Provide an unique truth for all theory cards

nnpdf_data/nnpdf_data/theory.py

@@ -0,0 +1,76 @@
+"""
+This module provides an unique source and definition for all the possible parameters
+that a theory card can contain.
+"""
+
+import dataclasses
+
+DEPRECATED_KEYS = ["MaxNfAs", "SxRes", "SxOrd" "EScaleVar", "Qedref", "global_nx"]
+
+
+@dataclasses.dataclass(frozen=True)
+class TheoryCard:
+    ID: int  # ID number of the theory
+    PTO: int  # Perturbative order (0 = LO, 1 = NLO, 2 = NNLO ...)
+    FNS: str  # Flavor number scheme (i.e. FONLL-C)
+    DAMP: int  # Whether a damping function is applied or not for FONLL
+    IC: int  # 0 = perturbative charm only , 1 = intrinsic charm allowed
+    ModEv: str  # DGLAP evolution solution method (EXA or TRN)
+    XIR: float  # Renormalization scale over the hard scattering scale ratio
+    XIF: float  # Factorization scale over the hard scattering scale ratio
+    NfFF: int  # Number of active flavors, only for FFNS or FFN0 schemes
+    QED: int  # QED correction to strong coupling: 0 = disabled , 1 = allowed
+    HQ: str  # Heavy quark mass scheme,  POLE for pole masses (default), MSBAR for running masses (used only in Eko).
+    mc: float  # [GeV] charm mass
+    Qmc: float  # [GeV] MSbar mass reference scale of the charm
+    kcThr: float  # Threshold ratio of the charm
+    mb: float  # # [GeV] bottom mass
+    Qmb: float  # [GeV] MSbar mass reference scale of the bottom
+    kbThr: float  # Threshold ratio of the bottom
+    mt: float  # # [GeV] top mass
+    Qmt: float  # [GeV] MSbar mass reference scale of the top
+    ktThr: float  # Threshold ratio of the top
+    CKM: list[float]  # CKM matrix elements
+    MZ: float  # [GeV] Mass of Z
+    MW: float  # [GeV] Mass of W
+    GF: float  # Fermi constant
+    SIN2TW: float
+    TMC: int  # Include target mass corrections: 0 = disabled, 1 = leading twist, 2 = higher twist approximated, 3 = higher twist exact
+    MP: float  # [GeV] Mass of the proton
+    Comments: str  # Comments on the theory
+    MaxNfPdf: int = 5  # Used by pineko to define the thresholds
+    # Fit theory parameters default
+    nf0: int = 4  # Number of active flavors at the parametrization scale Q0
+    Q0: float = 1.65  # [GeV] Parametrization scale
+    nfref: int = 5  # Number of active flavors at Qref
+    Qref: float = 91.2  # [GeV] Reference scale for alphas and alphaqed
+    alphas: float = 0.118  # Value of alpha_s at the scale Qref
+    alphaqed: float = 0.007496252  # Values of alpha QED at the scale Qref
+    # Evolution parameters
+    IterEv: int = None  # iterations for the evolution of the PDF. Defaults to 40 when ModEv = EXA
+    ModSV: str = None  # Scale variations method in EKO (expanded or exponentiated)
+    DAMPPOWERc: int = None  # Power of the damping factor in FONLL for the c
+    DAMPPOWERb: int = None  # Power of the damping factor in FONLL for the b
+    # N3LO parameters
+    n3lo_cf_variation: list = dataclasses.field(
+        default_factory=lambda: 7 * [0]
+    )  # N3LO coefficient functions variation: -1 = lower bound, 0 = central, 1 = upper bound
+    use_fhmruvv: bool = (
+        False  # N3LO splitting functions approximation: if True use the FHMRUVV parametrization, otherwise use EKO parametrization.
+    )
+    ###### Keys for compatibility with old NNPDF theories, their values will be dropped inmediately after reading to avoid problems
+    ###### they will be set to ``None`` inmediately after loading the theory
+    MaxNfAs: int = None
+    SxRes: int = None
+    SxOrd: str = None
+    EScaleVar: int = None
+    Qedref: float = None
+    global_nx: int = None
+
+    def __post_init__(self):
+        """Drop deprecated keys"""
+        for key in DEPRECATED_KEYS:
+            object.__setattr__(self, key, None)
+
+    def asdict(self):
+        return dataclasses.asdict(self)

nnpdf_data/nnpdf_data/theorydbutils.py

@@ -5,58 +5,15 @@
 low level utilities for querying the theory database file and representing the
 data as a python object.
 """
-from dataclasses import asdict, dataclass
 from functools import lru_cache
 from pathlib import Path
 
 import pandas as pd
 
+from .theory import TheoryCard
 from .utils import parse_yaml_inp
 
 
-@dataclass(frozen=True)
-class _TheoryCard:
-    ID: int
-    PTO: int
-    FNS: str
-    DAMP: int
-    IC: int
-    ModEv: str
-    XIR: float
-    XIF: float
-    NfFF: int
-    MaxNfAs: int
-    MaxNfPdf: int
-    Q0: float
-    alphas: float
-    Qref: float
-    QED: int
-    alphaqed: float
-    Qedref: float
-    SxRes: int
-    SxOrd: str
-    HQ: str
-    mc: float
-    Qmc: float
-    kcThr: float
-    mb: float
-    Qmb: float
-    kbThr: float
-    mt: float
-    Qmt: float
-    ktThr: float
-    CKM: list[float]
-    MZ: float
-    MW: float
-    GF: float
-    SIN2TW: float
-    TMC: int
-    MP: float
-    Comments: str
-    global_nx: int
-    EScaleVar: int
-
-
 class TheoryNotFoundInDatabase(Exception):
     pass
 
@@ -67,7 +24,8 @@ def parse_theory_card(theory_card):
     Returns the theory as a dictionary
     """
     if theory_card.exists():
-        return asdict(parse_yaml_inp(theory_card, _TheoryCard))
+        tcard = parse_yaml_inp(theory_card, TheoryCard)
+        return tcard.asdict()
     raise TheoryNotFoundInDatabase(f"Theory card {theory_card} not found")
 
 

nnpdf_data/nnpdf_data/utils.py

@@ -3,6 +3,11 @@
 import ruamel.yaml as yaml
 from validobj import ValidationError, parse_input
 
+try:
+    Loader = yaml.CLoader
+except AttributeError:
+    Loader = yaml.Loader
+
 
 def parse_yaml_inp(input_yaml, spec):
     """
@@ -12,7 +17,7 @@ def parse_yaml_inp(input_yaml, spec):
     https://validobj.readthedocs.io/en/latest/examples.html#yaml-line-numbers
     """
     input_yaml = pathlib.Path(input_yaml)
-    inp = yaml.round_trip_load(input_yaml.open("r", encoding="utf-8"))
+    inp = yaml.load(input_yaml.read_text(encoding="utf-8"), Loader=Loader)
     try:
         return parse_input(inp, spec)
     except ValidationError as e: