revert to last change

README.md

@@ -1,143 +1,50 @@
----
-title: AI Research Paper Generator
-emoji: 📚
-colorFrom: yellow
-colorTo: purple
-sdk: gradio
-sdk_version: 5.43.1
-app_file: app.py
-pinned: false
-short_description: AI-powered research paper generator for academic writing
----
+# Research Draft Generator
 
-# 📚 AI Research Paper Generator
+An AI-powered research paper generator that uses a multi-agent workflow to create comprehensive academic papers.
 
-An intelligent tool designed to assist researchers, students, and academics in generating comprehensive research paper structures, analyzing paper strength, and creating citations in multiple academic formats.
+## Features
 
-## 🚀 Features
+- **Multi-Agent AI System**: Specialized agents for topic analysis, research retrieval, outline building, draft writing, and citation formatting
+- **Human-in-the-Loop**: Interactive feedback system for outline review and approval
+- **Automated Workflow**: Streamlined process from topic to final draft
+- **PDF Export**: Generate downloadable PDFs of completed research papers
 
-### 📝 **Paper Structure Generation**
-- **Research Area Selection**: Choose from 10+ research domains including Machine Learning, Computer Vision, NLP, Robotics, and more
-- **AI-Powered Topic Generation**: Get intelligent topic suggestions based on your research area and keywords
-- **Comprehensive Outlines**: Generate complete paper structures with detailed sections and subsections
-- **Abstract Generation**: Create compelling abstracts that summarize your research focus
+## How It Works
 
-### 🔍 **Paper Strength Analysis**
-- **Content Assessment**: Analyze the completeness and strength of your research papers
-- **Scoring System**: Get a 1-10 score with detailed feedback
-- **Improvement Recommendations**: Receive specific suggestions for enhancing your paper
-- **Structure Analysis**: Identify missing sections and content gaps
+1. **Topic Input**: Enter your research topic
+2. **AI Analysis**: The system analyzes and refines your topic
+3. **Outline Generation**: Creates a structured research paper outline
+4. **Human Review**: You review and approve/revise the outline
+5. **Draft Generation**: AI generates the full research paper
+6. **PDF Export**: Download your completed paper
 
-### 📖 **Citation Generator**
-- **Multiple Formats**: Generate citations in APA, MLA, Chicago, IEEE, and Harvard styles
-- **Academic Standards**: Ensure your references meet publication requirements
-- **Easy Input**: Simple form-based citation creation
+## Technology Stack
 
-### 🎯 **Research Areas Supported**
-- Machine Learning & AI
-- Computer Vision
-- Natural Language Processing
-- Robotics & Automation
-- Data Science & Analytics
-- Cybersecurity
-- Quantum Computing
-- Bioinformatics
-- Climate Science
-- Economics & Social Sciences
+- **AI Models**: OpenAI GPT-4 for natural language generation
+- **Workflow Engine**: LangGraph for orchestrating multi-agent workflows
+- **User Interface**: Gradio for web-based interaction
+- **Document Generation**: ReportLab for PDF creation
 
-## 🛠️ How to Use
+## Usage
 
-### 1. **Generate Paper Structure**
-1. Select your research area from the dropdown
-2. Enter relevant keywords (comma-separated)
-3. Choose your paper type (Research, Review, Survey, etc.)
-4. Click "Generate Paper Structure"
-5. Get a complete outline with abstract, methodology, and next steps
+1. Enter your research topic in the input field
+2. Click "Generate Outline" to start the process
+3. Review the generated outline
+4. Approve, request revisions, or reject the outline
+5. Once approved, the system will generate the full draft
+6. Download your research paper as a PDF
 
-### 2. **Analyze Paper Strength**
-1. Paste your existing paper content
-2. Click "Analyze Paper"
-3. Receive detailed feedback and improvement suggestions
-4. Get a comprehensive score and recommendations
+## Environment Variables
 
