Building a Modular Health Information Crawler: Leveraging Apache Nutch for the Tala-Med Search Engine

Developed in 2020, the tala-med search engine provides high-quality, evidence-based health information from trustworthy German websites while ensuring user privacy. However, it still leaves room for technical improvements. In the present work, we are replacing the Fess crawler with Apache Nutch to improve scalability and customization, as Nutch offers greater flexibility for large-scale web crawling and indexing. The new system uses Docker to integrate five services: PostgreSQL for configurations, Nutch for crawling and indexing, and ElasticSearch for search operations; a Manager orchestrates the process with custom configurations via extended Nutch-REST interface. The configuration options include domain-specific URL filters to select the crawled content. A test crawl demonstrated the system's effectiveness, processing approximately 23k websites over 65 hours. Future work will focus on deploying the crawler long-term and generating a search index for further analysis. We have published our code under the MIT license at https://gitlab.com/mri-tum/aiim/search-platform/crawler.