Reformated the screen output
This commit is contained in:
@ -200,7 +200,7 @@ def get_lammps_build_configuration(lmp_binary):
|
||||
def execute(lmp_binary, config, input_file_name, generate_ref_yaml=False):
|
||||
cmd_str = config['mpiexec'] + " " + config['mpiexec_numproc_flag'] + " " + config['nprocs'] + " "
|
||||
cmd_str += lmp_binary + " -in " + input_file_name + " " + config['args']
|
||||
print(f"Executing: {cmd_str}")
|
||||
logger.info(f"Executing: {cmd_str}")
|
||||
p = subprocess.run(cmd_str, shell=True, text=True, capture_output=True)
|
||||
|
||||
return cmd_str, p.stdout, p.stderr, p.returncode
|
||||
@ -353,7 +353,7 @@ def iterate(lmp_binary, input_folder, input_list, config, results, progress_file
|
||||
|
||||
continue
|
||||
|
||||
str_t = "\nRunning " + input + f" ({test_id+1}/{num_tests})"
|
||||
str_t = " + " + input + f" ({test_id+1}/{num_tests})"
|
||||
|
||||
result = TestResult(name=input, output="", time="", status="passed")
|
||||
|
||||
@ -375,12 +375,12 @@ def iterate(lmp_binary, input_folder, input_list, config, results, progress_file
|
||||
generate_markers(input, input_markers)
|
||||
process_markers(input_markers, input_test)
|
||||
|
||||
str_t = "\nRunning " + input_test + f" ({test_id+1}/{num_tests})"
|
||||
str_t = "\n + " + input_test + f" ({test_id+1}/{num_tests})"
|
||||
else:
|
||||
input_test = input
|
||||
|
||||
print(str_t)
|
||||
print(f"-"*len(str_t))
|
||||
#print(f"-"*len(str_t))
|
||||
logger.info(str_t)
|
||||
logger.info(f"-"*len(str_t))
|
||||
|
||||
@ -467,7 +467,7 @@ def iterate(lmp_binary, input_folder, input_list, config, results, progress_file
|
||||
test_id = test_id + 1
|
||||
continue
|
||||
|
||||
print(f"Comparing thermo output from log.lammps against the reference log file {thermo_ref_file}")
|
||||
logger.info(f"Comparing thermo output from log.lammps against the reference log file {thermo_ref_file}")
|
||||
if num_runs != num_runs_ref:
|
||||
logger.info(f"ERROR: Number of runs in log.lammps ({num_runs}) is not the same as that in the reference log ({num_runs_ref})")
|
||||
result.status = "error"
|
||||
@ -554,19 +554,25 @@ def iterate(lmp_binary, input_folder, input_list, config, results, progress_file
|
||||
"{abs_diff_check.rjust(20)} {rel_diff_check.rjust(20)}")
|
||||
|
||||
if num_abs_failed > 0:
|
||||
print(f"{num_abs_failed} absolute diff checks failed with the specified tolerances.")
|
||||
msg = f" {num_abs_failed} abs diff checks failed."
|
||||
print(msg)
|
||||
logger.info(msg)
|
||||
result.status = "failed"
|
||||
if verbose == True:
|
||||
for i in failed_abs_output:
|
||||
print(f"- {i}")
|
||||
if num_rel_failed > 0:
|
||||
print(f"{num_rel_failed} relative diff checks failed with the specified tolerances.")
|
||||
msg = f" {num_rel_failed} rel diff checks failed."
|
||||
print(msg)
|
||||
logger.info(msg)
|
||||
result.status = "failed"
|
||||
if verbose == True:
|
||||
for i in failed_rel_output:
|
||||
print(f"- {i}")
|
||||
if num_abs_failed == 0 and num_rel_failed == 0:
|
||||
print(f"All {num_checks} checks passed.")
|
||||
msg = f" all {num_checks} checks passed."
|
||||
print(msg)
|
||||
logger.info(msg)
|
||||
result.status = "passed"
|
||||
num_passed = num_passed + 1
|
||||
|
||||
@ -575,9 +581,9 @@ def iterate(lmp_binary, input_folder, input_list, config, results, progress_file
|
||||
progress.write(f"{input}: {{ folder: {input_folder}, status: completed }}\n")
|
||||
progress.close()
|
||||
|
||||
str_t = f"Completed " + input_test
|
||||
print(str_t)
|
||||
print("-"*(5*width+4))
|
||||
#str_t = f"Completed " + input_test
|
||||
#print(str_t)
|
||||
#print("-"*(5*width+4))
|
||||
test_id = test_id + 1
|
||||
|
||||
# remove the annotated input script
|
||||
@ -678,10 +684,10 @@ if __name__ == "__main__":
|
||||
|
||||
if len(example_subfolders) > 0:
|
||||
print("\nExample folders to test:")
|
||||
print(example_subfolders)
|
||||
print(*example_subfolders, sep='\n')
|
||||
if example_toplevel != "":
|
||||
print("\nTop-level example folder:")
|
||||
print(example_toplevel)
|
||||
print(f" {example_toplevel}")
|
||||
|
||||
# Using in place input scripts
|
||||
inplace_input = True
|
||||
@ -764,7 +770,6 @@ if __name__ == "__main__":
|
||||
except Exception:
|
||||
print(f"Cannot open progress file {progress_file} to resume, rerun all the tests")
|
||||
|
||||
print("\nStart..")
|
||||
# default setting is to use inplace_input
|
||||
if inplace_input == True:
|
||||
|
||||
@ -785,7 +790,8 @@ if __name__ == "__main__":
|
||||
for directory in example_subfolders:
|
||||
|
||||
# change to the directory where the input script and data files are located
|
||||
print("\nEntering " + directory)
|
||||
print("-"*80)
|
||||
print("Entering " + directory)
|
||||
os.chdir(directory)
|
||||
|
||||
cmd_str = "ls in.*"
|
||||
@ -793,7 +799,7 @@ if __name__ == "__main__":
|
||||
input_list = p.stdout.split('\n')
|
||||
input_list.remove('')
|
||||
|
||||
print(f"List of {len(input_list)} input scripts: {input_list}")
|
||||
print(f"{len(input_list)} input script(s): {input_list}")
|
||||
total_tests += len(input_list)
|
||||
|
||||
# iterate through the input scripts
|
||||
|
||||
Reference in New Issue
Block a user