Merge pull request #2198 from NNPDF/update_th_covmat

Sum covmats from various point prescriptions

.github/workflows/tests.yml

@@ -13,9 +13,6 @@ jobs:
       matrix:
         os: [ubuntu-latest, macos-14]
         python-version: ["3.12"] # Test only the latest python in all system
-        include:
-          - os: ubuntu-latest # Test also the oldest python supported in ubuntu
-            python-version: "3.9"
       fail-fast: false
     runs-on: ${{ matrix.os }}
     env:
@@ -54,7 +51,7 @@ jobs:
           $CONDA_PREFIX/conda-bld/noarch/*.tar.bz2 \
           [email protected]:~/packages/conda/noarch
     - name: Build and upload sphinx documentation to NNPDF server
-      if: startsWith(matrix.os, 'ubuntu') && github.ref == 'refs/heads/master' && matrix.python-version == '3.9'
+      if: startsWith(matrix.os, 'ubuntu') && github.ref == 'refs/heads/master'
       shell: bash -l {0}
       run: |
         KEY=$( mktemp )

doc/sphinx/source/tutorials/general_th_covmat.rst

@@ -41,8 +41,6 @@ Instructions
 
     ############################################################################
     theorycovmatconfig:
-      use_scalevar_uncertainties: False
-      use_user_uncertainties: True
       user_covmat_path: "IeGM9CY8RxGcb5r6bIEYlQ==/topthcovmat.csv"
       use_thcovmat_in_sampling: True
       use_thcovmat_in_fitting: True
@@ -57,16 +55,10 @@ generation). The default is ``True`` for both.
       Changing either of these to ``False`` will affect the fit outcome and should
       be avoided unless you know what you are doing.
 
-
-4. The flag ``use_scalevar_uncertainties``  corresponds to whether or not a
-   theory covmat is to be constructed from a pointprescription with instructions
-   inside ``theorycovmatconfig``. Usually in cases where a custom covmat is
-   being loaded, this will be set to ``False``.
-
-5. Make sure that datasets are grouped under one big experiment called "BIGEXP",
+1. Make sure that datasets are grouped under one big experiment called "BIGEXP",
    just like in :ref:`vptheorycov-index`.
 
-6. For an example runcard, see `here <https://github.com/NNPDF/nnpdf/tree/master/validphys2/examples/fit_with_user_covmat.yaml.>`_
+2. For an example runcard, see `here <https://github.com/NNPDF/nnpdf/tree/master/validphys2/examples/fit_with_user_covmat.yaml.>`_
 
 
 Including both scale variation uncertainties and user uncertainties

n3fit/runcards/examples/Fit_with_theory_covmat.yml

@@ -27,9 +27,7 @@ datacuts:
 theory:
   theoryid: 708        # database id
 theorycovmatconfig:
-  point_prescription: "9 point"
-  theoryids:
-    from_: scale_variation_theories
+  point_prescriptions: ["9 point", "3 point"]
   pdf: NNPDF31_nlo_as_0118
   use_thcovmat_in_fitting: true
   use_thcovmat_in_sampling: true

n3fit/runcards/examples/nnpdfpol20_dis_nlo_mhou.yml

@@ -78,9 +78,7 @@ theory:
   theoryid: 41_100_010
 
 theorycovmatconfig:
-  point_prescription: '7 point'
-  theoryids:
-    from_: scale_variation_theories
+  point_prescriptions: ['7 point']
   pdf: NNPDFpol11_100
   unpolarized_bc: NNPDF40_nlo_pch_as_01180
   use_thcovmat_in_fitting: true

n3fit/src/n3fit/scripts/n3fit_exec.py

@@ -156,7 +156,6 @@ def from_yaml(cls, o, *args, **kwargs):
             N3FIT_FIXED_CONFIG['positivity_bound'] = None
 
