Structure
+
+```
+news_summariser/
+├── news_agents/
+│   ├── news_aggregator.py
+│   ├── summarizer.py
+│   └── audio_generator.py
+├── orchestrator.py
+├── main.py
+├── config/
+└── README.md
+```
+
+---
+
+## 🧠 Tech Stack
+
+| Category             | Technologies           |
+| :------------------- | :--------------------- |
+| **LLM**              | OpenAI GPT-4o-mini     |
+| **TTS**              | MiniMax API            |
+| **Framework**        | OpenAI Agents SDK      |
+| **Web UI**           | Gradio                 |
+| **Async Runtime**    | aiohttp · aiofiles     |
+| **Feeds**            | feedparser             |
+| **Config & Logging** | python-dotenv · loguru |
+
+---
+
+## 📊 Example Output
+
+```bash
+✓ Fetched 15 articles
+✓ Summary created (287 words)
+✓ Audio generated: news_summary_20251109.mp3
+```
+
+📰 **Text Summary:** Engaging 300-word brief (opening hook → top 3 stories → closing)
+🔊 **Audio File:** MP3 briefing ready for listening on the go
+
+---
+
+## 💡 Why This Matters
+
+| Traditional Pipelines     | This Project (Autonomous Agents)         |
+| :------------------------ | :--------------------------------------- |
+| Hard-coded function calls | Agents decide tools autonomously         |
+| Fixed sequence            | Dynamic reasoning + error recovery       |
+| Rigid logic               | Extensible and maintainable architecture |
+
+**Benefits**
+
+* 🔁 Change behavior by editing instructions (not code)
+* 🧩 Easily add new tools or agents
+* 🧱 Fewer hard-coded flows → cleaner design
+
+---
+
+## 🔗 Resources
+
+* 📘 [OpenAI Agents SDK Docs](https://openai.github.io/openai-agents-python/)
+* 🎧 [MiniMax TTS API](https://www.minimaxi.com/)
+* 🎨 [Gradio Docs](https://www.gradio.app/docs)
+
+---
+
+> **Built with the Official OpenAI Agents SDK 🚀**
+> Keep your `.env` file secure — it’s already ignored by Git.
+
+---
+
+Would you like me to include a **short “How to extend this project”** section (e.g., adding new tools or agents) at the end? It makes the README feel even more “developer-friendly” for open-source contributors.

2_openai/community_contributions/emmy_news_summariser/agents/__init__.py

@@ -0,0 +1,14 @@
+"""News agents module for news summarizer application."""
+
+# Our custom agent instances (they import from openai-agents SDK internally)
+from .news_aggregator import news_aggregator_agent, aggregate_news
+from .summarizer import summarizer_agent
+from .audio_generator import audio_generator_agent, synthesize_speech
+
+__all__ = [
+    "news_aggregator_agent",
+    "summarizer_agent",
+    "audio_generator_agent",
+    "aggregate_news",
+    "synthesize_speech",
+]

2_openai/community_contributions/emmy_news_summariser/agents/audio_generator.py

@@ -0,0 +1,95 @@
+"""Audio Generator Agent - Converts text to speech using MiniMax TTS."""
+
+import os
+import aiohttp
+import aiofiles
+from datetime import datetime
+from pathlib import Path
+from typing import Dict, Any
+import uuid
+
+from agents import Agent, function_tool
+
+
+@function_tool
+async def synthesize_speech(
+    text: str, 
+    voice_id: str = "Chinese (Mandarin)_News_Anchor",
+    speed: float = 0.95,
+    pitch: int = -1
+) -> str:
+    """Synthesize speech from text using MiniMax TTS.
+
+    Args:
+        text: Text to convert to speech
+        voice_id: Voice ID to use
+        speed: Speech speed (0.5-2.0)
+        pitch: Pitch adjustment (-20 to 20)
+
+    Returns:
+        Path to the generated audio file
+    """
+    api_key = os.getenv("MINIMAX_API_KEY")
+    if not api_key:
+        raise ValueError("MINIMAX_API_KEY must be set in environment variables")
+
+    url = "https://api.minimaxi.chat/v1/t2a_v2"
+
+    # Generate unique filename to avoid collisions
+    timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+    unique_id = str(uuid.uuid4())[:8]
+    audio_filename = f"news_summary_{timestamp}_{unique_id}.mp3"
+    audio_path = Path(audio_filename).absolute()
+
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json"
+    }
+
+    payload = {
+        "model": "speech-02-hd",
+        "text": text,
+        "voice_setting": {
+            "voice_id": voice_id,
+            "speed": speed,
+            "pitch": pitch,
+            "emotion": "neutral"
+        }
+    }
+
+    # Use aiohttp for async HTTP request
+    async with aiohttp.ClientSession() as session:
+        async with session.post(url, headers=headers, json=payload) as response:
+            if response.status == 200:
+                result = await response.json()
+
+                # MiniMax returns hex-encoded audio in result['data']['audio']
+                if 'data' in result and 'audio' in result['data']:
+                    # Convert hex-encoded audio to bytes
+                    audio_data = bytes.fromhex(result['data']['audio'])
+
+                    # Write audio file asynchronously
+                    async with aiofiles.open(audio_path, 'wb') as f:
+                        await f.write(audio_data)
+
+                    return str(audio_path)
+                else:
+                    raise Exception(f"Unexpected response format: {result}")
+            else:
+                error_text = await response.text()
+                raise Exception(f"Error: {response.status} - {error_text}")
+
+
+# Create the Audio Generator Agent with Gemini via LiteLLM
+AUDIO_GENERATOR_INSTRUCTIONS = """You are an audio generator agent. Your task is to convert text 
+summaries into audio files using the MiniMax TTS API. 
+
+When given text, use the synthesize_speech tool to generate high-quality Chinese audio.
+Return the path to the generated audio file."""
+
+audio_generator_agent = Agent(
+    name="Audio Generator",
+    instructions=AUDIO_GENERATOR_INSTRUCTIONS,
+    tools=[synthesize_speech],
+    model="gpt-4o-mini",  
+)

