mrmimic/covid-19-applet

covid-19-applet is a web application that lets you search through a large collection of COVID-19 research papers. Instead of manually reading thousands of academic articles, you can use this tool to find relevant information by searching for concepts and topics, similar to how you might search the web but focused specifically on pandemic-related research. Under the hood, the app pulls research data from a dataset called CORD-19, which was assembled by AI research institutes. It organizes this mountain of papers into a searchable database, and uses a technique called word2vec, essentially a way for computers to understand the meaning and relationships between words, to match your search queries with relevant content in the articles. When you type a search, the system finds papers that discuss what you're asking about, even if they don't use the exact same words. The results are then displayed through a simple web interface you access in your browser. This tool would be useful for researchers, public health professionals, journalists, or policy analysts trying to quickly locate specific information buried in the overwhelming volume of COVID-19 literature that emerged during the pandemic. For example, if someone needed to find what studies said about transmission routes in specific environments, they could search and get pointed to the right papers rather than scrolling through academic databases manually. Setting it up is somewhat involved, you need to download the research dataset, build a local database from it, and configure several file paths. The README notes that the full database would be about 22 GB if you included everything, but a filtered version focusing only on COVID-19-related papers brings it down to roughly 700 MB. There's also a caching system that holds search results for ten minutes at a time, so the first search after starting the server or after a period of inactivity takes about thirty seconds to process.