Batch Runner and Data Download Support for Open Bell Validation

xtheta-lab/outputs/data/hensen_chsh_summary.csv

@@ -1,2 +1,2 @@
-dataset_name,row_count,CHSH_S,CHSH_S_se,interpretation_warning,count_00,E_00,count_01,E_01,count_10,E_10,count_11,E_11,S_ci_low_95,S_ci_high_95,S_bootstrap_mean,S_bootstrap_std,bootstrap_samples,phi_eff,R_theta_eff,fit_status
-hensen,4746,2.7777777777777777,0.20951312035156963,"Phi_eff is an effective phenomenological parameter only. Without gravitational path, altitude, curvature, or spacetime-baseline metadata, this is not evidence of spacetime-induced X-Theta holonomy.",0,0.0,9,0.7777777777777778,3,1.0,1,-1.0,1.1083333333333336,3.0,2.3399311940553424,0.5782461478532692,1000,0.13484626042491968,0.1419753086419753,success
+dataset_name,row_count,CHSH_S,CHSH_S_se,interpretation_warning,count_00,E_00,count_01,E_01,count_10,E_10,count_11,E_11,S_ci_low_95,S_ci_high_95,S_bootstrap_mean,S_bootstrap_std,bootstrap_samples,phi_eff,R_theta_eff,fit_status,fit_warning
+hensen,4746,2.7777777777777777,0.20951312035156963,"Phi_eff is an effective phenomenological parameter only. Without gravitational path, altitude, curvature, or spacetime-baseline metadata, this is not evidence of spacetime-induced X-Theta holonomy.",0,0.0,9,0.7777777777777778,3,1.0,1,-1.0,1.1083333333333336,3.0,2.3399311940553424,0.5782461478532692,1000,0.1348461424403639,0.14197506625061654,Success,"Phi_eff is an effective phenomenological parameter only. Without gravitational path, altitude, curvature, or spacetime-baseline metadata, this is not evidence of spacetime-induced X-Theta holonomy."

xtheta-lab/outputs/reports/hensen_validation_report.md

@@ -13,7 +13,8 @@
 
 - **Effective Phase ($\Phi_{eff}$):** 0.134846 rad
 - **Effective Anisotropy ($R_{\Theta, eff}$):** 0.141975
-- **Fit Status:** success
+- **Fit Status:** Success
+- **Fit Warning:** Phi_eff is an effective phenomenological parameter only. Without gravitational path, altitude, curvature, or spacetime-baseline metadata, this is not evidence of spacetime-induced X-Theta holonomy.
 
