When classifying and tagging multi-million file sets milliseconds count. Flare uses a staged approach to optimise performance.
First off is a super-fast disk index application to extract key metadata about the files. This data gets fed into the Sirus graph-based system that uses reference taxonomies of ‘asset’ names (wells, fields etc.) and contextual classes and keywords to create a ‘first pass’ classification and tagging of the files.
If needed, files are then opened and further processing is carried out on the file contents using increasingly complex apps to classify/tag the files.
This means it is simple to use the right tools for the job, at the right time.