convertNoisyMap.py

Go to the documentation of this file.

#!/bin/python
# SPDX-License-Identifier: LGPL-3.0-or-later
# SPDX-FileCopyrightText: Copyright CERN for the benefit of the SHiP Collaboration
 
 
# Python script to convert a B field map from an ascii text file into
# a ROOT file for FairShip. The field map needs to use a regular
# x,y,z binning structure where the coordinates are assumed to
# be in ascending z, y and x, in that order. Each data line needs to
# contain the values "x y z Bx By Bz", with the field components in Tesla.
# Comment lines in the datafile starting with the hash symbol # are ignored.
# Distances for FairShip need to be in cm, so we use a scaling factor
# cmScale (default = 1.0) to convert the text file distances into cm.
# For example, if the input data uses mm for lengths, cmScale = 0.1.
 
import os
 
import pandas as pd
import ROOT
 
# Struct for the ROOT file TTree data: coord range and field binning
 
ROOT.gROOT.ProcessLine(
    "struct rangeStruct{\
       float xMin;\
       float xMax;\
       float dx;\
       float yMin;\
       float yMax;\
       float dy;\
       float zMin;\
       float zMax;\
       float dz;\
    };"
)
# The field map is assumed to obey the following coordinate bin ordering:
# z is increased first, y is increased 2nd, x is increased last.
# For the coordinate bin (iX, iY, iZ), the field bin = (iX*Ny + iY)*Nz + iZ,
# where Ny and Nz are the number of y and z bins
 
ROOT.gROOT.ProcessLine(
    "struct dataStruct{\
       float x;\
       float y;\
       float z;\
       float Bx;\
       float By;\
       float Bz;\
    };"
)
 
 
def run(
    inFileName="FieldTest.txt", rootFileName="BFieldTest.root", cmScale: float | int = 1.0, storeCoords: bool = False
) -> None:
    createRootMap(inFileName, rootFileName, cmScale, storeCoords)
 
 
def createRootMap(inFileName, rootFileName, cmScale, storeCoords) -> None:
    print(f"Create map {rootFileName} from {inFileName} using cmScale = {cmScale}")
    if storeCoords is True:
        print(f"We will also store the x,y,z field coordinates in {rootFileName}")
 
    rangeInfo = findRanges(inFileName, cmScale)
 
    # Define ROOT file and its TTree
    theFile = ROOT.TFile.Open(rootFileName, "recreate")
 
    rangeTree = ROOT.TTree("Range", "Range")
    rangeTree.SetDirectory(theFile)
 
    # Coordinate ranges
    rStruct = ROOT.rangeStruct()
    rangeTree.Branch("xMin", ROOT.addressof(rStruct, "xMin"), "xMin/F")
    rangeTree.Branch("xMax", ROOT.addressof(rStruct, "xMax"), "xMax/F")
    rangeTree.Branch("dx", ROOT.addressof(rStruct, "dx"), "dx/F")
    rangeTree.Branch("yMin", ROOT.addressof(rStruct, "yMin"), "yMin/F")
    rangeTree.Branch("yMax", ROOT.addressof(rStruct, "yMax"), "yMax/F")
    rangeTree.Branch("dy", ROOT.addressof(rStruct, "dy"), "dy/F")
    rangeTree.Branch("zMin", ROOT.addressof(rStruct, "zMin"), "zMin/F")
    rangeTree.Branch("zMax", ROOT.addressof(rStruct, "zMax"), "zMax/F")
    rangeTree.Branch("dz", ROOT.addressof(rStruct, "dz"), "dz/F")
 
    rStruct.xMin = rangeInfo["xMin"]
    rStruct.xMax = rangeInfo["xMax"]
    rStruct.dx = rangeInfo["dx"]
    rStruct.yMin = rangeInfo["yMin"]
    rStruct.yMax = rangeInfo["yMax"]
    rStruct.dy = rangeInfo["dy"]
    rStruct.zMin = rangeInfo["zMin"]
    rStruct.zMax = rangeInfo["zMax"]
    rStruct.dz = rangeInfo["dz"]
 
    # Centre the field map on the local origin (cm)
    x0 = 0  # .5 * (rStruct.xMin + rStruct.xMax)
    y0 = 0  # .5 * (rStruct.yMin + rStruct.yMax)
    z0 = 0.5 * (rStruct.zMin + rStruct.zMax)
 
    # Use this if we don't want to centre the field map
    # x0 = 0.0
    # y0 = 0.0
    # z0 = 0.0
 
    print(f"Centering field map using coordinate shift {x0} {y0} {z0} cm")
 
