Alphabet and param scan dev

.gitignore

@@ -191,3 +191,7 @@ docker_stderror
 data/.sites_map.json
 
 *.C
+
+src/bbtautau/postprocessing/templates/
+src/bbtautau/combine/bbtautau_ggfvbf.code-workspace
+src/bbtautau/combine/binder/

scripts/extract_qcd_tf.py

@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+"""
+Extract QCD transfer factor values from FitShapes.root.
+Usage: python scripts/extract_qcd_tf.py <FitShapes.root> [--region ggfbbtthmpass]
+"""
+import argparse
+import sys
+
+import numpy as np
+
+try:
+    import uproot
+except ImportError:
+    print("ERROR: uproot required. Install with: pip install uproot")
+    sys.exit(1)
+
+QCD_KEY = "CMS_bbtautau_boosted_qcd_datadriven"
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("fit_shapes", help="Path to FitShapes.root")
+    parser.add_argument("--region", default="ggfbbtthmpass", help="Pass region (e.g. ggfbbtthmpass)")
+    args = parser.parse_args()
+
+    f = uproot.open(args.fit_shapes)
+    pass_region = args.region
+    fail_region = pass_region.replace("pass", "fail")
+
+    pass_dir = f"{pass_region}_postfit"
+    fail_dir = f"{fail_region}_postfit"
+
+    try:
+        pass_qcd = f[f"{pass_dir}/{QCD_KEY}"]
+        fail_qcd = f[f"{fail_dir}/{QCD_KEY}"]
+    except KeyError as e:
+        print(f"ERROR: {e}")
+        print("Available keys:", [k for k in f.keys() if "postfit" in k][:20])
+        sys.exit(1)
+
+    pass_vals = np.array(pass_qcd.values())
+    fail_vals = np.array(fail_qcd.values())
+    axis = pass_qcd.axes[0]
+    edges = np.array(axis.edges()) if callable(getattr(axis, "edges", None)) else np.array(axis.edges)
+    centers = (edges[:-1] + edges[1:]) / 2
+
+    tf = np.where(fail_vals > 0, pass_vals / fail_vals, np.nan)
+
+    print(f"QCD Transfer Factor: {pass_region} / {fail_region}")
+    print("=" * 50)
+    print(f"{'Bin center':>12} | {'TF':>12} | Pass QCD  | Fail QCD")
+    print("-" * 50)
+    for i in range(len(centers)):
+        print(f"{centers[i]:12.4f} | {tf[i]:12.6e} | {pass_vals[i]:9.4e} | {fail_vals[i]:9.4e}")
+    print("-" * 50)
+    valid = ~np.isnan(tf)
+    if np.any(valid):
+        print(f"Min: {np.nanmin(tf):.6e}  Max: {np.nanmax(tf):.6e}  Mean: {np.nanmean(tf):.6e}  Std: {np.nanstd(tf):.6e}")
+        cv = np.nanstd(tf) / np.nanmean(tf) * 100 if np.nanmean(tf) != 0 else 0
+        print(f"Coefficient of variation: {cv:.2f}%")
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())