Dmitrov is a set of compiled code libraries and utilities to facilitate the use of multidimensional datasets in diverse formats and sizes from within code. Dmitrov does away with the traditional distinctions between files of different formats, and between input, creation, manipulation and output of datasets held in different locations. You use the same interface to read data from a file (whatever the format), augment and transform that data, create new data sets in memory, and output data (to whatever format). Dmitrov can work with local files and local memory, or can work over the cloud, scales well from small data sets / few variables, to huge data sets / large numbers of variables. Dmitrov also allows for the addition of metadata to files, in order to create self-describing data packages.
Multidimensional datasets are very common in scientific research: temperature or precipitation grids, satellite or medical imagery, mathematical matrices, for example. Such datasets can be stored in a variety of different file formats such as comma- or tab-separated text files, NetCDF or ASCIIGrids. They can also vary from a few values stored in a local file, to multi-terabyte data repositories held in the cloud. This diversity of data formats and sizes can make the utilisation of such data in code a complicated and time consuming task