    # Center coordinate range limits (cm)
    rStruct.xMin = rStruct.xMin - x0
    rStruct.xMax = rStruct.xMax - x0
 
    rStruct.yMin = rStruct.yMin - y0
    rStruct.yMax = rStruct.yMax - y0
 
    rStruct.zMin = rStruct.zMin - z0
    rStruct.zMax = rStruct.zMax - z0
 
    print(f"x range = {rStruct.xMin} to {rStruct.xMax}")
    print(f"y range = {rStruct.yMin} to {rStruct.yMax}")
    print(f"z range = {rStruct.zMin} to {rStruct.zMax}")
 
    # Fill info into range tree
    rangeTree.Fill()
 
    # Store field data components
    dataTree = ROOT.TTree("Data", "Data")
    dataTree.SetDirectory(theFile)
 
    # Field components with (x,y,z) coordinate binning ordered such that
    # z, then y, then x is increased. For the coordinate bin (iX, iY, iZ),
    # the field bin = (iX*Ny + iY)*Nz + iZ, where Ny and Nz are the number
    # of y and z bins
    dStruct = ROOT.dataStruct()
    if storeCoords is True:
        dataTree.Branch("x", ROOT.addressof(dStruct, "x"), "x/F")
        dataTree.Branch("y", ROOT.addressof(dStruct, "y"), "y/F")
        dataTree.Branch("z", ROOT.addressof(dStruct, "z"), "z/F")
 
    dataTree.Branch("Bx", ROOT.addressof(dStruct, "Bx"), "Bx/F")
    dataTree.Branch("By", ROOT.addressof(dStruct, "By"), "By/F")
    dataTree.Branch("Bz", ROOT.addressof(dStruct, "Bz"), "Bz/F")
 
    # Reopen the file and store the information in the ROOT file
 
    inData = pd.read_csv(inFileName, delim_whitespace=True, header=None)
    inData.columns = ["x", "y", "z", "bx", "by", "bz"]
    inData = inData.sort_values(by=["x", "y", "z"])
    inData = inData.astype(float)
 
    count = 0.0
    data_shape = float(inData.shape[0])
    for row in inData.itertuples():
        if row.Index / data_shape >= count:
            print(f"Processed: {count * 100} %")
            count += 0.1
 
        # Bin centre coordinates with origin shift (all in cm)
        if storeCoords is True:
            dStruct.x = row.x * cmScale - x0
            dStruct.y = row.y * cmScale - y0
            dStruct.z = row.z * cmScale - z0
 
        # B field components (Tesla)
        dStruct.Bx = row.bx
        dStruct.By = row.by
        dStruct.Bz = row.bz
 
        dataTree.Fill()
 
    theFile.cd()
    rangeTree.Write()
    dataTree.Write()
    theFile.Close()
 
 
def findRanges(inFileName, cmScale):
    # First read the data file to find the binning and coordinate ranges.
    # Store the unique (ordered) x, y and z values so we can then find the
    # bin widths, min/max limits and central offset
 
    xArray = []
    yArray = []
    zArray = []
    x_set = set()
    y_set = set()
    z_set = set()
 
    with open(inFileName) as f:
        # Read each line
        for line in f:
            # Ignore comment lines which begin with "#"
            if "#" not in line:
                sLine = line.split()
 
                x = float(sLine[0]) * cmScale
                y = float(sLine[1]) * cmScale
                z = float(sLine[2]) * cmScale
 
                if x not in x_set:
                    xArray.append(x)
                    x_set.add(x)
 
                if y not in y_set:
                    yArray.append(y)
                    y_set.add(y)
 
                if z not in z_set:
                    zArray.append(z)
                    z_set.add(z)
 
    Nx = len(xArray)
    Ny = len(yArray)
    Nz = len(zArray)
 
    xMin = 0.0
    xMax = 0.0
    dx = 0.0
 
    if Nx > 0:
        xMin = xArray[0]
        Nx1 = Nx - 1
        xMax = xArray[Nx1]
        dx = (xMax - xMin) / (Nx1 * 1.0)
 
    if Ny > 0:
        yMin = yArray[0]
        Ny1 = Ny - 1
        yMax = yArray[Ny1]
        dy = (yMax - yMin) / (Ny1 * 1.0)
 
    if Nz > 0:
        zMin = zArray[0]
        Nz1 = Nz - 1
        zMax = zArray[Nz1]
        dz = (zMax - zMin) / (Nz1 * 1.0)
 
    rangeInfo = {
        "Nx": Nx,
        "xMin": xMin,
        "xMax": xMax,
        "dx": dx,
        "Ny": Ny,
        "yMin": yMin,
        "yMax": yMax,
        "dy": dy,
        "Nz": Nz,
        "zMin": zMin,
        "zMax": zMax,
        "dz": dz,
    }
 
    print(f"rangeInfo = {rangeInfo}")
 
    return rangeInfo
 
 
if __name__ == "__main__":
    run(
        os.path.expandvars("$FAIRSHIP/files/noisy_fieldMap.csv"),
        os.path.expandvars("$FAIRSHIP/files/MuonShieldField.root"),
        1,
        False,
    )