TokSearch: A search engine for fusion experimental data

Fusion Engineering and Design(2018)

引用 11|浏览67
暂无评分
摘要
At a typical fusion research site, experimental data is stored using archive technologies that deal with each discharge as an independent set of data. These technologies (e.g. MDSplus or HDF5) are typically supplemented with a database that aggregates metadata for multiple shots to allow for efficient querying of certain predefined quantities. Often, however, a researcher will need to extract information from the archives, possibly for many shots, that is not available in the metadata store or otherwise indexed for quick retrieval. To address this need, a new search tool called TokSearch has been added to the General Atomics TokSys control design and analysis suite. This tool provides the ability to rapidly perform arbitrary, parallelized queries of archived tokamak shot data (both raw and analyzed) over large numbers of shots. The TokSearch query API borrows concepts from SQL, and users can choose to implement queries in either Matlab™ or Python.
更多
查看译文
关键词
Apache spark,Big data,Data archive,Parallel processing
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要