Spaces:
Running
Running
File size: 4,343 Bytes
da0ce7b | 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 94 95 96 97 | #!/usr/bin/env python3
"""Test + benchmark the bio pipeline on the data we legitimately have.
Two honest lanes, printed with ✅/❌:
1. **Bio anchor (synthetic / cited reference).** No open UPE dataset exists as a
clean download (biophoton data lives in paper supplements — see
``docs/reports/UPE_DATA_AVAILABILITY.md``), so the UPE path is exercised on the
cited reference profile + deterministic synthetic data: broadband/featureless
UPE must score NON-separable (the honest anchor), planted emission lines must
score separable, and the convergence map must run under governance.
2. **Real open molecular data.** The repo already holds genuine open-source spectra
(NIST WebBook IR peaks in ``data/spectra/nist_ir_peaks.csv``). We run the
falsifiable phenolic engine on them via ``connector.run_analysis`` and assert a
valid, deterministic result — the "open source data + test" demonstration.
Run: ``python scripts/validation/benchmark_bio_open_data.py``.
"""
from __future__ import annotations
import sys
from pathlib import Path
REPO_ROOT = Path(__file__).resolve().parents[2]
if str(REPO_ROOT) not in sys.path:
sys.path.insert(0, str(REPO_ROOT))
DATA = REPO_ROOT / "data" / "spectra"
FIXTURES = REPO_ROOT / "tests" / "fixtures"
def main(argv: list[str] | None = None) -> int:
"""Run the bio anchor + real-open-data checks; return 0 iff all pass."""
import numpy as np
from aureon.bio.convergence_map import analyze_convergence
from aureon.bio.upe_signal_adapter import score_upe, synthetic_upe
checks: list[tuple[bool, str]] = []
# ---- Lane 1: bio anchor (synthetic / cited reference) ------------------
broadband = score_upe(synthetic_upe("broadband"), consent=True,
provenance="open-data validation (synthetic UPE)", nulls=300)
structured = score_upe(synthetic_upe("structured"), consent=True,
provenance="open-data validation (synthetic UPE)", nulls=300)
checks.append((broadband.valid and not broadband.structure_present,
"broadband UPE reference -> structure ABSENT (honest anchor)"))
checks.append((structured.structure_present,
"structured UPE (planted lines) -> structure PRESENT"))
img = np.zeros((120, 120, 3), np.uint8)
img[:60, :60] = (230, 30, 30)
img[:60, 60:] = (30, 200, 30)
img[60:, :60] = (30, 30, 220)
img[60:, 60:] = (230, 220, 20)
cmap = analyze_convergence(img, consent=True, provenance="open-data validation (synthetic image)",
grid=4, nulls=200)
checks.append((cmap.valid and cmap.controls_pass
and all(c.converged == (c.channels_fired == 2) for c in cmap.cells),
f"convergence map valid; {cmap.n_converged}/{len(cmap.cells)} both-channel cells"))
# ---- Lane 2: real open molecular data (NIST IR) ------------------------
sources = [str(p) for p in (
FIXTURES / "weed_phenolic_spectral_map_codex.csv",
DATA / "nist_ir_peaks.csv",
DATA / "curated_open_access_peaks.csv",
) if p.exists()]
nist_present = (DATA / "nist_ir_peaks.csv").exists()
if sources:
import connector
r1 = connector.run_analysis(sources, nulls=300, seed=0)
r2 = connector.run_analysis(sources, nulls=300, seed=0)
d1, d2 = r1.to_dict(), r2.to_dict()
checks.append((d1["valid"] and len(d1["compounds"]) > 0,
f"real open data ({len(sources)} src, NIST={nist_present}) -> "
f"valid, {len(d1['compounds'])} compounds"))
checks.append((d1["compounds"] == d2["compounds"],
"phenolic run on open data is deterministic (seed=0)"))
else:
checks.append((False, "no open-data sources found on disk"))
print("Bio pipeline — open-source data test + benchmark")
ok = True
for passed, label in checks:
ok = ok and passed
print(f" {'✅' if passed else '❌'} {label}")
print(f" broadband A_p={broadband.test_A_p} | structured A_p={structured.test_A_p} "
f"B_p={structured.test_B_p}")
return 0 if ok else 1
if __name__ == "__main__":
raise SystemExit(main())
|