Basic TTN simulation (#41)

* Created new general.py file for logger and cutensornet handle. Updated copyright comment.
* Moved MPS code to make room for TTN.

* Moved ConfigMPS to Config in general.py

* Began implementation of TTN

* Added single-qubit gate support and some basic tests.

* Added vdot to TTN

* Implemented get_amplitude and get_statevector. Current tests passing.

* Implemented canonicalisation

* Canonicalisation debugged

* Exact apply 2q gate implemented and debugged

* Added truncation by chi and value_of_zero

* Unified TTN and MPS via StructuredState abstract class

* _get_qubit_partition now does recursive balanced bisections.

* Added truncation_fidelity approach to TTNxGate

Co-authored-by: Iakov Polyak <45295461+yapolyak@users.noreply.github.com>

* Updated docstring.

---------

Co-authored-by: PabloAndresCQ <pablo.andres-martinez@cambridgequantum.com>
Co-authored-by: Iakov Polyak <45295461+yapolyak@users.noreply.github.com>

docs/api.rst

@@ -3,4 +3,4 @@ API documentation
 
 .. toctree::
     modules/fullTN.rst
-    modules/mps.rst
+    modules/structured_state.rst

docs/modules/structured_state.rst

@@ -0,0 +1,61 @@
+Structured state evolution
+==========================
+
+.. automodule:: pytket.extensions.cutensornet.structured_state
+
+
+Simulation
+~~~~~~~~~~
+
+.. autofunction:: pytket.extensions.cutensornet.structured_state.simulate
+
+.. autoenum:: pytket.extensions.cutensornet.structured_state.SimulationAlgorithm()
+    :members:
+
+.. autoclass:: pytket.extensions.cutensornet.structured_state.Config()
+
+    .. automethod:: __init__
+
+.. autoclass:: pytket.extensions.cutensornet.structured_state.CuTensorNetHandle
+
+
+Classes
+~~~~~~~
+
+.. autoclass:: pytket.extensions.cutensornet.structured_state.StructuredState()
+
+    .. automethod:: __init__
+    .. automethod:: is_valid
+    .. automethod:: apply_gate
+    .. automethod:: apply_scalar
+    .. automethod:: vdot
+    .. automethod:: sample
+    .. automethod:: measure
+    .. automethod:: postselect
+    .. automethod:: expectation_value
+    .. automethod:: get_fidelity
+    .. automethod:: get_statevector
+    .. automethod:: get_amplitude
+    .. automethod:: get_qubits
+    .. automethod:: get_byte_size
+    .. automethod:: get_device_id
+    .. automethod:: update_libhandle
+    .. automethod:: copy
+
+.. autoclass:: pytket.extensions.cutensornet.structured_state.TTNxGate()
+
+    .. automethod:: __init__
+
+.. autoclass:: pytket.extensions.cutensornet.structured_state.MPSxGate()
+
+    .. automethod:: __init__
+
+.. autoclass:: pytket.extensions.cutensornet.structured_state.MPSxMPO()
+
+    .. automethod:: __init__
+
+
+Miscellaneous
+~~~~~~~~~~~~~
+
+.. autofunction:: pytket.extensions.cutensornet.structured_state.prepare_circuit_mps

examples/mpi/mpi_overlap_bcast_mps.py

@@ -42,10 +42,10 @@
 
 from pytket.circuit import Circuit, fresh_symbol
 
-from pytket.extensions.cutensornet.mps import (
+from pytket.extensions.cutensornet.structured_state import (
     simulate,
-    ConfigMPS,
-    ContractionAlg,
+    Config,
+    SimulationAlgorithm,
     CuTensorNetHandle,
 )
 
@@ -109,7 +109,7 @@
 this_proc_mps = []
 with CuTensorNetHandle(device_id) as libhandle:  # Different handle for each process
     for circ in this_proc_circs:
-        mps = simulate(libhandle, circ, ContractionAlg.MPSxGate, ConfigMPS())
+        mps = simulate(libhandle, circ, SimulationAlgorithm.MPSxGate, Config())
         this_proc_mps.append(mps)
 
 if rank == root: