Add HH model variations

This PR adds variations of the default HH physics model, each with different different selection in ggF and VBF samples. This is useful for testing the best possible combination of samples per analysis or even globally.

Also, this PR contains 2 further commits (improving the datacard manipulation scripts) which I pushed on this branch by accident. I guess this is fine for now.

dhi/models/HHModelPinv.py

@@ -292,7 +292,7 @@ class HHModel(PhysicsModel):
             couplings_in_expr = []
             if 'kl'  in s_expr: couplings_in_expr.append('kl')
             if 'kt'  in s_expr: couplings_in_expr.append('kt')
-            
+
             # no constant expressions are expected
             if len(couplings_in_expr) == 0:
                 raise RuntimeError('GGF HH : scaling expression has no coefficients')
@@ -426,51 +426,128 @@ class HHModel(PhysicsModel):
 
 #########################################
 
+# ggf samples with keys (kl, kt), ordered by kl
+ggf_samples = OrderedDict([
+    ((0, 1),    GGFHHSample(0,    1, val_xs=0.06007, label="ggHH_kl_0_kt_1")),
+    ((1, 1),    GGFHHSample(1,    1, val_xs=0.02675, label="ggHH_kl_1_kt_1")),
+    ((2.45, 1), GGFHHSample(2.45, 1, val_xs=0.01133, label="ggHH_kl_2p45_kt_1")),
+    ((5, 1),    GGFHHSample(5,    1, val_xs=0.07903, label="ggHH_kl_5_kt_1")),
 
-# GGF : val_kl, val_kt
-GGF_sample_list = [
-    GGFHHSample(1,1,      val_xs = 0.02675, label = 'ggHH_kl_1_kt_1'  ),
-    # GGFHHSample(0,1,      val_xs = 0.06007, label = 'ggHH_kl_0_kt_1'  ),
-    GGFHHSample(5,1,      val_xs = 0.07903, label = 'ggHH_kl_5_kt_1'  ),
-    GGFHHSample(2.45,1,   val_xs = 0.01133, label = 'ggHH_kl_2p45_kt_1'  ),
-]
+])
 
-# VBF : val_CV, val_C2V, val_kl
-VBF_sample_list = [
-    VBFHHSample(1,1,1,   val_xs = 0.00054/(0.3364), label = 'qqHH_CV_1_C2V_1_kl_1'  ),
-    VBFHHSample(1,2,1,   val_xs = 0.00472/(0.3364), label = 'qqHH_CV_1_C2V_2_kl_1'  ),
-    VBFHHSample(1,1,2,   val_xs = 0.00044/(0.3364), label = 'qqHH_CV_1_C2V_1_kl_2'  ),
-    VBFHHSample(1,1,0,   val_xs = 0.00145/(0.3364), label = 'qqHH_CV_1_C2V_1_kl_0'  ),
-    VBFHHSample(0.5,1,1, val_xs = 0.00353/(0.3364), label = 'qqHH_CV_0p5_C2V_1_kl_1'),
-    VBFHHSample(1.5,1,1, val_xs = 0.02149/(0.3364), label = 'qqHH_CV_1p5_C2V_1_kl_1'),
-]
+# vbf samples with keys (CV, C2V, kl), SM point first, then ordered by kl, then C2V, then CV
+vbf_samples = OrderedDict([
+    ((1,   1, 1), VBFHHSample(1,   1, 1, val_xs=0.00054 / 0.3364, label="qqHH_CV_1_C2V_1_kl_1")),
+    ((1,   1, 0), VBFHHSample(1,   1, 0, val_xs=0.00145 / 0.3364, label="qqHH_CV_1_C2V_1_kl_0")),
+    ((1,   1, 2), VBFHHSample(1,   1, 2, val_xs=0.00044 / 0.3364, label="qqHH_CV_1_C2V_1_kl_2")),
+    ((1,   0, 1), VBFHHSample(1,   0, 1, val_xs=0.00880 / 0.3364, label="qqHH_CV_1_C2V_0_kl_1")),
+    ((1,   2, 1), VBFHHSample(1,   2, 1, val_xs=0.00472 / 0.3364, label="qqHH_CV_1_C2V_2_kl_1")),
+    ((0.5, 1, 1), VBFHHSample(0.5, 1, 1, val_xs=0.00353 / 0.3364, label="qqHH_CV_0p5_C2V_1_kl_1")),
+    ((1.5, 1, 1), VBFHHSample(1.5, 1, 1, val_xs=0.02149 / 0.3364, label="qqHH_CV_1p5_C2V_1_kl_1")),
+])
 
-HHdefault = HHModel(
-    ggf_sample_list = GGF_sample_list,
-    vbf_sample_list = VBF_sample_list,
-    name            = 'HHdefault'
-)
 
-#bbZZ_model = False
+def get_ggf_samples(keys):
+    """
+    Returns a list of :py:class:`GGFHHSample` instances mapped to certain *keys* in the ordered
+    *ggf_samples* dictionary above. A key can either be a tuple of (kl, kt) values as used in dict,
+    or an index. Example:
 
-#if(bbZZ_model):
+    .. code-block:: python
 
-#    GGF_sample_list_bbZZ = [
-#        GGFHHSample(1,1,   val_xs = 0.02675, label = 'ggHH_hbbhzz_process'  ),
-#        GGFHHSample(2,1,   val_xs = 0.01238, label = 'klambda2_process'  ),
-        #GGFHHSample(0,1,   val_xs = 0.06007, label = 'klambda0_process'  ),
-#        GGFHHSample(5,1,   val_xs = 0.07903, label = 'klambda5_process'  ),
-        # GGFHHSample(2.45,1,   val_xs = 0.01133, label = 'ggHH_kl_2p45_kt_1'  ),
-#    ]
+        get_ggf_samples([2, (5, 1)])
+        # -> [GGFHHSample:ggHH_kl_2p45_kt_1, GGFHHSample:ggHH_kl_5_kt_1]
+    """
+    keys = [
+        (key if isinstance(key, tuple) else ggf_samples.keys()[key])
+        for key in keys
+    ]
+    return [ggf_samples[key] for key in keys]
 
-#    HHbbZZ4l = HHModel(
-#        ggf_sample_list = GGF_sample_list_bbZZ,
-#        vbf_sample_list = VBF_sample_list,
-#        name            = 'HH_bbZZ_default'
-#    )
 