         # Theorycovmat flags and defaults
-        N3FIT_FIXED_CONFIG['theory_covmat_flag'] = False
         N3FIT_FIXED_CONFIG['use_thcovmat_in_fitting'] = False
         N3FIT_FIXED_CONFIG['use_thcovmat_in_sampling'] = False
         if (thconfig := file_content.get('theorycovmatconfig')) is not None:
@@ -166,17 +165,6 @@ def from_yaml(cls, o, *args, **kwargs):
             N3FIT_FIXED_CONFIG['use_thcovmat_in_sampling'] = thconfig.get(
                 'use_thcovmat_in_sampling', True
             )
-            if (
-                N3FIT_FIXED_CONFIG['use_thcovmat_in_sampling']
-                or N3FIT_FIXED_CONFIG['use_thcovmat_in_fitting']
-            ):
-                N3FIT_FIXED_CONFIG['theory_covmat_flag'] = True
-            N3FIT_FIXED_CONFIG['use_user_uncertainties'] = thconfig.get(
-                'use_user_uncertainties', False
-            )
-            N3FIT_FIXED_CONFIG['use_scalevar_uncertainties'] = thconfig.get(
-                'use_scalevar_uncertainties', True
-            )
         # Fitting flag
         file_content.update(N3FIT_FIXED_CONFIG)
         return cls(file_content, *args, **kwargs)

n3fit/src/n3fit/scripts/vp_setupfit.py

@@ -153,7 +153,14 @@ def from_yaml(cls, o, *args, **kwargs):
             filter_action = 'datacuts::theory::fitting filter'
             check_n3fit_action = 'datacuts::theory::fitting n3fit_checks_action'
         SETUPFIT_FIXED_CONFIG['actions_'] += [check_n3fit_action, filter_action]
-        if file_content.get('theorycovmatconfig') is not None:
+        if (thconfig := file_content.get('theorycovmatconfig')) is not None:
+            if thconfig.get('point_prescription') is not None:
+                raise ConfigError(
+                    "`point_prescription` has been removed in favor of a list of "
+                    "`point_prescriptions`. The options that can be included in the list are found "
+                    "in pointprescriptions.yaml. E.g. \n"
+                    "`point_prescriptions: ['9 point', '3 point']`"
+                )
             SETUPFIT_FIXED_CONFIG['actions_'].append(
                 'datacuts::theory::theorycovmatconfig nnfit_theory_covmat'
             )

validphys2/examples/data_theory_comparison_w_sv.yaml

@@ -7,8 +7,6 @@ use_cuts: "internal"
 
 theoryid: 717 # define the central theory
 point_prescription: "3 point"
-theoryids:
-  from_: scale_variation_theories
 use_theorycovmat: true
 
 use_pdferr: true

validphys2/examples/theory_covariance/covariance_matrix_plots_3pt.yaml

@@ -10,8 +10,6 @@ theoryid: 163
 point_prescription: '3 point'
 
 
-theoryids:
-   from_: scale_variation_theories
 
 dataspecs:
         - theoryid: 163

validphys2/examples/theory_covariance/covariance_matrix_plots_9pt_reduced.yaml

@@ -9,9 +9,6 @@ default_theory:
 theoryid: 163
 point_prescription: '9 point'
 
-theoryids:
-   from_: scale_variation_theories
-
 
 dataspecs:
         - theoryid: 163

validphys2/examples/theory_covariance/validation_report_3pt.yaml

@@ -46,8 +46,6 @@ theoryconfig:  # For generating theory covariance matrix
 
    theoryid: 163
 
-   theoryids:
-     from_: scale_variation_theories
 
    point_prescription: '3 point'
 

validphys2/examples/theory_covariance/validation_report_9pt.yaml

@@ -46,9 +46,6 @@ theoryconfig:  # For generating theory covariance matrix
 
    theoryid: 163
 
-   theoryids:
-     from_: scale_variation_theories
-
    point_prescription: '9 point'
 
    use_cuts: fromfit

validphys2/src/validphys/config.py

@@ -210,15 +210,8 @@ def produce_replicas(self, nreplica: int):
         """Produce a replicas array"""
         return NSList(range(1, nreplica + 1), nskey="replica")
 
-    def produce_inclusive_use_scalevar_uncertainties(
-        self, use_scalevar_uncertainties: bool = False, point_prescription: (str, None) = None
-    ):
-        """Whether to use a scale variation uncertainty theory covmat.
-        Checks whether a point prescription is included in the runcard and if so
-        assumes scale uncertainties are to be used."""
-        if (not use_scalevar_uncertainties) and (point_prescription is not None):
-            use_scalevar_uncertainties = True
-        return use_scalevar_uncertainties
+    def parse_point_prescriptions(self, point_prescriptions):
+        return NSList(point_prescriptions, nskey="point_prescription")
 
     # TODO: load fit config from here
     @element_of("fits")
@@ -736,9 +729,7 @@ def produce_experiment_from_input(self, experiment_input, theoryid, use_cuts, fi
         }
 
     @configparser.explicit_node
-    def produce_dataset_inputs_fitting_covmat(
-        self, theory_covmat_flag=False, use_thcovmat_in_fitting=False
-    ):
+    def produce_dataset_inputs_fitting_covmat(self, use_thcovmat_in_fitting=False):
         """
         Produces the correct covmat to be used in fitting_data_dict according
         to some options: whether to include the theory covmat, whether to
@@ -747,7 +738,7 @@ def produce_dataset_inputs_fitting_covmat(
         """
         from validphys import covmats
 
-        if theory_covmat_flag and use_thcovmat_in_fitting:
+        if use_thcovmat_in_fitting:
             return covmats.dataset_inputs_t0_total_covmat
         return covmats.dataset_inputs_t0_exp_covmat
 
@@ -758,7 +749,7 @@ def produce_sep_mult(self, separate_multiplicative=False):
 
     @configparser.explicit_node
     def produce_dataset_inputs_sampling_covmat(
-        self, sep_mult=False, theory_covmat_flag=False, use_thcovmat_in_sampling=False
+        self, sep_mult=False, use_thcovmat_in_sampling=False
     ):
         """
         Produces the correct covmat to be used in make_replica according
@@ -767,7 +758,7 @@ def produce_dataset_inputs_sampling_covmat(
         """
         from validphys import covmats
 
-        if theory_covmat_flag and use_thcovmat_in_sampling:
+        if use_thcovmat_in_sampling:
             if sep_mult:
                 return covmats.dataset_inputs_total_covmat_separate
             else:
@@ -782,20 +773,21 @@ def produce_loaded_theory_covmat(
         self,
         output_path,
         data_input,
-        theory_covmat_flag=False,
-        use_user_uncertainties=False,
-        use_scalevar_uncertainties=True,
+        user_covmat_path=None,
+        point_prescriptions=None,
+        use_thcovmat_in_sampling=False,
+        use_thcovmat_in_fitting=False,
     ):
         """
         Loads the theory covmat from the correct file according to how it
         was generated by vp-setupfit.
         """
-        if theory_covmat_flag is False:
+        if not use_thcovmat_in_sampling and not use_thcovmat_in_fitting:
             return 0.0
         # Load correct file according to how the thcovmat was generated by vp-setupfit
         generic_path = "datacuts_theory_theorycovmatconfig_theory_covmat_custom.csv"
-        if use_user_uncertainties is True:
-            if use_scalevar_uncertainties is True:
+        if user_covmat_path is not None:
+            if point_prescriptions is not None:
                 generic_path = "datacuts_theory_theorycovmatconfig_total_theory_covmat.csv"
             else:
                 generic_path = "datacuts_theory_theorycovmatconfig_user_covmat.csv"
@@ -1152,15 +1144,13 @@ def parse_lumi_channel(self, ch: str):
     def produce_all_lumi_channels(self):
         return {"lumi_channels": self.parse_lumi_channels(list(LUMI_CHANNELS))}
 
-    def produce_loaded_user_covmat_path(
-        self, user_covmat_path: str = "", use_user_uncertainties: bool = False
-    ):
+    def produce_loaded_user_covmat_path(self, user_covmat_path: str = ""):
         """
         Path to the user covmat provided by user_covmat_path in the runcard.
         If no path is provided, returns None.
         For use in theorycovariance.construction.user_covmat.
         """
-        if not use_user_uncertainties:
+        if user_covmat_path is None:
             return None
         else:
             l = self.loader
@@ -1169,26 +1159,26 @@ def produce_loaded_user_covmat_path(
 
     @configparser.explicit_node
     def produce_nnfit_theory_covmat(
-        self, inclusive_use_scalevar_uncertainties, use_user_uncertainties: bool = False
+        self, point_prescriptions: list = None, user_covmat_path: str = None
     ):
         """
         Return the theory covariance matrix used in the fit.
 