-### 3. **Generate Citations**
-1. Enter paper title, authors, year, and journal
-2. Click "Generate Citations"
-3. Copy citations in your preferred academic format
-
-## 🔧 Technical Details
-
-- **Framework**: Gradio 5.43.1
-- **Language**: Python 3.8+
-- **Architecture**: Modular design with separate components for each functionality
-- **AI Integration**: Ready for integration with actual AI models (currently uses intelligent templates)
-
-## 🚀 Getting Started
-
-### Local Development
+Set your OpenAI API key in the environment:
 ```bash
-# Clone the repository
-git clone <your-repo-url>
-cd Research_Draft_Generator
-
-# Install dependencies
-pip install -r requirements.txt
-
-# Run the application
-python app.py
+OPENAI_API_KEY=your_api_key_here
 ```
 
-### Hugging Face Spaces
-1. **Fork this Space** to your Hugging Face account
-2. **Customize** the app.py file as needed
-3. **Deploy** automatically with Hugging Face's infrastructure
-4. **Share** your Space with the research community
-
-## 📊 Use Cases
-
-- **Students**: Generate research paper outlines for academic assignments
-- **Researchers**: Structure complex research projects and proposals
-- **Academics**: Analyze and improve existing papers before submission
-- **Writers**: Create citations in multiple formats for publications
-- **Educators**: Teach research methodology and paper structure
-
-## 🔮 Future Enhancements
-
-- **Real AI Model Integration**: Connect with GPT, Claude, or other LLMs
-- **Multi-Language Support**: Generate papers in different languages
-- **Collaborative Features**: Multi-user paper editing and review
-- **Export Options**: PDF, LaTeX, and Word document generation
-- **Reference Management**: Integration with Zotero, Mendeley, etc.
-- **Plagiarism Detection**: Built-in content originality checking
-
-## 🤝 Contributing
-
-We welcome contributions! Here's how you can help:
-
-1. **Fork** the repository
-2. **Create** a feature branch
-3. **Make** your changes
-4. **Test** thoroughly
-5. **Submit** a pull request
-
-## 📄 License
-
-This project is open source and available under the [MIT License](LICENSE).
-
-## 🙏 Acknowledgments
-
-- Built with [Gradio](https://gradio.app/) for beautiful web interfaces
-- Inspired by the research community's need for better writing tools
-- Designed for academic excellence and research productivity
-
-## 📞 Support
-
-- **Issues**: Report bugs and request features on GitHub
-- **Discussions**: Join community discussions about research writing
-- **Documentation**: Check our comprehensive guides and tutorials
+## License
 
----
+This project is for research and educational purposes.
 
-**Ready to revolutionize your research paper writing? Try the AI Research Paper Generator today! 🚀**
+## Author
 
-*Built with ❤️ for the research community*
+Anushka0007

app.py

@@ -315,5 +315,11 @@ with gr.Blocks(css=css, title="Research Draft Generator") as demo:
 
 
 if __name__ == "__main__":
+    # Hugging Face Spaces provides PORT; bind to 0.0.0.0
     port = int(os.getenv("PORT", 7860))
-    demo.queue().launch(server_name="0.0.0.0", server_port=port)
+    demo.queue().launch(
+        server_name="0.0.0.0", 
+        server_port=port, 
+        share=False,
+        reload=False  # Disable reload for production deployment
+    )

orchestrator/research_orchestrator.py

@@ -545,20 +545,17 @@ class AStreamCallbackHandler(BaseCallbackHandler):
 
 def human_feedback_node(state: ResearchState) -> ResearchState:
     """
-    Human feedback node using proper LangGraph interrupt mechanism
+    Human feedback node using current LangGraph interrupt() function
     """
-    # Check if we have an outline to review
-    if not state.get("outline"):
-        state["outline_approved"] = False
-        state["revision_feedback"] = "No outline available for review"
-        return state
-    
-    # Set the state to indicate we're waiting for human feedback
-    state["waiting_for_feedback"] = True
-    state["workflow_status"] = "paused_for_feedback"
+    # Use the current LangGraph interrupt() function
+    # This will pause execution and wait for human input
+    human_input = interrupt(
+        "Please review the outline and provide your feedback. "
+        "Options: 'approve', 'revise [feedback]', or 'reject'"
+    )
     
-    # Return the state to trigger the interrupt in the workflow
-    return state
+    # Process the human input directly in this node
+    return process_human_feedback_inline(state, human_input)
 
 
 def process_human_feedback_inline(state: ResearchState, human_input: str) -> ResearchState:
@@ -569,8 +566,7 @@ def process_human_feedback_inline(state: ResearchState, human_input: str) -> Res
     feedback_text = str(human_input).strip()
     feedback_lower = feedback_text.lower()
     
