Research analyzer with bedrock

README.md

@@ -20,10 +20,11 @@ By [@joaomdmoura](https://x.com/joaomdmoura).
 - [Create Instagram Post](https://github.com/joaomdmoura/crewAI-examples/tree/main/instagram_post)
 - [Markdown Validator](https://github.com/joaomdmoura/crewAI-examples/tree/main/markdown_validator)
 - [Using Azure OpenAI API](https://github.com/joaomdmoura/crewAI-examples/tree/main/azure_model)
+- [Research Paper Analyzer with Bedrock](https://github.com/crewAIInc/crewAI-examples/tree/main/research_analyzer_with_bedrock)
 
 Starting your own example
   - [Starter Template](https://github.com/joaomdmoura/crewAI-examples/tree/main//starter_template)
 ### Advanced Examples
 - [Stock Analysis](https://github.com/joaomdmoura/crewAI-examples/tree/main/stock_analysis)
 - [Landing Page Generator](https://github.com/joaomdmoura/crewAI-examples/tree/main/landing_page_generator)
-- [CrewAI + LangGraph](https://github.com/joaomdmoura/crewAI-examples/tree/main/CrewAI-LangGraph)
+- [CrewAI + LangGraph](https://github.com/joaomdmoura/crewAI-examples/tree/main/CrewAI-LangGraph)

research_analyzer_with_bedrock/README.md

@@ -0,0 +1,71 @@
+# Research Paper Analyzer
+
+A sophisticated research paper analysis system powered by [CrewAI](https://crewai.com) and [Amazon Bedrock](https://aws.amazon.com/bedrock/) that analyzes academic papers and generates comprehensive reports with executive summaries, key insights, and actionable recommendations.
+
+## Overview
+
+The Research Paper Analyzer uses a team of specialized AI agents to analyze academic papers through several phases:
+
+1. **Technical Analysis**: Detailed examination of methodology, statistical approaches, and results
+2. **Literature Review**: Contextualization within the broader academic landscape
+3. **Summary Creation**: Synthesis of key findings and contributions
+4. **Final Report Generation**: Comprehensive report with insights and recommendations
+
+### High-Level Architecture
+
+The system operates through four specialized agents:
+
+- **Research Director**: Coordinates analysis and ensures quality
+- **Technical Expert**: Analyzes methodologies and statistical significance
+- **Literature Specialist**: Reviews research context and connections
+- **Summary Writer**: Creates clear, accessible summaries
+
+## Installation
+
+### Prerequisites
+
+- Python 3.10+
+- AWS Account with Amazon Bedrock access
+- An AWS IAM role with permissions to access Amazon Bedrock, configured following [AWS IAM security best practices](https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html)
+- [Access to Claude-3.5 Sonnet model](https://docs.aws.amazon.com/bedrock/latest/userguide/model-access-modify.html) in Amazon Bedrock
+
+### Setup
+
+1. Clone the repository:
+```bash
+git clone https://github.com/crewAIInc/crewAI-examples.git
+cd research_analyzer_with_bedrock
+```
+2. Install dependencies using Poetry:
+```bash
+poetry install --no-root
+```
+3. Create a .env file in the project root:
+```bash
+AWS_REGION_NAME=us-west-2
+BEDROCK_MODEL_ID=bedrock/anthropic.claude-3-5-sonnet-20240620-v1:0
+MAX_OUTPUT_TOKENS=4096
+TEMPERATURE=0.7
+```
+### Usage
+
+Analyze a research paper using either method:
+Direct Python execution:
+
+    poetry run python main.py path/to/your/paper.pdf
+
+Using CrewAI CLI:
+
+    poetry run crew run --pdf_path path/to/your/paper.pdf
+
+Features
+
+    Comprehensive Analysis: Multi-faceted examination of research papers
+    AWS Bedrock Integration: Leverages powerful Claude-3 model
+    Modular Architecture: Separate specialized agents for different aspects
+    PDF Processing: Automatic extraction and processing of PDF papers
+    Structured Output: Clear, organized analysis reports
+
+## License
+
+This project is licensed under the MIT License - see the LICENSE file for details.

