Merge pull request #39 from cta-observatory/add_longitudinal

Add reader and fit function for longitudinal files

Author: maxnoe

corsikaio/__init__.py

@@ -1,11 +1,14 @@
 from .file import CorsikaFile, CorsikaCherenkovFile, CorsikaParticleFile
+from .longitudinal import read_longitudinal_distributions, longitudinal_fit_function
 from .version import __version__
 
 
 __all__ = [
     'CorsikaFile',
     'CorsikaCherenkovFile',
     'CorsikaParticleFile',
+    'read_longitudinal_distributions',
+    'longitudinal_fit_function',
     'as_dict',
     '__version__',
 ]

corsikaio/longitudinal.py

@@ -0,0 +1,114 @@
+"""
+Functions related to the CORSIKA longitudinal distribution
+"""
+import re
+import numpy as np
+
+PARTICLE_HEADER_RE = re.compile(r"LONGITUDINAL DISTRIBUTION IN\s+(\d+)\s+(SLANT|VERTICAL)\s+STEPS OF\s+(\d+(?:.\d*)?) G\/CM\*\*2 FOR SHOWER\s+(\d+)")
+ENERGY_HEADER_RE = re.compile(r"LONGITUDINAL ENERGY DEPOSIT IN\s+(\d+)\s+(SLANT|VERTICAL)\s+STEPS OF\s+(\d+(?:.\d*)?) G\/CM\*\*2 FOR SHOWER\s+(\d+)")
+
+ENERGY_COLUMNS = [
+    "depth",
+    "gamma", "em_ioniz", "em_cut", "mu_ioniz",
+    "mu_cut", "hadr_ioniz", "hadr_cut", "neutrino", "sum"
+]
+PARTICLE_COLUMNS = [
+    "depth",
+    "gammas", "positrons", "electrons", "mu_plus", "mu_minus",
+    "hadrons", "charged", "nuclei", "cherenkov"
+]
+
+
+def longitudinal_fit_function(depth, n_max, depth_0, depth_max, a, b, c):
+    """
+    The function CORSIKA fits to the longitudinal distribution.
+
+    Parameters of this function are stored in the event end block
+    in case the corresponding options of the LONGI keyword are set.
+
+    Use like this:
+
+    >>> depth = np.linspace(0, 1000, 100)
+    >>> longitudinal_fit_function(depth, *event.end["longitudinal_fit_parameters"])
+    """
+    denominator =  (a + b * depth + c * depth**2)
+    exponent = (depth_max - depth_0) / denominator
+    power = ((depth - depth_0) / (depth_max - depth_0))**exponent
+    exp = np.exp((depth_max - depth) / denominator)
+    return n_max * power *  exp
+
+
+def read_longitudinal_distributions(path):
+    """
+    Read longitudinal profiles from CORSIKA longitudinal file.
+
+    This function returns a generator that iterates over air showers.
+
+    Parameters
+    ----------
+    path : str, Path
+        Path to CORSIKA longitudinal output file (DATXXXXXX.long)
+
+    Yields
+    ------
+    longitudinal : dict
+        Dict with the information for one air shower.
+        Contains the longitudinal tables for "particles" and "energy_deposition"
+        and the "parameters", "chi2_ndf" and "average_deviation" values of the 
+        fit to the distribution if available.
+    """
+    first = True
+    with open(path, "r") as f:
+        try:
+            line = f.readline().strip()
+        except UnicodeDecodeError:
+            raise IOError(f"Inputfile {path} does not seem to be a longitudinal file")
+
+        while line:
+            match = PARTICLE_HEADER_RE.match(line)
+            if not match:
+                if first:
+                    raise IOError(f"Inputfile {path} does not seem to be a longitudinal file")
+                else:
+                    raise IOError(f"Error reading file, expected header line, got: {line}")
+            first = False
+
+            n_steps = int(match.group(1))
+            longi = dict(
+                shower=int(match.group(4)),
+                n_steps=n_steps,
+                slant=match.group(2) == "SLANT",
+                step_width=float(match.group(3)),
+            )
+
+            # skip line with names
+            f.readline()
+            longi["particles"] = np.genfromtxt(
+                f, max_rows=n_steps, names=PARTICLE_COLUMNS, dtype=None
+            )
+
+            line = f.readline().strip()
+            match = ENERGY_HEADER_RE.match(line)
+            if not match:
+                raise IOError(f"Error reading file, expected energy deposition header line, got: {line}")
+            n_steps = int(match.group(1))
+
+            # skip
+            f.readline()
+            longi["energy_deposition"] = np.genfromtxt(f, max_rows=n_steps, names=ENERGY_COLUMNS, dtype=None)
+
+            line = f.readline()
+            if line.strip().startswith("FIT"):
+                f.readline()
+                parameters = f.readline().partition(" = ")[2]
+                longi["parameters"] = np.array([float(p) for p in parameters.split()])
+
+                chi2_ndf = f.readline().partition(" = ")[2]
+                longi["chi2_ndf"] = float(chi2_ndf)
+
+                deviation = f.readline().partition(" = ")[2]
+                longi["average_deviation"] = float(deviation)
+
+            f.readline()
+            yield longi
+            line = f.readline().strip()