The Rope is a binary-tree based data structure designed for performing string-manipulation operations efficiently on large amounts of text.
This is an implementation of a rope in C++. While there is an existing implementation of ropes in the C++ STL, that
implementation is restricted to ropes using either char or wchar as the underlying type. Further, the STL
implementation is not unicode-aware, which increases the complexity of working with UTF-8 strings (for example,
iterating by words / code-points). This implementation provides both.
- This rope does not specify an underlying type, the underlying type must be provided as a template parameter.
- Random-access and slicing operations on the rope are supported by the concept of a "Measure", which is similar ot a monoid.
- Pretty much all interaction is via iterators, which are constructed from measure identifiers.
The "Measure" type is an abstract class with the following requirements:
- It must be constructable from a vector of the rope's underlying type
- It must be composable (i.e., there must be a function of type (T, T) -> T, such as '+')
- It must have an 'index' function, which, when given a vector and a target measure, will be capable of locating the point in the vector at which the accumulative measurement matches the target measure.
In the case of counting bytes (given char as the rope's underlying type), those functions might be implemented as:
using ByteMeasure = uint64_t;
ByteMeasure byte_measure(std::vector<char> &vec) { return vec.size(); }
ByteMeasure byte_compose(ByteMeasure &lhs, ByteMeasure &rhs) { return lhs + rhs; }
uint64_t byte_index(std::vector<char> &vec, ByteMeasure &measure) { return measure; }