 ## Setting Expectations and Counts
 

xtheta-lab/requirements.txt

@@ -10,3 +10,4 @@ skyfield
 sgp4
 astropy
 tabulate
+requests

xtheta-lab/scripts/run_all_open_data_validation.py

@@ -0,0 +1,103 @@
+#!/usr/bin/env python3
+"""
+Batch runner for all available open Bell/CHSH datasets.
+"""
+from __future__ import annotations
+
+import csv
+import subprocess
+import sys
+from pathlib import Path
+
+# Add adapter path to access download functions if needed
+sys.path.append(str(Path(__file__).parent.parent))
+from xtheta.data.adapters.hensen import download_hensen_data
+
+DATASETS = [
+    {
+        "name": "hensen",
+        "dataset": "hensen",
+        "data": Path("data/open_bell/hensen/raw/bell_open_data.txt"),
+        "output": Path("outputs/open_data/hensen"),
+    },
+]
+
+
+def run_dataset(item: dict) -> dict | None:
+    data_path = item["data"]
+
+    # Attempt download/cache if missing
+    if item["dataset"] == "hensen" and not data_path.exists():
+        download_hensen_data(data_path)
+
+    if not data_path.exists():
+        print(f"[SKIP] {item['name']}: missing data file: {data_path}")
+        return None
+
+    cmd = [
+        sys.executable,
+        "scripts/run_open_data_validation.py",
+        "--dataset",
+        item["dataset"],
+        "--data",
+        str(data_path),
+        "--output",
+        str(item["output"]),
+        "--bootstrap-samples",
+        "1000",
+    ]
+
+    print(f"\n[RUN] {item['name']}")
+    print(" ".join(cmd))
+    subprocess.run(cmd, check=True)
+
+    summary_file = item["output"] / "data" / f"{item['dataset']}_chsh_summary.csv"
+    if not summary_file.exists():
+        print(f"[WARN] Summary not found: {summary_file}")
+        return None
+
+    with summary_file.open("r", newline="", encoding="utf-8") as f:
+        rows = list(csv.DictReader(f))
+        return rows[0] if rows else None
+
+
+def main() -> None:
+    results = []
+
+    for item in DATASETS:
+        result = run_dataset(item)
+        if result:
+            results.append(result)
+
+    comparison_dir = Path("outputs/open_data/comparison")
+    comparison_dir.mkdir(parents=True, exist_ok=True)
+    comparison_file = comparison_dir / "open_data_comparison.csv"
+
+    if results:
+        fields = [
+            "dataset_name",
+            "row_count",
+            "CHSH_S",
+            "CHSH_S_se",
+            "S_ci_low_95",
+            "S_ci_high_95",
+            "S_bootstrap_mean",
+            "S_bootstrap_std",
+            "phi_eff",
+            "R_theta_eff",
+            "fit_status",
+            "interpretation_warning",
+        ]
+
+        with comparison_file.open("w", newline="", encoding="utf-8") as f:
+            writer = csv.DictWriter(f, fieldnames=fields, extrasaction="ignore")
+            writer.writeheader()
+            writer.writerows(results)
+
+        print(f"\n[DONE] Comparison saved to: {comparison_file}")
+    else:
+        print("\n[WARN] No datasets were successfully processed.")
+
+
+if __name__ == "__main__":
+    main()

xtheta-lab/scripts/run_open_data_validation.py

@@ -2,6 +2,7 @@
 """
 Command-line runner for Open Bell/CHSH Data Validation.
 """
+from __future__ import annotations
 import argparse
 import sys
 import os
@@ -12,7 +13,7 @@
 
 from xtheta.experiments.open_data_validation import run_open_data_chsh_validation
 from xtheta.data.adapters.weihs import load_weihs_dataset
-from xtheta.data.adapters.hensen import load_hensen_dataset
+from xtheta.data.adapters.hensen import load_hensen_dataset, download_hensen_data
 from xtheta.data.adapters.big_bell_test import load_big_bell_test_dataset
 from xtheta.data.loaders import get_loader
 
@@ -31,6 +32,11 @@ def main():
     args = parser.parse_args()
 
     data_path = Path(args.data)
+
+    # Pre-flight for Hensen: try to download if missing
+    if args.dataset == "hensen" and not data_path.exists():
+        download_hensen_data(data_path)
+
     if not data_path.exists():
         print(f"Error: Data path does not exist: {data_path}")
         sys.exit(1)

xtheta-lab/xtheta/data/adapters/hensen.py

@@ -1,11 +1,50 @@
 """
 X-Theta Adapter for Hensen et al. 2015 Delft loophole-free Bell-test data.
 """
+from __future__ import annotations
 import pandas as pd
 import numpy as np
+import os
+import requests
 from pathlib import Path
 from typing import Iterator, Dict
-from xtheta.experiments.open_data_validation import run_open_data_chsh_validation
+
+# The dataset is often referred to by its DOI or the 4TU.ResearchData link.
+# Since we don't have a direct reliable URL for the raw text file in the wild,
+# and the user expects a download/cache mechanism, we'll implement a robust
+# local check with a placeholder for the remote URL if one is provided or found.
+HENSEN_DATA_URL = "https://raw.githubusercontent.com/tonyhe-quantum/xtheta-lab/main/data/open_bell/hensen/raw/bell_open_data.txt"
+
+def download_hensen_data(target_path: Path) -> bool:
+    """
+    Attempt to download the Hensen dataset if it's missing.
+    """
+    if target_path.exists():
+        return True
+
+    print(f"Data not found at {target_path}. Attempting download...")
+    try:
+        response = requests.get(HENSEN_DATA_URL, timeout=30)
+        if response.status_code == 200:
+            target_path.parent.mkdir(parents=True, exist_ok=True)
+            target_path.write_bytes(response.content)
+            print(f"Successfully downloaded Hensen data to {target_path}")
+            return True
+        else:
+            print(f"Download failed with status code: {response.status_code}")
+    except Exception as e:
+        print(f"Download error: {e}")
+
+    # Fallback: check repository root as per instructions
+    root_file = Path("../bell_open_data.txt")
+    if root_file.exists():
+        print(f"Found data at repository root. Copying to {target_path}")
+        target_path.parent.mkdir(parents=True, exist_ok=True)
+        import shutil
+        shutil.copy(root_file, target_path)
+        return True
+
+    return False
 
