Qibo

What's Changed
* Test that code does not contain `print` statements (391) stavros11
* fixing hep import in sphinx (390) scarrazza
* removing undesired prints (387) scarrazza
* renaming cma function (383) scarrazza

What's Changed
* Bell inequalities variational - Example (371) igres26
* Fix variational models for numpy backends (379) stavros11
* Improving performance for circuits with many gates (363) stavros11
* Use single gates.py file for all backends (372) stavros11
* Simplify backend switch (370) stavros11
* Fix CPU custom operators for more than 32 qubits (377) stavros11
* Update main benchmark script (375) stavros11
* Fix variational classifier (368) DiegoGM91
* Proposal for new model code structure (365) stavros11
* Conditioning gates on measurement outcomes (352) stavros11
* Standarizing optimizers input and output (358) scarrazza
* Multiple collapse measurements with single circuit call (355) stavros11
* Update semiclassical Shor example to use measurements with collapse (356) stavros11
* Updating multiline circuit draw (354) scarrazza
* Fix `.on_qubits()` with `.controlled_by()` gates (349) stavros11
* Use random seed for frequency custom operator (350) stavros11
* Fix Metropolis for sparse distributions (347) stavros11

What's Changed
* Fix issue with tf slicing (345) stavros11
* Hamiltonian expectation for density matrices (344) stavros11
* Measurement fixes (325) stavros11
* Fix controlled by Y gates for custom backend (343) stavros11
* Metropolis algorithm for sampling shots (332) stavros11
* Speeding up Metropolis for frequency (338) scarrazza
* Move state reshape in gates (340) stavros11
* Add nshot switcher (341) scarrazza
* PartialTrace gate for density matrices (335) stavros11
* Fixing circuit.draw misalignment (334) scarrazza
* Fix gap callback for degenerate Hamiltonians (329) stavros11
* Fix einsum in density matrix traceout (328) stavros11
* Make DistributedState custom operator free (323) stavros11
* adding Windows support (316) scarrazza
* Fix old tests for Windows (319) stavros11
* Detach custom operators (317) scarrazza
* Fix custom operator imports in tests (318) stavros11
* upgrading to tf 2.4.1 (315) scarrazza
* State abstraction (314) stavros11
* MaxCut Hamiltonian (311) scarrazza
* Symbolic Hamiltonian fixes (312) scarrazza
* Tests that run without Tensorflow (310) stavros11
* Circuit viewer (308) scarrazza
* small fixes (307) scarrazza
* Backend abstraction (303) stavros11
* Parallel circuit evaluation (304) scarrazza
* custom initial_state (305) scarrazza
* custom initial_state GPU (306) scarrazza
* Documenting custom VQC optimizations (298) scarrazza
* Updating wheels with cuda 11 (301) scarrazza
* Upgrading tf to 2.4.0 (280) scarrazza
* Create base callbacks (296) stavros11
* Fix test failure on GPU for parallel lbfgs (297) scarrazza
* Trainable flag for parametrized gates (294) stavros11

Major New Features

- Collapse gate.
- Shor (2n+3) example.
- qPDF model and example.
- Density matrix simulation using custom operators.
- Post-measurement bit-flip errors.
- Thermal relaxation error channel.
- Parallel optimization using multiprocessing for L-BFGS-B on CPU.
- Improved performance when repeating circuit execution.
- Replacing TensorFlow ThreadPool with OpenMP for custom operators on CPU.
- General code upgrades including tests, wheels and documentation.

Fixes
- Cost function for autoencoder tutorial.
- Circuit depth calculation.
- Backend return name.
- Trotter Hamiltonian from_symbolic and make_compatible.
- Scipy missing requirement in pip package.

Major Features
- Implemented new entaglement spectrum callback
- Extended circuit summary, fixed circuit depth definition.
- New method for circuit inversion
- New general unitary gates.
- Updating QASM and testing its compatibility.
- Adding possibility to add circuits to larger ones.
- Including symbolic interface for Hamiltonian allocation.
- Upgraded TensorFlow pip dependency to 2.3.1
- Updated documentation, fixed typos.
- Adjusting documentation urls to stable.

First Qibo stable release.

Major Features
- Basic circuit model containing gates and/or measurements (Circuit).
- Circuits that can be executed on multiple devices such as CPU/GPUs (DistributedCircuit).
- Circuit implementing the Quantum Fourier Transform (QFT).
- Variational Quantum Eigensolver (VQE), with support for optimization of the variational parameters.
- Quantum Approximate Optimization Algorithm (QAOA), with support for optimization of the variation parameters.
- Unitary time evolution of quantum states under a Hamiltonian (StateEvolution).
- Adiabatic time evolution of quantum states, with support for optimization of the scheduling function (AdiabaticEvolution).