Interfaces

These containers formally follow and extend on the named requirements of the C++ standard library. If you know how to use std::map, you already know how to use 90% any of these containers. You can use m.erase(it), m.insert(v), m.empty(), m.size(), m.begin() , and m.end() like you would with any other associative container.

Python Bindings

Although this library is completely implemented in C++17, because data scientists love Python, we also include Python bindings for all these data structures. We further replicate the syntax of the native Python data structures, so that m.erase(k) becomes del m[k], if m.empty() becomes if m:, and m.insert(k,v) becomes m[k] = v. If you're a C++ programmer using Python, the C++ container syntax is still available in Python.

C++ Concepts / Named Requirements

Formally, these containers implement the Container, ReversibleContainer, AllocatorAwareContainer, and AssociativeContainer Concepts / Named Requirements. Their iterators also implement the LegacyBidirectionalIterator concepts and they can use memory allocators that follow the Allocator concept. The extensions are formally defined as the concepts SpatialContainer, FrontContainer, and ArchiveContainer, whose pre- and post- conditions are checked with our unit tests.

All that means they work transparently with other native data structures. We include lots of unit tests, benchmarks, and continuous integration to make sure this compatibility is maintained. This also means they're easy to integrate with other libraries. For instance, the source file examples/matplotpp_example.cpp and the headers in source/pareto/matplot exemplify how to create the plots you are seeing in this documentation with Matplot++.