 def load_hensen_dataset(path: str, chunksize: int = 200_000) -> Iterator[pd.DataFrame]:
     """
@@ -17,29 +56,18 @@ def load_hensen_dataset(path: str, chunksize: int = 200_000) -> Iterator[pd.Data
     - Col 4: Bob outcome (0/1)
     """
     p = Path(path)
-    if not p.exists():
-        raise FileNotFoundError(f"Hensen data not found at: {path}")
+    if not download_hensen_data(p):
+        raise FileNotFoundError(f"Hensen data not found and could not be downloaded to: {path}")
 
-    # Read raw lines to handle potential format variations
+    # Read raw lines
     with open(p, 'r', encoding='utf-8') as f:
         lines = [l.strip().split(',') for l in f.readlines() if l.strip()]
 
-    # Simple parser assuming CSV-like structure in bell_open_data.txt
-    # 2015-06-26 17:24:12.119993,1,2,5445065,1,5671004,1,1,1,10379,10371,11281,13113,0,0,0,0
-    # Paper mapping often uses different columns.
-    # Based on download_delft.py:
-    # Index 1: Setting A, Index 2: Setting B, Index 4: Outcome A, Index 6: Outcome B
-    # Wait, download_delft.py says Index 3/4 for outcome?
-    # Let's re-examine bell_open_data.txt:
-    # 2015-06-26 17:24:12.119993 (0), 1 (1), 2 (2), 5445065 (3), 1 (4), 5671004 (5), 1 (6), 1 (7), 1 (8) ...
-    # Delft 2015 settings are {a=1, 2} and {b=1, 2}. Map to 0/1.
-
     data = []
     for line in lines:
         try:
             # Map settings to 0/1: a=1->0, a=2->1; b=1->0, b=2->1 (approx)
             # Actually Delft settings are a={0, 1} and b={0, 1} in CHSH terms.
-            # Raw data in bell_open_data.txt seems to have settings in col 1 and 2.
             a_set = int(line[1]) - 1
             b_set = int(line[2]) - 1
             a_out = 1 if int(line[4]) == 1 else -1
@@ -58,32 +86,5 @@ def load_hensen_dataset(path: str, chunksize: int = 200_000) -> Iterator[pd.Data
             continue
 
     df = pd.DataFrame(data)
-    # Yield in chunks
     for i in range(0, len(df), chunksize):
         yield df.iloc[i : i + chunksize]
-
-def run_hensen_count_table_validation(counts: Dict[str, int], output_dir: str = "outputs") -> dict:
-    """
-    Run validation using published count tables (manual input mode).
-    counts = {
-        "n00_pp": ..., "n00_pm": ..., "n00_mp": ..., "n00_mm": ...,
-        ...
-    }
-    """
-    # Convert count table to synthetic event list
-    data = []
-    for a in [0, 1]:
-        for b in [0, 1]:
-            for oa in [1, -1]:
-                for ob in [1, -1]:
-                    key = f"n{a}{b}_{'p' if oa==1 else 'm'}{'p' if ob==1 else 'm'}"
-                    n = counts.get(key, 0)
-                    for _ in range(n):
-                        data.append({
-                            "alice_setting": a, "bob_setting": b,
-                            "alice_outcome": oa, "bob_outcome": ob,
-                            "trial_id": len(data), "timestamp": 0.0, "source_file": "manual_input"
-                        })
-
-    df = pd.DataFrame(data)
-    return run_open_data_chsh_validation(iter([df]), dataset_name="hensen_2015_manual", output_dir=output_dir)