-    # Reset the waiting state
-    state["waiting_for_feedback"] = False
+    # print(f"\n Processing human feedback: '{feedback_text}'")
     
     # Check for topic change requests
     topic_change_keywords = [
@@ -589,6 +585,7 @@ def process_human_feedback_inline(state: ResearchState, human_input: str) -> Res
         # Extract new topic from feedback
         new_topic = extract_new_topic_from_feedback(feedback_text)
         if new_topic:
+            # print(f"🔄 Topic change detected! New topic: '{new_topic}'")
             state["topic"] = new_topic
             state["refined_topic"] = None  # Reset refined topic
             state["research_notes"] = None  # Reset research notes
@@ -599,6 +596,7 @@ def process_human_feedback_inline(state: ResearchState, human_input: str) -> Res
             state["workflow_status"] = "topic_change"
             return state
         else:
+            # print("❌ Topic change requested but no new topic found. Please specify the new topic.")
             state["outline_approved"] = False
             state["revision_feedback"] = "Please specify the new topic you want to research"
             return state
@@ -607,11 +605,12 @@ def process_human_feedback_inline(state: ResearchState, human_input: str) -> Res
     if feedback_lower.startswith("approve") or feedback_lower in ["a"]:
         state["outline_approved"] = True
         state["feedback"] = ""
-        state["workflow_status"] = "approved"
+        # print(" Outline approved! Proceeding to draft writing...")
         
     elif feedback_lower.startswith("reject") or feedback_lower in ["r", "no", "n", "abort"]:
         state["outline_approved"] = False
         state["workflow_status"] = "aborted"
+        # print("Workflow aborted by user.")
         
     elif feedback_lower.startswith("revise") or feedback_lower.startswith("v"):
         # Extract feedback from the input
@@ -630,6 +629,7 @@ def process_human_feedback_inline(state: ResearchState, human_input: str) -> Res
             # Extract new topic from the revision feedback
             new_topic = extract_new_topic_from_feedback(feedback)
             if new_topic:
+                # print(f"🔄 Topic change detected in revision feedback! New topic: '{new_topic}'")
                 state["topic"] = new_topic
                 state["refined_topic"] = None  # Reset refined topic
                 state["research_notes"] = None  # Reset research notes
@@ -640,21 +640,24 @@ def process_human_feedback_inline(state: ResearchState, human_input: str) -> Res
                 state["workflow_status"] = "topic_change"
                 return state
             else:
+                # print("❌ Topic change requested in revision but no new topic found.")
                 state["outline_approved"] = False
                 state["revision_feedback"] = "Please specify the new topic you want to research"
                 return state
         else:
-            # Regular revision request
+            # Regular revision feedback
             state["outline_approved"] = False
             state["revision_feedback"] = feedback
-            state["workflow_status"] = "revision_requested"
-    
+            # print(f" Revision requested with feedback: {feedback}")
+        
     else:
-        # Unrecognized input
+        # print(f" Invalid response: '{feedback_text}'. Defaulting to revision request.")
         state["outline_approved"] = False
-        state["revision_feedback"] = f"Unrecognized input: '{feedback_text}'. Please use: approve, reject, or revise [feedback]"
+        state["revision_feedback"] = "Please improve the outline structure and content"
     
     return state
+
+
 def extract_new_topic_from_feedback(feedback_text: str) -> str:
     """
     Extract new topic from feedback text using various patterns

requirements.txt

@@ -1,25 +1,9 @@
-gradio==5.43.1
-transformers
-torch
-numpy
-pandas
-scikit-learn
-matplotlib
-seaborn
-plotly
-streamlit
-langchain
-openai
-langchain-community
-langchain-openai
-anthropic
-tiktoken
-python-dotenv
-requests
-beautifulsoup4
-markdown
-fastapi
-uvicorn
-pydantic>=2
-langgraph>=0.2.30
-reportlab
+gradio>=4.0.0
+langgraph>=0.2.0
+langchain>=0.2.0
+langchain-openai>=0.2.0
+openai>=1.0.0
+python-dotenv>=1.0.0
+reportlab>=4.0.0
+asyncio
+typing