ClaudeandClaude Opus 5.5
74169e1bfc
The local baseline still claimed "lahiri_assumed_by_skill" after the engine default moved to Raman, so the PyJHora comparison reported 59% instead of ~95% (every longitude 1.446 deg off). The baseline now passes --ayanamsa and --node-mode to the engine and records what the engine reports; the compare takes --ayanamsa and stops when the baseline was built differently. The nine public Rodden AA backtest figures join the fixed samples. Raman run (19 samples): ascendant, planets, nakshatras, D2/D4/D9/D10 and Ashtakavarga all 100%; Shadbala and three old-chart dashas differ (separate). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01DbBXoyeZrYn6ryNcSe5Y2Y
110 lines
4.8 KiB
Python
110 lines
4.8 KiB
Python
import subprocess
|
|
import sys
|
|
from datetime import datetime, timezone
|
|
from pathlib import Path
|
|
|
|
from benchmarks.jyotish.scripts.run_pyjhora_compare import write_report
|
|
from benchmarks.jyotish.scripts.run_pyjhora_compare import compare_one
|
|
|
|
|
|
ROOT = Path(__file__).resolve().parents[1]
|
|
RUNNER = ROOT / "benchmarks" / "jyotish" / "scripts" / "run_pyjhora_compare.py"
|
|
|
|
|
|
def test_pyjhora_compare_help_is_non_executing():
|
|
result = subprocess.run(
|
|
[sys.executable, str(RUNNER), "--help"], cwd=ROOT, capture_output=True, text=True, timeout=15
|
|
)
|
|
|
|
assert result.returncode == 0
|
|
assert "--build-local" in result.stdout
|
|
assert "--refresh-local" in result.stdout
|
|
assert "--output-prefix" in result.stdout
|
|
assert "FileNotFoundError" not in result.stderr
|
|
|
|
|
|
def test_pyjhora_report_uses_supplied_utc_generation_timestamp():
|
|
report = write_report(
|
|
[],
|
|
[],
|
|
generated_at=datetime(2026, 7, 15, 4, 30, tzinfo=timezone.utc),
|
|
)
|
|
|
|
assert "生成时间：2026-07-15T04:30:00+00:00" in report
|
|
assert "生成时间：2026-06-03" not in report
|
|
|
|
|
|
def test_pyjhora_comparison_includes_d2_d4_bav_sav_and_shadbala_rows():
|
|
bodies = ["Ascendant", "Sun", "Moon", "Mars", "Mercury", "Jupiter", "Venus", "Saturn", "Rahu", "Ketu"]
|
|
chart = {body: {"sign": "Aries", "degree_in_sign": 1.0} for body in bodies}
|
|
planet = {body: {"sign": "Aries", "degree_in_sign": 1.0, "nakshatra": "Ashwini", "nakshatra_pada": 1} for body in bodies[1:]}
|
|
local = {
|
|
"ascendant": chart["Ascendant"], "planets": planet,
|
|
"varga": {"D2": chart, "D4": chart, "D9": chart, "D10": chart},
|
|
"dasha": {},
|
|
"ashtakavarga": {"bav": {name: [1] * 12 for name in ["Sun", "Moon", "Mars", "Mercury", "Jupiter", "Venus", "Saturn", "Lagna"]}, "sav": [7] * 12},
|
|
"shadbala": {name: 100.0 for name in ["Sun", "Moon", "Mars", "Mercury", "Jupiter", "Venus", "Saturn"]},
|
|
}
|
|
pyjhora = {
|
|
"ascendant": chart["Ascendant"], "planets": planet,
|
|
"varga": {"D2": chart, "D4": chart, "D9": chart, "D10": chart},
|
|
"dasha": {},
|
|
"ashtakavarga": local["ashtakavarga"], "shadbala": local["shadbala"],
|
|
}
|
|
|
|
sections = {row["section"] for row in compare_one("fixture", local, pyjhora)}
|
|
|
|
assert {"D2", "D4", "Ashtakavarga_BAV", "Ashtakavarga_SAV", "Shadbala"} <= sections
|
|
|
|
|
|
def test_baseline_records_the_engine_ayanamsa_and_the_compare_refuses_a_mismatch():
|
|
# 2026-11-04: the engine default moved from Lahiri to Raman while the
|
|
# baseline still claimed "lahiri_assumed_by_skill"; the comparison then
|
|
# reported 59% instead of ~95% because every longitude was 1.446 deg off.
|
|
from benchmarks.jyotish.scripts.run_pyjhora_compare import baseline_settings_error
|
|
from benchmarks.jyotish.scripts.run_skill_baseline import canonicalize
|
|
|
|
data = {"birth_info": {"ayanamsa_name": "raman", "node_mode": "mean"}, "modules": {"chart": {}}}
|
|
canon = canonicalize({"id": "fixture", "label": "x", "category": "smoke", "birth": {}}, data)
|
|
assert canon["parameters"]["ayanamsa"] == "raman"
|
|
assert canon["parameters"]["node_mode"] == "mean"
|
|
assert baseline_settings_error(canon, ayanamsa="raman", node_mode="mean") is None
|
|
assert "lahiri" in baseline_settings_error(canon, ayanamsa="lahiri", node_mode="mean")
|
|
assert "unrecorded" in baseline_settings_error({"parameters": {}}, ayanamsa="lahiri", node_mode="mean")
|
|
|
|
|
|
def test_baseline_passes_ayanamsa_and_node_mode_to_the_engine(monkeypatch, tmp_path):
|
|
from benchmarks.jyotish.scripts import run_skill_baseline as rsb
|
|
|
|
seen = {}
|
|
|
|
class Done:
|
|
returncode = 1
|
|
stdout = ""
|
|
stderr = "stub"
|
|
|
|
def fake_run(cmd, **kwargs):
|
|
seen["cmd"] = cmd
|
|
return Done()
|
|
|
|
monkeypatch.setattr(rsb, "RAW", tmp_path / "raw")
|
|
monkeypatch.setattr(rsb, "CANON", tmp_path / "canon")
|
|
monkeypatch.setattr(rsb.subprocess, "run", fake_run)
|
|
sample = {"id": "fixture", "birth": {"year": 2000, "month": 1, "day": 1, "hour": 0, "minute": 0, "lat": 0, "lon": 0, "tz": 0}}
|
|
rsb.run_sample(sample, ayanamsa="raman", node_mode="true")
|
|
cmd = seen["cmd"]
|
|
assert cmd[cmd.index("--ayanamsa") + 1] == "raman"
|
|
assert cmd[cmd.index("--node-mode") + 1] == "true"
|
|
rsb.run_sample(sample)
|
|
assert seen["cmd"][seen["cmd"].index("--ayanamsa") + 1] == "lahiri"
|
|
|
|
|
|
def test_fixed_samples_include_the_public_rodden_aa_backtest_figures():
|
|
import json
|
|
|
|
samples = json.loads((ROOT / "benchmarks" / "jyotish" / "data" / "benchmark_samples.json").read_text())
|
|
ids = {sample["id"] for sample in samples}
|
|
for figure in ["steve_jobs", "barack_obama", "elizabeth_taylor", "marilyn_monroe", "judy_garland", "edith_piaf", "frida_kahlo", "george_w_bush", "zinedine_zidane"]:
|
|
assert figure in ids
|
|
assert all(sample.get("privacy") in {"fictional_or_public_test", "public_rodden_aa"} for sample in samples)
|