Describe the problem:
Calling the optimize_stochastic method for the PojectedSchrodingerPyCI class fails.
Code that reproduces the bug:
The following test on the 108-test-interface-submodule branch fails if run in the test folder:
import numpy as np
import pyci
from fanpy.interface.fanci.pyci import ProjectedSchrodingerPyCI
from interface_utils import FakeHamiltonian, FakeWavefunction, FakeSchrodinger
def make_test_instance(**overrides):
"""make test instance of ProjectedSchrodingerPyCI with fake fanpy objective and fake pyci hamiltonian and wavefunction
This helps set up a class that requires a lot of parameters.
"""
# build Fake fanpy objective
wfn = FakeWavefunction(2, 4, np.ones(4))
nocc = wfn.nelec // 2
ham = FakeHamiltonian(np.ones((2, 2)), np.ones((2, 2, 2, 2)))
obj = FakeSchrodinger(wfn, ham)
# build fake pyci hamiltonian
energy_nuc = 0.0
pyci_ham = pyci.hamiltonian(energy_nuc, ham.one_int, ham.two_int)
# build pyci wavefunction
# use FCI pspace wavefunction
pyci_wfn = pyci.fullci_wfn(pyci_ham.nbasis, wfn.nelec - nocc, nocc)
defaults = {
"fanpy_objective" : obj,
"ham" : pyci_ham,
"wfn" : pyci_wfn,
"nocc" : 2,
"seniority": wfn.seniority,
"nproj": 1,
"fill": "excitation",
"mask": np.ones(wfn.params.shape[0]+1, dtype=bool),
"constraints": {},
"param_selection": obj.indices_component_params,
"norm_param": None,
"norm_det": None,
"max_memory": 8000,
"step_print": False,
"step_save": False,
"tmpfile": ""
}
defaults.update(overrides)
return ProjectedSchrodingerPyCI(**defaults)
def test_optimize_stochasitc_lstsq():
""" Check if optimize method runs without errors and energy is one of the keys"""
mask = np.ones(5, dtype=int)
pyci_obj = make_test_instance(mask=mask)
initial_guess = np.random.rand(pyci_obj.fanpy_wfn.nparams+1)
results = pyci_obj.optimize_stochastic(nsamp=3, x0=initial_guess, mode="lstsq", fill = pyci_obj.fill)
assert "energy" in results.keys()Traceback error:
=================================================================================== test session starts ====================================================================================
platform linux -- Python 3.10.16, pytest-8.3.4, pluggy-1.5.0
rootdir: /home/quintana/Documents/code/Fanpy
configfile: pyproject.toml
plugins: cov-6.0.0
collected 1 item
test.py F [100%]
========================================================================================= FAILURES =========================================================================================
______________________________________________________________________________ test_optimize_stochasitc_lstsq ______________________________________________________________________________
def test_optimize_stochasitc_lstsq():
""" Check if optimize method runs without errors and energy is one of the keys"""
mask = np.ones(5, dtype=int)
pyci_obj = make_test_instance(mask=mask)
initial_guess = np.random.rand(pyci_obj.fanpy_wfn.nparams+1)
> results = pyci_obj.optimize_stochastic(nsamp=3, x0=initial_guess, mode="lstsq", fill = pyci_obj.fill)
test.py:51:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <fanpy.interface.fanci.pyci.ProjectedSchrodingerPyCI object at 0x7e0fca5b93c0>, nsamp = 3, x0 = array([0.2028639 , 0.28284929, 0.95316465, 0.30245556, 0.24879663]), mode = 'lstsq'
use_jac = False, fill = 'excitation', kwargs = {}, ham = <pyci._pyci.secondquant_op object at 0x7e0fca57e9b0>, nproj = 1, nparam = 5, nbasis = 2, nocc_up = 1, nocc_dn = 1, constraints = ()
def optimize_stochastic(
self,
nsamp: int,
x0: np.ndarray,
mode: str = "lstsq",
use_jac: bool = False,
fill: str = "excitation",
**kwargs: Any,
) -> List[Tuple[np.ndarray]]:
"""
Run a stochastic optimization of a FanCI wave function.
