import pathlib
import numpy as np
from validphys.utils import yaml_safe
from .q2grids import Q2GRID_NNPDF40
LAMBDA2 = 0.0625
[docs]def read_runcard(usr_path):
"""Read the runcard and return the relevant information for evolven3fit"""
return yaml_safe.load((usr_path / "filter.yml").read_text(encoding="UTF-8"))
[docs]def get_theoryID_from_runcard(usr_path):
"""Return the theoryID from the runcard"""
# read the runcard
my_runcard = read_runcard(usr_path)
return my_runcard["theory"]["theoryid"]
[docs]def Q2_to_t(q2: float, lambda2) -> float:
"""Map to define loglog spacing of the Q2grid"""
return np.log(np.log(q2 / lambda2))
[docs]def t_to_Q2(t: float, lambda2) -> float:
"""Map to go from the loglog spacing back to Q2"""
return lambda2 * np.exp(np.exp(t))
[docs]def generate_q2grid(Q0, Qmin, Qmax, match_dict, total_points, total_points_ic, legacy40=False):
"""Generate the q2grid used in the final evolved pdfs or use the default grid if legacy40 is set.
The grid uses $\log(\log(Q^2/\text{`LAMBDA2`})) spacing between the points.
The grid from `Q2_min` --> `mc` is made separately from the rest to be able to let it contain at
least 5 points. The reason for this is to always have some points in the intrinsic charm regime.
The rest of the grids contains the same loglog spacing from threshold to threshold.
Since the grid needs to contain the $Q^2$ values `Q2_min`, `mc`^2, `Q0`^2, `mb`^2 and `Q2_max`,
we divide the grid in batches that have these values as boundaries, and add them together
at the end. The batches ad boundaries are thus like this:
Q2_min --> mc^2 --> Q0^2 --> mb^2 --> Q2_max
`match_dict` contains the quark mass thresholds and factors. This "factor" is the number to be
multiplied to mass in order to obtain the relative matching scale (e.g. `match_dict["kbThr"]`
in the case of the bottom threshold).
"""
# If flag --legacy40 is set return handmade legacy grid
if legacy40:
return Q2GRID_NNPDF40
# Otherwise dynamically create the grid from Q2_min --> Q2_max
if Q0 < Qmin:
raise ValueError("Q0 cannot be smaller than Qmin because the grid needs to contain Q0")
if Qmax < Qmin:
raise ValueError("Qmax cannot be smaller than Qmin")
if total_points <= 5:
raise ValueError("You need minimally 6 points (total_points > 5)")
if total_points_ic < 2 and Qmin < 1.502:
raise ValueError("You need minimally 2 points below Q0 (total_points_ic > 1)")
Q2_min = Qmin**2 # 1.0**2
Q2_max = Qmax**2 # 1e5**2
# Collect all node Q2's from Q0^2 --> Q2_max
q0_2 = Q0**2
node_Q2 = [q0_2, (match_dict["mb"] * match_dict["kbThr"]) ** 2, Q2_max]
# Make initial uniform grid in t from Q0^2 --> Q2_max
t_min = Q2_to_t(q0_2, LAMBDA2)
t_max = Q2_to_t(Q2_max, LAMBDA2)
t_vals = np.linspace(t_min, t_max, total_points)
q2_vals = t_to_Q2(t_vals, LAMBDA2)
# Count how many points fall into each subgrid
n_intervals = len(node_Q2) - 1
nQpoints = np.zeros(n_intervals, dtype=int)
subgridindex = 0
for q2 in q2_vals:
while subgridindex < n_intervals - 1 and q2 >= node_Q2[subgridindex + 1]:
subgridindex += 1
nQpoints[subgridindex] += 1
# Make t grid from Q2_min --> Q0^2
t_min_ic = Q2_to_t(Q2_min, LAMBDA2)
t_max_ic = Q2_to_t((match_dict["mc"] * match_dict["kcThr"]) ** 2, LAMBDA2)
t_vals_ic = np.linspace(t_min_ic, t_max_ic, total_points_ic)
q2_vals_ic = t_to_Q2(t_vals_ic, LAMBDA2)
# Now build each subgrid to contain the points we want
grids = []
grids.append(q2_vals_ic)
for i in range(len(node_Q2) - 1):
q2_lo, q2_hi = node_Q2[i], node_Q2[i + 1]
t_lo, t_hi = Q2_to_t(q2_lo, LAMBDA2), Q2_to_t(q2_hi, LAMBDA2)
npts = int(nQpoints[i])
t_subgr = np.linspace(t_lo, t_hi, npts)
q2_subgr = t_to_Q2(t_subgr, LAMBDA2)
if i < n_intervals - 1:
q2_subgr = q2_subgr[:-1]
grids.append(q2_subgr)
# Combine all subgrids and return as an array
q2_full = np.concatenate(grids)
return q2_full
[docs]def check_is_a_fit(config_folder):
"""Check if config_folder is a fit folder, i.e. if it contains the filter.yml file
and the nnfit folder."""
usr_path = pathlib.Path(config_folder)
filter_path = usr_path / "filter.yml"
if not filter_path.is_file():
raise ValueError(
"filter.yaml file not found: the path" + str(filter_path.absolute()) + " is not valid"
)
nnfitpath = usr_path / "nnfit"
if not nnfitpath.is_dir():
raise ValueError("nnfit folder not found: provided path is not valid")
[docs]def check_filter(config_folder):
"""Check if config_folder contains a filter.yml file."""
usr_path = pathlib.Path(config_folder)
filter_path = usr_path / "filter.yml"
if not filter_path.is_file():
raise ValueError(
f"filter.yaml file not found: the path {filter_path.absolute()} is not valid"
)
[docs]def check_nnfit_folder(config_folder):
"""Check if config_folder contains a nnfit folder."""
usr_path = pathlib.Path(config_folder)
nnfitpath = usr_path / "nnfit"
if not nnfitpath.is_dir():
raise ValueError("nnfit folder not found: provided path is not valid")