QPanda: Quantum Programming Architecture for NISQ Device Application
QPanda-lite should be a simple, easy, and transparent python-native version for QPanda.
Developing. Unstable.
- A clear, and tranparent way to assemble/execute a quantum program
- Support sync/async modes for execution on a quantum hardware
- Clear error hints
- Full, and better documentations
- Visualization of the quantum program
- Be able to migrate to different quantum backends
- Windows
- Linux (not fully tested)
- MacOS (not fully tested)
- Python >= 3.8
manually install via pip :
- pyquafu (pip install pyquafu)
manually install via pip :
- qiskit (pip install qiskit) and
- qiskit-ibm-provider (pip install qiskit-ibm-provider) and
- qiskit-ibmq-provider (pip install qiskit-ibmq-provider)
- CMake >= 3.1
- C++ compiler (with C++ 14 support), including MSVC, gcc, clang, etc...
# Clone the code
git clone https://github.com/Agony5757/QPanda-lite.git
cd QPanda-lite
# install
python setup.py install --no-cppgit clone https://github.com/Agony5757/QPanda-lite.git
cd QPanda-lite
# install
python setup.py develop
With C++ enabled (quantum circuit simulator written in C++, ensure that CMAKE is included in your environment variables.)
git clone https://github.com/Agony5757/QPanda-lite.git
cd QPanda-lite
# install
python setup.py install
Will be released in the future.
For python 3.8 to 3.10
pip install qpandalite
There are several ways to use QPanda-lite now.
- Circuit building
- Run circuit on several backends / dummies (classical-simulation backends)
- Circuit simulation
Refer to test/demo
from qpandalite import Circuit
c = Circuit()
c.rx(1, 0.1)
c.cnot(1, 0)
measure(0, 1, 2, 3)
print(c.circuit)| Function | Code sample | Explanation |
|---|---|---|
| Circuit initialization | c = qpandalite.Circuit() | |
| Qubit/cbit initialization | No need to specify the number | |
| Gate (like CNOT) | c.cnot(1,2) | Directly use the qubit number |
| Measure | c.measure(0,1,2) | Directly use the qubit number (no support mid-circuit measurement) |
| Remap | c = c.remapping({0:10, 1:11, 2:12}) | Input a python dict to indicate the mapping. It creates a new Circuit object. |
| Output as str | c.circuit / c.originir | Return a python str |
| Function | Code sample | Explanation |
|---|---|---|
| "Import" the platform | import qpandalite.task.originq as originq | This importing is independent from "import qpandalite". Available platforms are under qpandalite.task |
| Prepare the account | See qcloud_config_template | |
| Task submission | taskid = originq.submit_task(circuits) | Circuits is str or List[str]. Returned taskid can be either list or one str, depending on the number of inputting circuits. All returns are native python data structures. See Circuit build. |
| Query (synchronously) | results = originq.query_by_taskid_sync(taskid) | Inputting the taskid by the return of submit_task. The results are always a list (even if you only submit one circuit). All returns are native python data structures. |
| Query (asynchronously) | status_and_result = originq.query_by_taskid(taskid) | Inputting the taskid by the return of submit_task. This will immediately return without waiting. Use status_and_result['status'] to see if the computing is finished; use status_and_result['result'] to view results (the same with Query (synchronously), always being a list). All returns are native python data structures. |
| Handle measurement result | results = originq.convert_originq_result(results, style = 'keyvalue', prob_or_shots = 'prob', key_style = 'bin') | Convert the raw data to a more human-friendly format. Style includes "keyvalue" and "list", prob_or_shots includes "prob" and "shots". When inputting a list, the output is also a list corresponding to all inputs. All returns are native python data structures. |
| Calculate expectation | exps = [calculate_expectation(result, ['ZII', 'IZI', 'IIZ']) for result in results] | Calculate the Z/I expectations accroding to the measurement results. Note that it only accepts the diagonal Hamiltonians. The hamiltonians can be a list, where the output is also a list. However, the input "result" cannot be a list. |
Refer to qcloud_config_template/originq_template.py
- Input the necessary information (token, urls, group_size) to call create_originq_online_config
- You will have the originq_online_config.json in your current working directory (cwd).
- Now you can submit task to the online chip!
Dummy server is used to emulate the behavior of an online-avaiable quantum computing server, without really accessing the system but with your local computer to simulate the quantum circuit.
-
Input the necessary information (available_qubits and available_topology) to call create_originq_dummy_config.
-
If you want both mode, use create_originq_config and inputting all needed information.
Refer to test/demo
Refer to qcloud_config_template/quafu_template.py
- Input the necessary information (token, urls, group_size) to call create_quafu_online_config
- You will have the quafu_online_config.json in your cwd.
- Now you can submit task to the online chip!
Todo.
Todo.
Refer to test/draft_test/originir_simulator_test.py
import qpandalite.simulator as qsim
sim = qsim.OriginIR_Simulator(reverse_key=False)
originir = '''
QINIT 72
CREG 2
RY q[45],(0.9424777960769379)
RY q[46],(0.9424777960769379)
CZ q[45],q[46]
RY q[45],(-0.25521154)
RY q[46],(0.26327053)
X q[46]
MEASURE q[45],c[0]
MEASURE q[46],c[2]
MEASURE q[52],c[1]
'''
res = sim.simulate(originir)
print(res)
print(sim.state)
# Expect output:
# [0.23218757036469517, 0.04592184582945769, 0.0, 0.0, 0.6122094271102275, 0.10968115669561962, 0.0, 0.0]
# [(0.4818584546987789+0j), (-0.21429383059121812+0j), (0.7824381298928546+0j), (0.33118145584500897+0j), 0j, 0j, 0j, 0j]Readthedocs: https://qpanda-lite.readthedocs.io/
The doc is based on
cd docs
pip install -r requirements.txt
make html