initial commit

README.md

@@ -1,419 +1,20 @@
-# 🚀 LangGraph Post Generator
-
-An intelligent LinkedIn post generator powered by **LangGraph** and **Azure OpenAI** that creates engaging, professional content by intelligently routing queries through web search and content generation workflows.
-
-## 📋 Table of Contents
-
-- [Overview](#overview)
-- [Features](#features)
-- [Architecture](#architecture)
-- [Tech Stack](#tech-stack)
-- [Installation](#installation)
-- [Configuration](#configuration)
-- [Usage](#usage)
-- [Project Structure](#project-structure)
-- [API Documentation](#api-documentation)
-- [Technical Details](#technical-details)
-- [Testing](#testing)
-- [Contributing](#contributing)
-- [License](#license)
-
-## 🌟 Overview
-
-The **LangGraph Post Generator** is an AI-powered application that transforms simple topic inputs into compelling LinkedIn posts. It leverages the power of **LangGraph** to create an intelligent workflow that decides whether to search for additional context or directly generate content based on the input query.
-
-### Key Capabilities:
-- **Intelligent Routing**: Automatically determines if web search is needed for better context
-- **Web Search Integration**: Uses DuckDuckGo to gather recent, relevant information
-- **Professional Content Generation**: Creates LinkedIn-optimized posts with proper tone and structure
-- **Real-time Processing**: Streamlit-based web interface for immediate results
-- **Comprehensive Logging**: Detailed logging for debugging and monitoring
-
-## ✨ Features
-
-- 🧠 **Smart Query Routing**: AI-powered decision making for content enhancement
-- 🔍 **Real-time Web Search**: Integration with DuckDuckGo for up-to-date information
-- 📝 **Professional Content Creation**: LinkedIn-optimized posts with engaging tone
-- 🎯 **Context-Aware Generation**: Uses web search results to create informed content
-- 📊 **Enhanced Streamlit Interface**: Beautiful, modern UI with sidebar configurations
-- ⚙️ **Configurable Settings**: Customize post length, tone, hashtags, and call-to-action
-- 🎨 **Professional Styling**: LinkedIn-inspired design with gradients and modern components
-- 📈 **Post Analytics**: Track generated posts and usage statistics
-- 💡 **Smart Suggestions**: Hashtag recommendations and example topics
-- 🔧 **Modular Architecture**: Clean, maintainable codebase with separation of concerns
-- 📋 **Comprehensive Logging**: Detailed logs for monitoring and debugging
-- 🔄 **Flexible Workflow**: LangGraph-powered state management
-
-## 🏗️ Architecture
-
-The application follows a **LangGraph-based workflow** with the following components:
-
-```mermaid
-graph TD
-    A[User Input] --> B[Question Router]
-    B --> C{Route Decision}
-    C -->|Need Context| D[Query Transform]
-    C -->|Direct Generation| E[Content Generator]
-    D --> F[Web Search]
-    F --> E
-    E --> G[LinkedIn Post Output]
-```
-
-### Workflow Steps:
-1. **Input Processing**: User provides a topic through Streamlit interface
-2. **Question Routing**: AI determines if web search is needed
-3. **Query Transformation** (if needed): Optimizes the query for web search
-4. **Web Search** (if needed): Gathers relevant context using DuckDuckGo
-5. **Content Generation**: Creates professional LinkedIn post using Azure OpenAI
-6. **Output**: Returns formatted LinkedIn post to user
-
-## 🛠️ Tech Stack
-
-### Core Technologies:
-- **Python 3.8+** - Primary programming language
-- **LangGraph** - Workflow orchestration and state management
-- **LangChain** - LLM framework and prompt management
-- **Azure OpenAI** - Language model for content generation
-- **Streamlit** - Web application framework
-- **DuckDuckGo Search** - Web search functionality
-
-### Dependencies:
-- `streamlit` - Web interface
-- `langchain` - LLM framework
-- `langchain-core` - Core LangChain functionality
-- `langchain-community` - Community tools and utilities
-- `langchain-openai` - Azure OpenAI integration
-- `langgraph` - Graph-based workflow management
-- `langsmith` - LangChain monitoring and debugging
-- `duckduckgo-search` - Web search capabilities
-- `python-dotenv` - Environment variable management
-
-## 📦 Installation
-
-### Prerequisites
-- Python 3.8 or higher
-- Azure OpenAI API access
-- Git (for cloning the repository)
-
-### Step 1: Clone the Repository
-```bash
-git clone https://github.com/your-username/langgraph-post-generator.git
-cd langgraph-post-generator/post-generator-agent
-```
-
-### Step 2: Create Virtual Environment
-```bash
-# Create virtual environment
-python -m venv venv
-
-# Activate virtual environment
-# On Windows:
-venv\Scripts\activate
-# On macOS/Linux:
-source venv/bin/activate
-```
-
-### Step 3: Install Dependencies
-```bash
-pip install -r requirements.txt
-```
-
-### Step 4: Run the Application
-```bash
-streamlit run app.py
-```
-
-### Step 5: Configure Credentials
-1. Open the app in your browser
-2. Go to the sidebar configuration
-3. Enter your Azure OpenAI credentials:
-   - **API Key**: Your Azure OpenAI API key
-   - **Endpoint**: Your Azure OpenAI endpoint (e.g., `https://your-resource.openai.azure.com/`)
-   - **Deployment Name**: Your model deployment name (e.g., `gpt-35-turbo`)
-   - **API Version**: API version (default: `2024-02-01`)
-4. Click "Save Credentials"
-5. Test the connection using the "Test Connection" button
-
-## ⚙️ Configuration
-
-### Azure OpenAI Setup
-
-1. **Create Azure OpenAI Resource**:
-   - Go to [Azure Portal](https://portal.azure.com)
-   - Create a new OpenAI resource
-   - Note your endpoint and API key
-
-2. **Deploy a Model**:
-   - Deploy a GPT model (GPT-3.5-turbo or GPT-4)
-   - Note your deployment name
-
-3. **Update Configuration**:
-   - Add your credentials to the `.env` file
-   - Ensure the API version matches your Azure OpenAI service
-
-### Environment Variables
-
-| Variable | Description | Example |
-|----------|-------------|---------|
-| `AZURE_OPENAI_API_KEY` | Your Azure OpenAI API key | `abc123...` |
-| `AZURE_OPENAI_ENDPOINT` | Your Azure OpenAI endpoint URL | `https://example.openai.azure.com/` |
-| `OPENAI_API_VERSION` | API version for Azure OpenAI | `2024-02-01` |
-| `AZURE_DEPLOYMENT` | Name of your model deployment | `gpt-35-turbo` |
-
-## 🚀 Usage
-
-### Running the Application
-
-1. **Start the Streamlit App**:
-```bash
-streamlit run app.py
-```
-
-2. **Access the Interface**:
-   - Open your browser to `http://localhost:8501`
-   - Configure settings in the sidebar (post length, tone, hashtags)
-   - Enter a topic in the main input area
-   - Add optional context for better results
-   - Click "Generate LinkedIn Post" to create content
-   - View the generated post in a styled container
-   - Use example topics for quick testing
-
-### UI Features
-
-- **📊 Sidebar Configuration**: 
-  - **🔑 Azure Credentials Management**: 
-    - Manual credential input with password masking
-    - Real-time credential validation for Azure OpenAI format
-    - Connection testing functionality
-    - Credential persistence across sessions
-  - API status indicators with masked API keys
-  - Post length options (Short, Medium, Long)
-  - Tone selection (Professional, Casual, Enthusiastic, etc.)
-  - Hashtag and call-to-action toggles
-  - Hashtag suggestions based on topic
-
-- **🎯 Main Interface**:
-  - Large text area for topic input
-  - Optional context expander
-  - Generate and Clear buttons
-  - Tips for better posts
-  - Post analytics tracking
-  - Example topics for quick testing
-
-- **📄 Output Display**:
-  - Styled output container
-  - Success/error messages
-  - Copy to clipboard functionality
-
-### Example Usage
-
-**Input**: "Artificial Intelligence trends in 2024"
-
-**Output**: A professionally formatted LinkedIn post including:
-- Current AI trends and insights
-- Professional tone suitable for LinkedIn
-- Actionable insights for professionals
-- Engaging call-to-action or discussion points
-
-### Command Line Testing
-
-You can also test individual components:
-
-```bash
-# Test LLM configuration
-python tests/llm_test.py
-
-# Test web search functionality
-python tests/search_test.py
-
-# Test the complete agent workflow
-python utils/run_agent.py
-```
-
-## 📁 Project Structure
-
-```
-post-generator-agent/
-├── 📄 app.py                     # Main Streamlit application
-├── 📁 config/
-│   ├── configs.py                # Environment configuration
-│   └── generation_config.py      # Generation settings and prompts
-├── 📁 helper/
-│   ├── configure_llm.py          # Azure OpenAI LLM setup
-│   ├── graphs.py                 # LangGraph workflow definition
-│   ├── model_load.py             # Graph state model
-│   └── web_search_agent.py       # DuckDuckGo search configuration
-├── 📁 loggers/
-│   └── logger.py                 # Logging configuration
-├── 📁 template/
-│   ├── response_prompt.py        # Content generation prompts
-│   ├── router_prompt.py          # Query routing prompts
-│   └── transform_prompt.py       # Query transformation prompts
-├── 📁 tests/
-│   ├── llm_test.py              # LLM functionality tests
-│   └── search_test.py           # Search functionality tests
-├── 📁 utils/
-│   ├── credential_manager.py     # Azure credentials management
-│   ├── generate_content.py       # Content generation logic
-│   ├── query_transform.py        # Query transformation logic