research_analyzer_with_bedrock/agents.py

@@ -0,0 +1,81 @@
+from crewai import Agent
+from utils.Bedrock import load_bedrock_llm
+
+# Initialize Bedrock LLM instances for all agents
+bedrock_llm = load_bedrock_llm()
+
+
+class ResearchAgents:
+    """
+    Factory class for creating specialized research analysis agents.
+    Each agent is configured with specific roles, goals, and expertise
+    for different aspects of research paper analysis.
+    """
+
+    @staticmethod
+    def get_research_director() -> Agent:
+        """
+        Creates a Research Director agent responsible for overseeing the entire
+        analysis process and ensuring quality of outputs.
+
+        Returns:
+            Agent: Configured Research Director agent instance
+        """
+        return Agent(
+            role='Research Director',
+            goal='Coordinate research analysis and ensure high-quality comprehensive output',
+            backstory='Former Research Director at top universities with 15+ years of experience in managing research teams and projects',
+            verbose=True,
+            llm=bedrock_llm
+        )
+
+    @staticmethod
+    def get_technical_expert() -> Agent:
+        """
+        Creates a Technical Expert agent specialized in analyzing methodologies,
+        statistical approaches, and technical aspects of research papers.
+
+        Returns:
+            Agent: Configured Technical Expert agent instance
+        """
+        return Agent(
+            role='Technical Analysis Expert',
+            goal='Analyze technical methodologies, results, and statistical significance in research papers',
+            backstory='PhD in Data Science with 10+ years experience in research analysis and statistical methods',
+            verbose=True,
+            llm=bedrock_llm
+        )
+
+    @staticmethod
+    def get_literature_specialist() -> Agent:
+        """
+        Creates a Literature Specialist agent focused on contextualizing research
+        within the broader academic landscape and identifying connections.
+
+        Returns:
+            Agent: Configured Literature Specialist agent instance
+        """
+        return Agent(
+            role='Literature Review Specialist',
+            goal='Analyze research context and identify connections with existing literature',
+            backstory='Academic librarian with expertise in systematic reviews and meta-analyses',
+            verbose=True,
+            llm=bedrock_llm
+        )
+
+    @staticmethod
+    def get_summary_writer() -> Agent:
+        """
+        Creates a Summary Writer agent specialized in distilling complex
+        research into clear, accessible summaries.
+
+        Returns:
+            Agent: Configured Summary Writer agent instance
+        """
+        return Agent(
+            role='Summary Writer',
+            goal='Create clear, concise summaries of complex research papers',
+            backstory='Science communication expert with experience in top scientific journals',
+            verbose=True,
+            llm=bedrock_llm
+        )

research_analyzer_with_bedrock/main.py

@@ -0,0 +1,128 @@
+from crewai import Crew
+from agents import ResearchAgents
+from tasks import ResearchTasks
+from utils.PdfReader import PdfReader
+import sys
+
+
+class ResearchAnalyzerCrew:
+    """
+    Orchestrates the research paper analysis process using a crew of AI agents.
+    Manages the workflow of technical analysis, literature review, summarization,
+    and final report generation.
+    """
+
+    def __init__(self, paper_content: str):
+        """
+        Initialize the analyzer with research paper content.
+
+        Args:
+            paper_content (str): The extracted text content of the research paper
+        """
+        self.paper_content = paper_content
+
+    def run(self) -> str:
+        """
+        Executes the complete research paper analysis workflow.
+
+        The workflow consists of four main steps:
+        1. Technical analysis of the paper's methodology and results
+        2. Literature review and context analysis
+        3. Creation of a comprehensive summary
+        4. Generation of a final report with insights and recommendations
+
+        Returns:
+            str: The final analysis report
+        """
+        # Initialize factory classes for agents and tasks
+        agents = ResearchAgents()
+        tasks = ResearchTasks()
+
+        # Initialize specialized agents for different aspects of analysis
+        research_director = agents.get_research_director()
+        technical_expert = agents.get_technical_expert()
+        literature_specialist = agents.get_literature_specialist()
+        summary_writer = agents.get_summary_writer()
+
+        # Define sequential analysis tasks with dependencies
+        technical_analysis_task = tasks.analyze_technical_components(
+            agent=technical_expert,
+            paper_content=self.paper_content
+        )
+
+        literature_review_task = tasks.review_literature_context(
+            agent=literature_specialist,
+            paper_content=self.paper_content,
+            technical_analysis="{{ technical_analysis_task.output }}"  # Depends on technical analysis
+        )
+
+        summary_task = tasks.create_research_summary(
+            agent=summary_writer,
+            technical_analysis="{{ technical_analysis_task.output }}",
+            literature_review="{{ literature_review_task.output }}"  # Depends on both previous tasks
+        )
+
+        final_report_task = tasks.create_final_report(
+            agent=research_director,
+            summary="{{ summary_task.output }}",
+            technical_analysis="{{ technical_analysis_task.output }}",
+            literature_review="{{ literature_review_task.output }}"  # Integrates all previous outputs
+        )
+
+        # Configure the CrewAI workflow with agents and their tasks
+        crew = Crew(
+            agents=[
+                research_director,
+                technical_expert,
+                literature_specialist,
+                summary_writer
+            ],
+            tasks=[
+                technical_analysis_task,
+                literature_review_task,
+                summary_task,
+                final_report_task
+            ],
+            verbose=True  # Enable detailed logging of the analysis process
+        )
+
+        # Execute the analysis workflow
+        result = crew.kickoff()
+        return result
+
+
+if __name__ == "__main__":
+    print("## Welcome to Research Analyzer AI")
+    print("----------------------------------")
+
+    # Validate command line arguments
+    if len(sys.argv) != 2:
+        print("Usage: python main.py <path_to_pdf>")
+        print("Example: python main.py research_paper.pdf")
+        sys.exit(1)
+
+    pdf_path = sys.argv[1]
+    print(f"\nReading PDF file: {pdf_path}")
+
+    try:
+        # Extract and process PDF content
+        pdf_reader = PdfReader(pdf_path)
+        paper_content = pdf_reader.read()
+
+        print("\nPDF content extracted successfully!")
+        print(f"Content length: {len(paper_content)} characters")
+
+        # Initialize and execute the analysis workflow
+        print("\nStarting analysis...")
+        research_crew = ResearchAnalyzerCrew(paper_content)
+        result = research_crew.run()
+
+        # Display analysis results
+        print("\n\n########################")
+        print("## Research Analysis Results:")
+        print("########################\n")
+        print(result)
+
+    except Exception as e:
+        print(f"Error: {str(e)}")
+        sys.exit(1)