TODO

TODO list
=========
[ ] : not started
[W] : work in progress
[X] : done
[-] : cancelled

End-user:
=========

[X] Cumulative resources: resources that can process multiple tasks at the same time

[X] Optional tasks: tasks that may or may not be scheduled

[X] Buffers: input/output for tasks

[X] Add a plotly/pandas gantt renderer with real dates and times

[ ] JSON importer/exporter for a scheduling problem

[ ] Export a networkx model to a set of tasks/taks precedences

[W] Voila based user interface

[ ] Add a penalty parameter to the Task class. And the related optimization (minimization) function.

[X] Cost function: currently the cost of a task is a constant rate. It is defined, for example, as 5$/hour, or
540€/day etc. It should be possible to express the cost as a function of the time, for example a linear or affine
function (cost(t)=a*t+b) for example. Indeed many industrial costs may change over time as quadratic or exponential
functions, seasonality etc.

[X] incremental solver: optimizer that uses a different method than the Optimize solver, which can
be quite long to converge to a solution

Dev:
====

[X] Performance benchmark

[X] Improve class hierarchy and inheritance

[ ] Categorize task constraints

[ ] Developer guide, especially how to add constraints

Misc/unordered:
===============

[ ] Resource status. A status is a set of real/boolean parameters (for instance the position of a point, the heat of an oven). That would enable to create tasks that change a resource status, from example "Move a crane from point A to point B" or "heat the oven from 25°C to 1300°C", given a certain time or speed.

[ ] Add a 'MobileWorker' class. A mobile worker has a parameter position that changes across time. Position
can be 1d (linear), 2d or 3d. Possible to use this class to represent cranes or trucks.

[ ] Test CapacityConstraint for cumulative workers