         This function is only used in vp-setupfit to store the necessary covmats as .csv files in
         the tables directory.
         """
-        if inclusive_use_scalevar_uncertainties:
-            if use_user_uncertainties:
+        if point_prescriptions is not None:
+            if user_covmat_path is not None:
                 # Both scalevar and user uncertainties
                 from validphys.theorycovariance.construction import total_theory_covmat_fitting
 
                 f = total_theory_covmat_fitting
             else:
                 # Only scalevar uncertainties
-                from validphys.theorycovariance.construction import theory_covmat_custom_fitting
+                from validphys.theorycovariance.construction import theory_covmat_custom
 
-                f = theory_covmat_custom_fitting
-        elif use_user_uncertainties:
+                f = theory_covmat_custom
+        elif user_covmat_path is not None:
             # Only user uncertainties
             from validphys.theorycovariance.construction import user_covmat_fitting
 
@@ -1223,21 +1213,17 @@ def produce_fitthcovmat(
                 thcovmat_present = False
 
         if use_thcovmat_if_present and thcovmat_present:
-            # Expected directory of theory covmat hardcoded
-            covmat_path = fit.path / "tables"
+            # Expected path of theory covmat hardcoded
+            covmat_path = (
+                fit.path / "tables" / "datacuts_theory_theorycovmatconfig_theory_covmat_custom.csv"
+            )
             # All possible valid files
-            covfiles = sorted(covmat_path.glob("*theory_covmat*.csv"))
-            if not covfiles:
+            if not covmat_path.exists():
                 raise ConfigError(
                     "Fit appeared to use theory covmat in fit but the file was not at the "
                     f"usual location: {covmat_path}."
                 )
-            if len(covfiles) > 1:
-                raise ConfigError(
-                    "More than one valid theory covmat file found at the "
-                    f"usual location: {covmat_path}. These are {covfiles}."
-                )
-            fit_theory_covmat = ThCovMatSpec(covfiles[0])
+            fit_theory_covmat = ThCovMatSpec(covmat_path)
         else:
             fit_theory_covmat = None
         return fit_theory_covmat
@@ -1671,11 +1657,10 @@ def produce_group_dataset_inputs_by_experiment(self, data_input):
     def produce_group_dataset_inputs_by_process(self, data_input):
         return self.produce_group_dataset_inputs_by_metadata(data_input, "nnpdf31_process")
 
-    def produce_scale_variation_theories(self, point_prescription, t0id):
+    def produce_theoryids(self, t0id, point_prescription):
         """Produces a list of theoryids given a theoryid at central scales and a point
         prescription. The options for the latter are defined in pointprescriptions.yaml.
         This hard codes the theories needed for each prescription to avoid user error."""
-        pp = point_prescription
         th = t0id.id
 
         lsv = yaml.safe_load(read_text(validphys.scalevariations, "scalevariationtheoryids.yaml"))
@@ -1697,7 +1682,7 @@ def produce_scale_variation_theories(self, point_prescription, t0id):
         )
 
         try:
-            scales = pp_scales_dict[pp]
+            scales = pp_scales_dict[point_prescription]
         except KeyError:
             valid_pps = ", ".join(pp_scales_dict.keys())
             raise ConfigError(
@@ -1727,13 +1712,13 @@ def produce_scale_variation_theories(self, point_prescription, t0id):
                 + "available. To use this point prescription for this central theoryid, theoryids "
                 + "that correspond to the following scale choices must be created and added to "
                 + "validphys2/src/validphys/scalevariations/scalevariationtheoryids.yaml: "
-                + f"(k_F, k_R) = {missing_scales_string}."
+                + f"missing variations = {missing_scales_string}."
             )
 
         # Check each theory is loaded
         theoryids = [self.loader.check_theoryID(thid) for thid in thids]
-        # NSList needs to be used for theoryids to be recognised as a namespace
-        return {"theoryids": NSList(theoryids, nskey="theoryid")}
+
+        return NSList(theoryids, nskey="theoryid")
 
     @configparser.explicit_node
     def produce_filter_data(self, fakedata: bool = False, theorycovmatconfig=None):