-│   ├── question_route.py         # Question routing logic
-│   ├── run_agent.py             # Main agent execution
-│   └── search_web_content.py     # Web search implementation
-├── 📄 requirements.txt           # Python dependencies
-└── 📄 README.md                 # This documentation
-```
-
-## 📖 API Documentation
-
-### Core Functions
-
-#### `execute_agent(query: str) -> str`
-Main function to execute the LangGraph workflow.
-
-**Parameters:**
-- `query` (str): The topic or question for LinkedIn post generation
-
-**Returns:**
-- `str`: Generated LinkedIn post content
-
-**Example:**
-```python
-from utils.run_agent import execute_agent
-
-result = execute_agent("Latest trends in machine learning")
-print(result)
-```
-
-#### `route_question(state: dict) -> str`
-Routes questions to appropriate workflow paths.
-
-**Parameters:**
-- `state` (dict): Current graph state containing the question
-
-**Returns:**
-- `str`: Routing decision ("websearch" or "generate")
-
-#### `transform_query(state: dict) -> dict`
-Transforms user queries for optimal web search.
-
-**Parameters:**
-- `state` (dict): Current graph state
-
-**Returns:**
-- `dict`: Updated state with optimized search query
-
-#### `web_search(state: dict) -> dict`
-Performs web search using DuckDuckGo.
-
-**Parameters:**
-- `state` (dict): Current graph state with search query
-
-**Returns:**
-- `dict`: Updated state with search results context
-
-#### `generate(state: dict) -> dict`
-Generates final LinkedIn post content.
-
-**Parameters:**
-- `state` (dict): Current graph state with question and optional context
-
-**Returns:**
-- `dict`: Updated state with generated content
-
-## 🔧 Technical Details
-
-### LangGraph Workflow
-
-The application uses **LangGraph** to create a stateful workflow:
-
-```python
-# Graph state definition
-class GraphState(TypedDict):
-    question: str      # Original user question
-    generation: str    # Generated LinkedIn post
-    search_query: str  # Optimized search query
-    context: str       # Web search results
-```
-
-### Prompt Engineering
-
-The system uses three specialized prompt templates:
-
-1. **Router Prompt**: Determines routing strategy
-2. **Transform Prompt**: Optimizes queries for web search
-3. **Response Prompt**: Generates LinkedIn-optimized content
-
-### Logging System
-
-Comprehensive logging with:
-- **File Logging**: Timestamped log files in `logs/` directory
-- **Console Logging**: Real-time feedback during execution
-- **Multiple Log Levels**: DEBUG, INFO, WARNING, ERROR
-
-### Error Handling
-
-Robust error handling including:
-- Graceful fallbacks when web search fails
-- Retry mechanisms for API calls
-- Comprehensive error logging
-
-## 🧪 Testing
-
-### Running Tests
-
-```bash
-# Test Azure OpenAI connection
-python tests/llm_test.py
-
-# Test web search functionality
-python tests/search_test.py
-
-# Test complete workflow
-python utils/run_agent.py
-```
-
-### Test Coverage
-
-- **LLM Integration**: Validates Azure OpenAI connectivity
-- **Web Search**: Tests DuckDuckGo search functionality
-- **End-to-End**: Complete workflow validation
-
-## 🤝 Contributing
-
-We welcome contributions! Please follow these steps:
-
-1. **Fork the Repository**
-2. **Create Feature Branch**: `git checkout -b feature/amazing-feature`
-3. **Commit Changes**: `git commit -m 'Add amazing feature'`
-4. **Push to Branch**: `git push origin feature/amazing-feature`
-5. **Open Pull Request**
-
-### Development Guidelines
-
-- Follow PEP 8 style guidelines
-- Add comprehensive docstrings
-- Include tests for new features
-- Update documentation as needed
-
-## 📄 License
-
-This project is licensed under the MIT License. See the `LICENSE` file for details.
-
-## 🆘 Support
-
-For support and questions:
-
-- **Issues**: Open an issue on GitHub
-- **Documentation**: Check this README and code comments
-- **Community**: Join our discussions
-
-## 🙏 Acknowledgments
+---
+title: LinkedIn Post Generator
+emoji: 🚀
+colorFrom: blue
+colorTo: green
+sdk: docker
+pinned: false
+license: mit
+---
 
-- **LangChain Team** for the excellent LLM framework
-- **Azure OpenAI** for powerful language models
-- **Streamlit** for the intuitive web framework
-- **DuckDuckGo** for search capabilities
+# 🚀 LinkedIn Post Generator
 
----
+An AI-powered LinkedIn post generator built with **LangGraph**, **Azure OpenAI**, and **Streamlit**.  
+It intelligently decides whether to search the web for fresh content before generating professional LinkedIn posts.
 
-**Built with ❤️ using LangGraph and Azure OpenAI**
+## Features
+- Smart query routing with LangGraph
+- DuckDuckGo real-time web search
+- Azure OpenAI content generation
+- Streamlit web interface