-# f = GGFHHFormula(GGF_sample_list)
-# f = VBFHHFormula(VBF_sample_list)
+def get_vbf_samples(keys):
+    """
+    Returns a list of :py:class:`VBFHHSample` instances mapped to certain *keys* in the ordered
+    *vbf_samples* dictionary above. A key can either be a tuple of (CV, C2V, kl) values as used in
+    dict, or an index. Example:
+
+    .. code-block:: python
+
+        get_vbf_samples([2, (1, 2, 1)])
+        # -> [VBFHHSample:qqHH_CV_1_C2V_1_kl_2, VBFHHSample:qqHH_CV_1_C2V_2_kl_1]
+    """
+    keys = [
+        (key if isinstance(key, tuple) else vbf_samples.keys()[key])
+        for key in keys
+    ]
+    return [vbf_samples[key] for key in keys]
+
+
+def create_model(name, skip_ggf=None, skip_vbf=None):
+    """
+    Returns a new :py:class:`HHModel` instance named *name*. Its ggf and vbf sample lists are
+    using all availabe samples except those defined in *skip_ggf* and *skip_vbf*, respectively,
+    through :py:func:`get_ggf_samples` and :py:func:`get_vbf_samples`. The order of passed keys to
+    be skipped does not matter.
+    """
+    # get all use sample keys
+    ggf_keys = list(ggf_samples.keys())
+    vbf_keys = list(vbf_samples.keys())
+    for key in skip_ggf or []:
+        ggf_keys.remove(key)
+    for key in skip_vbf or []:
+        vbf_keys.remove(key)
+
+    # create the return the model
+    return HHModel(
+        ggf_sample_list=get_ggf_samples(ggf_keys),
+        vbf_sample_list=get_vbf_samples(vbf_keys),
+        name=name,
+    )
+
+
+# standard of models
+model_all     = create_model("model_all")
+model_default = create_model("model_default", skip_ggf=[(0, 1)], skip_vbf=[(1, 0, 1)])
+model_bbgg    = create_model("model_bbgg",    skip_ggf=[(0, 1)], skip_vbf=[(0.5, 1, 1)])
+
+# ggf test models
+model_no_ggf_kl0    = create_model("model_no_ggf_kl0",    skip_ggf=[(0, 1)],    skip_vbf=[(1, 0, 1)])
+model_no_ggf_kl1    = create_model("model_no_ggf_kl1",    skip_ggf=[(1, 1)],    skip_vbf=[(1, 0, 1)])
+model_no_ggf_kl2p45 = create_model("model_no_ggf_kl2p45", skip_ggf=[(2.45, 1)], skip_vbf=[(1, 0, 1)])
+model_no_ggf_kl5    = create_model("model_no_ggf_kl5",    skip_ggf=[(5, 1)],    skip_vbf=[(1, 0, 1)])
+
+# vbf test models
+model_no_vbf_sm    = create_model("model_no_vbf_sm",    skip_ggf=[(0, 1)], skip_vbf=[(1, 1, 1)])
+model_no_vbf_kl0   = create_model("model_no_vbf_kl0",   skip_ggf=[(0, 1)], skip_vbf=[(1, 1, 0)])
+model_no_vbf_kl2   = create_model("model_no_vbf_kl2",   skip_ggf=[(0, 1)], skip_vbf=[(1, 1, 2)])
+model_no_vbf_C2V0  = create_model("model_no_vbf_C2V0",  skip_ggf=[(0, 1)], skip_vbf=[(1, 0, 1)])
+model_no_vbf_C2V2  = create_model("model_no_vbf_C2V2",  skip_ggf=[(0, 1)], skip_vbf=[(1, 2, 1)])
+model_no_vbf_CV0p5 = create_model("model_no_vbf_CV0p5", skip_ggf=[(0, 1)], skip_vbf=[(0.5, 1, 1)])
+model_no_vbf_CV1p5 = create_model("model_no_vbf_CV1p5", skip_ggf=[(0, 1)], skip_vbf=[(1.5, 1, 1)])
+
+# legacy objects for development
+GGF_sample_list = [
+    # GGFHHSample(0,    1, val_xs=0.06007, label="ggHH_kl_0_kt_1"),
+    GGFHHSample(1,    1, val_xs=0.02675, label="ggHH_kl_1_kt_1"),
+    GGFHHSample(2.45, 1, val_xs=0.01133, label="ggHH_kl_2p45_kt_1"),
+    GGFHHSample(5,    1, val_xs=0.07903, label="ggHH_kl_5_kt_1"),
+]
+VBF_sample_list = [
+    VBFHHSample(1,   1, 1, val_xs=0.00054 / 0.3364, label="qqHH_CV_1_C2V_1_kl_1"),
+    VBFHHSample(1,   1, 0, val_xs=0.00145 / 0.3364, label="qqHH_CV_1_C2V_1_kl_0"),
+    VBFHHSample(1,   1, 2, val_xs=0.00044 / 0.3364, label="qqHH_CV_1_C2V_1_kl_2"),
+    # VBFHHSample(1,   0, 1, val_xs=0.00880 / 0.3364, label="qqHH_CV_1_C2V_0_kl_1"),
+    VBFHHSample(1,   2, 1, val_xs=0.00472 / 0.3364, label="qqHH_CV_1_C2V_2_kl_1"),
+    VBFHHSample(0.5, 1, 1, val_xs=0.00353 / 0.3364, label="qqHH_CV_0p5_C2V_1_kl_1"),
+    VBFHHSample(1.5, 1, 1, val_xs=0.02149 / 0.3364, label="qqHH_CV_1p5_C2V_1_kl_1"),
+]
+HHdefault = HHModel(
+    ggf_sample_list=GGF_sample_list,
+    vbf_sample_list=VBF_sample_list,
+    name="HHdefault",
+)
+
 
 def create_ggf_xsec_func(formula=None, nnlo=True):
     """
@@ -543,3 +620,26 @@ get_ggf_xsec = create_ggf_xsec_func()
 
 #: Default function for getting VBF cross sections using the formula of the HHdefault model.
 get_vbf_xsec = create_vbf_xsec_func()
+
+
+# debug
+#bbZZ_model = False
+
+#if(bbZZ_model):
+
+#    GGF_sample_list_bbZZ = [
+#        GGFHHSample(1,1,   val_xs = 0.02675, label = 'ggHH_hbbhzz_process'  ),
+#        GGFHHSample(2,1,   val_xs = 0.01238, label = 'klambda2_process'  ),
+        #GGFHHSample(0,1,   val_xs = 0.06007, label = 'klambda0_process'  ),
+#        GGFHHSample(5,1,   val_xs = 0.07903, label = 'klambda5_process'  ),
+        # GGFHHSample(2.45,1,   val_xs = 0.01133, label = 'ggHH_kl_2p45_kt_1'  ),
+#    ]
+
+#    HHbbZZ4l = HHModel(
+#        ggf_sample_list = GGF_sample_list_bbZZ,
+#        vbf_sample_list = VBF_sample_list,
+#        name            = 'HH_bbZZ_default'
+#    )
+
+# f = GGFHHFormula(GGF_sample_list)
+# f = VBFHHFormula(VBF_sample_list)