Document tools; add tools for converting native sampler outputs

tools/convert_native_sample_to_time_series.py

@@ -0,0 +1,158 @@
+import numpy as np
+import matplotlib.pyplot as plt
+import sys
+import glob
+import os
+import argparse
+
+def main():
+    """Convert native data for a single sampler point to a time series in ASCII."""
+    parser = argparse.ArgumentParser(
+        description="Convert native data for a single sampler point to a time series in ASCII"
+    )
+    parser.add_argument(
+        "-aw",
+        "--amr_wind_path",
+        help="path to amr-wind directory",
+        required=True,
+        type=str,
+    )
+    parser.add_argument(
+        "-d",
+        "--run_dir",
+        help="Directory where run took place, where post-processing directory is",
+        required=True,
+        type=str,
+    )
+    parser.add_argument(
+        "-p",
+        "--post_proc_dir",
+        help="Name of post-processing directory",
+        default="post_processing",
+        type=str,
+    )
+    parser.add_argument(
+        "-l",
+        "--label",
+        help="Top-level sampling label name",
+        default="sampling",
+        type=str,
+    )
+    parser.add_argument(
+        "-s",
+        "--sampler",
+        help="Bottom-level sampling label name",
+        default="* use first *",
+        type=str,
+    )
+    parser.add_argument(
+        "-i",
+        "--index",
+        help="Particle index within specified sampler",
+        default=0,
+        type=int,
+    )
+    args = parser.parse_args()
+
+    ########## Input arguments from command line ##############
+    print("\nParameters to be used:")
+    print("   run directory path: ", os.path.realpath(args.run_dir))
+    print("   AMR-Wind path: ", os.path.realpath(args.amr_wind_path))
+    print("   post-processing directory name: ", args.post_proc_dir)
+    print("   top-level sampling label name: ", args.label)
+    ###########################################################
+
+    sys.path.append(args.amr_wind_path+'/tools/')
+    import amrex_particle
+    from amrex_particle import AmrexParticleFile
+
+    post_proc_dir = args.run_dir + "/" + args.post_proc_dir
+    pfile = AmrexParticleFile(args.run_dir)
+
+    os.chdir(post_proc_dir)
+    contents = glob.glob(args.label + "*")
+    directories = [path for path in contents if os.path.isdir(path)]
+    sdirs = sorted(directories)
+    n_dir = len(sdirs)
+    if (n_dir == 0):
+        print("ERROR: No matching sampling directories found, exiting!")
+        sys.exit(1)
+
+    # Separate by length, then order again (because 10000 would be next to 100000)
+    max_len = len(sdirs[0])
+    min_len = max_len
+    for n in range(len(sdirs)):
+        max_len = max(max_len,len(sdirs[n]))
+        min_len = min(min_len,len(sdirs[n]))
+
+    sdirs_init = list(sdirs)
+    if (max_len != min_len):
+        cur_len = min_len
+        idx = 0
+        for inc_len in range(max_len - min_len + 1):
+            cur_len += inc_len
+            for n in range(len(sdirs_init)):
+                if (len(sdirs_init[n]) == cur_len):
+                    sdirs[idx] = sdirs_init[n]
+                    idx += 1
+
+    max_step = int(sdirs[n_dir-1][len(args.label):len(sdirs[n_dir-1])])
+
+    # Get stuff for header and array sizes using first available directory
+    n0 = 0
+    step = int(sdirs[n0][len(args.label):len(sdirs[n0])])
+    pt = pfile.load(step, label=args.label, root_dir = ".")
+    pt.parse_header()
+    pt.parse_info()
+    output = np.zeros((n_dir, 1 + 3 + pt.num_reals + pt.num_ints - 3))
+    v_str = "x y z"
+    for rv in range(pt.num_reals):
+        v_str += " " + pt.real_var_names[rv]
+    for iv in range(pt.num_ints - 3):
+        v_str += " " + pt.int_var_names[iv + 3]
+    header = "t " + v_str
+    samplers = pt.info["samplers"]
+    s_index = 0
+    if (args.sampler[0] != "*"):
+        for s in range(len(samplers)):
+            if (args.sampler == samplers[s]["label"]):
+                s_index = s
+    else:
+        args.sampler = samplers[0]["label"]
+
+    ########## Input arguments from command line ##############
+    print("   bottom-level sampling label name: ", args.sampler)
+    print("   particle index within sampler: ", args.index)
+    ###########################################################
+
+    fname = args.label + "_" + args.sampler + "_" + str(args.index) + "_time_series.txt"
+    print("Output file name: ", fname)
+    print("Output file name with path: ", os.path.realpath(fname))
+    print()
+
+    # Loop through files to get time series data
+    for n in range(n_dir):
+        step = int(sdirs[n][len(args.label):len(sdirs[n])])
+        print(str(step) + " out of " + str(max_step))
+        pt = pfile.load(step, label=args.label, root_dir = ".")
+        pt.parse_header()
+        pt.parse_info()
+        pt.load_binary_data()
+        data = pt.df
+
+        output[n, 0] = pt.info["time"]
+
+        for p in range(pt.num_particles):
+            if (data.set_id[p] == s_index and data.probe_id[p] == args.index):
+                output[n, 1] = data.xco[p]
+                output[n, 2] = data.yco[p]
+                output[n, 3] = data.zco[p]
+                for rv in range(pt.num_reals):
+                    output[n, 4+rv] = data[pt.real_var_names[rv]][p]
+                for iv in range(pt.num_ints - 3):
+                    output[n, 4+pt.num_reals+iv] = data[pt.int_var_names[iv+3]][p]
+
+    np.savetxt(fname, output[0:len(sdirs)],header=header)
+
+if __name__ == "__main__":
+    main()  