Parameters
----------
nsamp: int
Number of samples to compute.
x0 : np.ndarray
Initial guess for wave function parameters.
mode : ('lstsq' | 'root' | 'cma'), default='lstsq'
Solver mode.
use_jac : bool, default=False
Whether to use the Jacobian function or a finite-difference approximation.
fill : ('excitation' | 'seniority' | None)
Whether to fill the projection ("P") space by excitation level, by seniority, or not
at all (in which case ``wfn`` must already be filled).
kwargs : Any, optional
Additional keyword arguments to pass to optimizer.
Returns
-------
result : List[Tuple[np.ndarray]]
List of (occs, coeffs, params) vectors for each solution.
"""
# Get wave function information
ham = self.ham
nproj = self.nproj
nparam = self.nparam
nbasis = self.wfn.nbasis
nocc_up = self.wfn.nocc_up
nocc_dn = self.wfn.nocc_dn
constraints = self.constraints
mask = self.mask
ci_cls = self.wfn.__class__
# Start at sample 1
isamp = 1
result = []
# Iterate until nsamp samples are reached
# **kwargs: Any,
while True:
# Optimize this FanCI wave function and get the result
opt = self.optimize(x0, mode=mode, use_jac=use_jac, **kwargs)
energy = opt.x[-1]
if opt.success:
print("Optimization was successful")
else:
print("Optimization was not successful: {}".format(opt.message))
> print("Final Electronic Energy for sample {isamp}: {}".format(energy))
E KeyError: 'isamp'
../fanpy/interface/fanci/pyci.py:860: KeyError
----------------------------------------------------------------------------------- Captured stdout call -----------------------------------------------------------------------------------
WARNING: System is underdetermined with dimensions 1, 5. Continuing anyways
Iteration Total nfev Cost Cost reduction Step norm Optimality
0 1 3.0041e+01 7.75e+00
1 2 2.2201e+01 7.84e+00 1.09e+00 6.66e+00
2 3 1.0072e+01 1.21e+01 2.18e+00 4.49e+00
3 4 9.4402e-03 1.01e+01 4.35e+00 1.37e-01
4 8 1.0442e-06 9.44e-03 1.36e-01 1.45e-03
5 13 7.3321e-08 9.71e-07 1.06e-03 3.83e-04
6 15 1.0975e-08 6.23e-08 5.31e-04 1.48e-04
7 17 6.8904e-09 4.08e-09 2.66e-04 1.17e-04
8 19 1.1831e-10 6.77e-09 1.33e-04 1.54e-05
9 22 7.3713e-13 1.18e-10 1.66e-05 1.21e-06
10 25 3.7015e-13 3.67e-13 2.07e-06 8.60e-07
11 27 1.5646e-14 3.55e-13 1.04e-06 1.77e-07
12 30 1.1154e-15 1.45e-14 1.30e-07 4.72e-08
13 32 1.5488e-16 9.61e-16 6.48e-08 1.76e-08
`xtol` termination condition is satisfied.
Function evaluations 32, initial cost 3.0041e+01, final cost 1.5488e-16, first-order optimality 1.76e-08.
Optimization was successful
================================================================================= short test summary info ==================================================================================
FAILED test.py::test_optimize_stochasitc_lstsq - KeyError: 'isamp'
==================================================================================== 1 failed in 0.18s =====================================================================================Package versions:
python: 3.10
fanpy: N/A
numpy:1.26
Additonal comments
No response
Describe the problem:
Calling the
optimize_stochasticmethod for thePojectedSchrodingerPyCIclass fails.Code that reproduces the bug:
The following test on the
108-test-interface-submodulebranch fails if run in thetestfolder:Traceback error:
Package versions:
python: 3.10
fanpy: N/A
numpy:1.26
Additonal comments
No response