| --- |
| title: Kryonex Biology Knowledge Engine |
| emoji: 🚀 |
| colorFrom: red |
| colorTo: red |
| sdk: docker |
| app_port: 8501 |
| tags: |
| - streamlit |
| pinned: false |
| short_description: Streamlit template space |
| --- |
| |
| # 🚀 NASA Bioscience Explorer |
|
|
| A Streamlit web application for exploring and analyzing NASA life sciences publications with AI-powered summarization capabilities. |
|
|
|  |
|
|
| ## 📋 Overview |
|
|
| NASA Bioscience Explorer provides an interactive platform to explore 608 NASA life sciences publications. The application features advanced filtering, visualization, and AI-powered section-by-section paper summarization to help researchers quickly understand key findings. |
|
|
| ## ✨ Features |
|
|
| ### 🔍 Search & Filter |
| - **Keyword Search**: Search across titles, topics, and organisms |
| - **Topic Filtering**: Filter by research areas (Bone Health, Muscle Physiology, Immune System, etc.) |
| - **Organism Filtering**: Filter by studied organisms (Mouse, Human, Arabidopsis, etc.) |
| - **Real-time Filtering**: Instant results as you type and select filters |
|
|
| ### 📊 Research Dashboard |
| - **Interactive Visualizations**: Pie charts and bar graphs showing research distribution |
| - **Publication Metrics**: Key statistics on publications, topics, and organisms |
| - **Publication Browser**: Browse all filtered publications with direct paper links |
| - **Research Insights**: Identify most studied areas and research gaps |
|
|
| ### 📄 AI Paper Summarizer |
| - **Section-wise Summarization**: AI-generated summaries for specific paper sections |
| - **Smart Content Extraction**: Automatically extracts Introduction, Methods, Results, Discussion, and Conclusion sections |
| - **Multi-model Support**: Uses Facebook's BART-large-CNN model for high-quality summarization |
| - **URL-based Processing**: Summarize any scientific paper by providing its URL |
|
|
| ### 🤖 AI-Powered Features |
| - **Section Detection**: Intelligent identification of academic paper sections |
| - **Chunk Processing**: Handles long documents through smart text chunking |
| - **Cached Results**: Efficient caching for faster repeated access |
| - **Fallback Mechanisms**: Robust error handling with fallback options |
|
|
| ## 🛠️ Installation |
|
|
| ### Prerequisites |
| - Python 3.8+ |
| - pip package manager |
|
|
| ### Setup Instructions |
|
|
| 1. **Clone the repository** |
| ```bash |
| git clone https://github.com/your-username/nasa-bioscience-explorer.git |
| cd nasa-bioscience-explorer |
| ``` |
|
|
| 2. **Create a virtual environment** (recommended) |
| ```bash |
| python -m venv venv |
| source venv/bin/activate # On Windows: venv\Scripts\activate |
| ``` |
|
|
| 3. **Install dependencies** |
| ```bash |
| pip install -r requirements.txt |
| ``` |
|
|
| 4. **Set up data directory** |
| ```bash |
| mkdir data |
| # Place your SB_publication_PMC.csv file in the data directory |
| ``` |
|
|
| 5. **Run the application** |
| ```bash |
| streamlit run app.py |
| ``` |
|
|
| ## 📁 Project Structure |
|
|
| ``` |
| nasa-bioscience-explorer/ |
| │ |
| ├── app.py # Main Streamlit application |
| ├── requirements.txt # Python dependencies |
| ├── README.md # Project documentation |
| ├── data/ # Data directory |
| │ └── SB_publication_PMC.csv # Publication dataset |
| ``` |
|
|
| ## 📊 Data Format |
|
|
| The application expects a CSV file with the following columns: |
| - `Title`: Publication title |
| - `Link`: URL to the publication |
| - Additional columns for metadata |
|
|
| ## 🔧 Configuration |
|
|
| ### Environment Variables |
| No environment variables required. All configuration is handled within the application. |
|
|
| ### Model Configuration |
| - **Summarization Model**: `facebook/bart-large-cnn` |
| - **Text Chunk Size**: 1000 characters |
| - **Max Input Length**: 8000 characters |
| - **Section Limit**: 2000 characters per section |
|
|
| ## 🚀 Usage |
|
|
| ### Research Dashboard |
| 1. Use the search bar and filters to find relevant publications |
| 2. View distribution charts to understand research trends |
| 3. Click on individual publications to generate AI summaries |
| 4. Explore research insights and identified gaps |
|
|
| ### Paper Summarizer |
| 1. Navigate to the "Paper Summarizer" tab |
| 2. Enter any scientific paper URL (PMC, PubMed, etc.) |
| 3. Click "Summarize Paper" to generate section-wise summaries |
| 4. View summaries for Introduction, Methods, Results, Discussion, and Conclusion |
|
|
| ## 🎯 Supported Research Areas |
|
|
| - **Bone Health**: Bone density, skeletal research, osteoporosis |
| - **Muscle Physiology**: Muscle atrophy, physiology studies |
| - **Immune System**: Immune response, infections, microbiome |
| - **Plant Biology**: Arabidopsis, plant growth, root systems |
| - **Radiation Effects**: DNA damage, genomic stability |
| - **Microgravity Adaptation**: Gravity effects, space adaptation |
| - **Other**: Miscellaneous life sciences research |
|
|
| ## 🧬 Supported Organisms |
|
|
| - Mouse |
| - Human/Astronaut |
| - Arabidopsis |
| - Drosophila |
| - Rat |
| - Various (multiple organisms) |
|
|
| ## 🔍 API & Libraries Used |
|
|
| - **Streamlit**: Web application framework |
| - **Plotly**: Interactive visualizations |
| - **Transformers**: Hugging Face AI models (BART) |
| - **Trafilatura**: Web content extraction |
| - **Pandas**: Data manipulation and analysis |
| - **Textwrap**: Text formatting utilities |
|
|
