File size: 3,786 Bytes
b9d7591 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 | """
Executes the actual Python code blocks published in docs/getting-started.md
and docs/examples.md, instead of a hand-maintained copy of them.
The docs snippets were written by copying from real, already-tested code
(experiments/matrix_healing_zne.py, test_mps.py, test_autodiff.py) -- this
file closes the other direction: if a future API change (renamed function,
changed signature, removed argument) breaks what's *published*, these tests
fail immediately instead of the breakage sitting undiscovered until someone
rereads the site by hand.
Each block is located by its section heading (not "first ```python fence in
the file"), since docs/getting-started.md also has a Colab-only block under
"## Install" that uses IPython magics (`!git clone`, `%cd`) and is not valid
standalone Python -- it's intentionally not selected here.
"""
import re
import pathlib
import numpy as np
import pytest
REPO_ROOT = pathlib.Path(__file__).parent
def _extract_section_code(md_path, heading):
text = (REPO_ROOT / md_path).read_text(encoding='utf-8')
heading_pos = text.index(heading)
remainder = text[heading_pos + len(heading):]
match = re.search(r"```python\n(.*?)\n```", remainder, re.DOTALL)
assert match is not None, f"no python fence found under {heading!r} in {md_path}"
return match.group(1)
def _run_section(md_path, heading):
code = _extract_section_code(md_path, heading)
namespace = {'__name__': '__doc_example__'}
exec(compile(code, f"{md_path}::{heading}", 'exec'), namespace)
return namespace
class TestGettingStartedExamples:
def test_quick_start(self):
ns = _run_section('docs/getting-started.md', '## Quick start')
probs = np.asarray(ns['probs'])
sv = np.asarray(ns['sv'])
assert probs.sum() == pytest.approx(1.0, abs=1e-6)
assert np.linalg.norm(sv) == pytest.approx(1.0, abs=1e-6)
def test_anti_oom_chunk(self):
ns = _run_section('docs/getting-started.md', '## Anti-OOM for large circuits')
sim = ns['sim']
assert sim.n == 27
assert sim.num_chunks >= 1
def test_zero_noise_extrapolation(self):
ns = _run_section('docs/getting-started.md', '## Zero-Noise Extrapolation')
raw_fidelity = float(ns['raw_fidelity'])
corrected_fidelity = float(ns['corrected_fidelity'])
assert 0.0 <= raw_fidelity <= 1.0
assert 0.0 <= corrected_fidelity <= 1.0
# fixed rng seed=0 -- this is a real regression check, not just "ran
# without raising": the correction must still improve fidelity.
assert corrected_fidelity > raw_fidelity
class TestExamplesPage:
def test_density_matrix_zne_healing(self):
ns = _run_section('docs/examples.md', '## Density-matrix ZNE healing')
raw_fidelity = float(ns['raw_fidelity'])
corrected_fidelity = float(ns['corrected_fidelity'])
assert 0.0 <= raw_fidelity <= 1.0
assert 0.0 <= corrected_fidelity <= 1.0
assert corrected_fidelity > raw_fidelity
def test_mps_low_entanglement(self):
ns = _run_section('docs/examples.md', '## MPS for low-entanglement circuits')
prob = np.asarray(ns['prob'])
n = ns['n']
assert prob[0] == pytest.approx(0.5, abs=1e-6)
assert prob[2 ** n - 1] == pytest.approx(0.5, abs=1e-6)
assert prob.sum() == pytest.approx(1.0, abs=1e-6)
def test_differentiable_vqe(self):
ns = _run_section('docs/examples.md', '## Differentiable VQE')
energy = float(ns['energy'])
ground_state = float(np.min(np.diag(ns['h_matrix'])))
# fixed rng seeds (theta=3, hamiltonian=7) -- 30 Adam epochs must land
# meaningfully below the mid-spectrum, not just "somewhere".
assert energy < ground_state / 2
|