List view
improve ease of use: (a) use marshmallow serializer to validate data sources; (b) use standardized format for all data sources; (c) update data readers to use standard serialized data source format, (d) use fields, eg: `DataField(name=, description=, units=, isconstant=, dtype=, ...)`, (e) ditto for all other types, fields have attributes, (f) new batch processing, parameterized study, optimization, data viz, standard reporting, data scrubbing layers, etc.
Due by October 21, 2016•22/22 issues closedBig Top release * comprehensive documentation * more concrete structure, guidelines, base classes and examples for simulation layer and models * quick start project script * add dictionary construct as allowable types to group data together as new data * use serializer instead of data source class as option * allow hot swappable serializer packages * absract calculators instead of hard coding into calculation layer, allow subclassing and provide more generic examples * provide parallel or mapping calculator for embarrassingly parallel calculations, eg: performance of multiple layouts * provide non-linear calculator * provide non-linear methods for simulating each timestep * provide examples of parametric studies or optimizing simulations
Due by May 10, 2016•14/14 issues closed