2_openai/community_contributions/emmy_news_summariser/agents/news_aggregator.py

@@ -0,0 +1,100 @@
+"""News Aggregator Agent - Fetches news from RSS feeds."""
+
+import asyncio
+import aiohttp
+import feedparser
+from typing import Dict, List, Any
+
+from agents import Agent, function_tool
+
+
+FEEDS = {
+    'tech': [
+        'https://techcrunch.com/feed/',
+        'https://www.theverge.com/rss/index.xml',
+        'https://feeds.arstechnica.com/arstechnica/index'
+    ],
+    'world': [
+        'http://feeds.bbci.co.uk/news/world/rss.xml',
+        'https://rss.nytimes.com/services/xml/rss/nyt/World.xml'
+    ],
+    'business': [
+        'http://feeds.bbci.co.uk/news/business/rss.xml',
+        'https://www.cnbc.com/id/100003114/device/rss/rss.html'
+    ],
+    'politics': [
+        'https://rss.nytimes.com/services/xml/rss/nyt/Politics.xml',
+        'http://feeds.bbci.co.uk/news/politics/rss.xml'
+    ],
+    'sports': [
+        'https://rss.nytimes.com/services/xml/rss/nyt/Sports.xml',
+        'http://feeds.bbci.co.uk/sport/rss.xml'
+    ]
+}
+
+
+async def fetch_feed(session: aiohttp.ClientSession, url: str) -> List[Dict[str, str]]:
+    """Fetch articles from a single RSS feed.
+
+    Args:
+        session: aiohttp client session
+        url: RSS feed URL
+
+    Returns:
+        List of article dictionaries with title, summary, link, and published date
+    """
+    try:
+        async with session.get(url, timeout=10) as response:
+            content = await response.text()
+            feed = feedparser.parse(content)
+            return [
+                {
+                    'title': entry.title,
+                    'summary': entry.get('summary', ''),
+                    'link': entry.link,
+                    'published': entry.get('published', '')
+                }
+                for entry in feed.entries[:3]
+            ]
+    except Exception as e:
+        print(f"Error fetching {url}: {e}")
+        return []
+
+
+@function_tool
+async def aggregate_news(topic: str, num_sources: int = 5) -> Dict[str, Any]:
+    """Aggregate news from RSS feeds concurrently.
+
+    Args:
+        topic: News topic (tech, world, business, politics, sports)
+        num_sources: Number of RSS sources to fetch from
+
+    Returns:
+        Dictionary containing list of articles
+    """
+    feed_urls = FEEDS.get(topic, FEEDS['tech'])[:num_sources]
+
+    async with aiohttp.ClientSession() as session:
+        tasks = [fetch_feed(session, url) for url in feed_urls]
+        results = await asyncio.gather(*tasks)
+
+    # Flatten results
+    articles = [article for feed_articles in results for article in feed_articles]
+
+    return {"articles": articles[:15]}
+
+
+# Create the News Aggregator Agent with Gemini via LiteLLM
+NEWS_AGGREGATOR_INSTRUCTIONS = """You are a news aggregator agent. Your task is to fetch and aggregate 
+news articles from RSS feeds for a given topic. 
+
+When asked to get news on a topic, use the aggregate_news tool with the topic name 
+(tech, world, business, politics, or sports). The tool will fetch articles from multiple sources 
+and return a list of the most recent articles."""
+
+news_aggregator_agent = Agent(
+    name="News Aggregator",
+    instructions=NEWS_AGGREGATOR_INSTRUCTIONS,
+    tools=[aggregate_news],
+    model="gpt-4o-mini", 
+)