tools/convert_native_sampling_to_structured_ascii.py

@@ -0,0 +1,112 @@
+import numpy as np
+import matplotlib.pyplot as plt
+import sys
+import glob
+import os
+import argparse
+
+def main():
+    """Convert native data for a single sampler point to a time series in ASCII."""
+    parser = argparse.ArgumentParser(
+        description="Convert native data for a single sampler point to a time series in ASCII"
+    )
+    parser.add_argument(
+        "-aw",
+        "--amr_wind_path",
+        help="path to amr-wind directory",
+        required=True,
+        type=str,
+    )
+    parser.add_argument(
+        "-d",
+        "--run_dir",
+        help="Directory where run took place, where post-processing directory is",
+        required=True,
+        type=str,
+    )
+    parser.add_argument(
+        "-p",
+        "--post_proc_dir",
+        help="Name of post-processing directory",
+        default="post_processing",
+        type=str,
+    )
+    parser.add_argument(
+        "-l",
+        "--label",
+        help="Top-level sampling label name",
+        default="sampling",
+        type=str,
+    )
+    args = parser.parse_args()
+
+    ########## Input arguments from command line ##############
+    print("\nParameters to be used:")
+    print("   run directory path: ", os.path.realpath(args.run_dir))
+    print("   AMR-Wind path: ", os.path.realpath(args.amr_wind_path))
+    print("   post-processing directory name: ", args.post_proc_dir)
+    print("   top-level sampling label name: ", args.label)
+    ###########################################################
+
+    sys.path.append(args.amr_wind_path+'/tools/')
+    import amrex_particle
+    from amrex_particle import AmrexParticleFile
+
+    post_proc_dir = args.run_dir + "/" + args.post_proc_dir
+    pfile = AmrexParticleFile(args.run_dir)
+
+    os.chdir(post_proc_dir)
+    contents = glob.glob(args.label + "*")
+    directories = [path for path in contents if os.path.isdir(path)]
+    sdirs = sorted(directories)
+
+    max_dir = len(sdirs)-1
+    print(max_dir)
+    if (max_dir < 0):
+        print("ERROR: No matching sampling directories found, exiting!")
+        sys.exit(1)
+    max_step = int(sdirs[max_dir][len(args.label):len(sdirs[max_dir])])
+
+    for n in range(len(sdirs)):
+        step = int(sdirs[n][len(args.label):len(sdirs[n])])
+        print(str(step) + " out of " + str(max_step))
+        pt = pfile.load(step, label=args.label, root_dir = ".")
+        pt.parse_header()
+        pt.parse_info()
+        pt.load_binary_data()
+        data = pt.df
+        samplers = pt.info["samplers"]
+
+        t_str = str(pt.info["time"])
+        v_str = "x y z"
+
+        for rv in range(pt.num_reals):
+            v_str += " " + pt.real_var_names[rv]
+        for iv in range(pt.num_ints - 3):
+            v_str += " " + pt.int_var_names[iv + 3]
+
+        header = "t = " + t_str + "\n" + v_str
+
+        # Have to go with max possible size, subsampler size not stored
+        output = np.zeros((pt.num_particles, 3 + pt.num_reals + pt.num_ints - 3))
+
+        for s in range(len(samplers)):
+            np_s = 0
+            for p in range(pt.num_particles):
+                if data.set_id[p] == s:
+                    np_s += 1
+                    pid = data.probe_id[p]
+                    output[pid, 0] = data.xco[p]
+                    output[pid, 1] = data.yco[p]
+                    output[pid, 2] = data.zco[p]
+                    for rv in range(pt.num_reals):
+                        output[pid, 3+rv] = data[pt.real_var_names[rv]][p]
+                    for iv in range(pt.num_ints - 3):
+                        output[pid, 3+pt.num_reals+iv] = data[pt.int_var_names[iv+3]][p]
+
+
+            fname = sdirs[n] + "_" + samplers[s]["label"]+ ".txt"
+            np.savetxt(fname,output[0:np_s,:],header=header)
+
+if __name__ == "__main__":
+    main()