def generate_data(model, inputdir, outputdir, cluster_type="pp", pp_cpus=2, runs=1):
"""
The first pipeline step: data generation.
:param model: the model to process
:param inputdir: the directory containing the model
:param outputdir: the directory containing the output files
:param cluster_type: pp for local Parallel Python, lsf for Load Sharing Facility, sge for Sun Grid Engine.
:param pp_cpus: the number of CPU used by Parallel Python.
:param runs: the number of model simulation
:return: no output
"""
if runs < 1:
logger.error("variable " + str(runs) + " must be greater than 0. Please, check your configuration file.")
return
if not os.path.isfile(os.path.join(inputdir, model)):
logger.error(os.path.join(inputdir, model) + " does not exist.")
return
copasi = get_copasi()
if copasi is None:
logger.error(
"CopasiSE not found! Please check that CopasiSE is installed and in the PATH environmental variable.")
return
# folder preparation
refresh_directory(outputdir, model[:-4])
# execute runs simulations.
logger.info("Simulating model " + model + " for " + str(runs) + " time(s)")
# Replicate the copasi file and rename its report file
groupid = "_" + get_rand_alphanum_str(20) + "_"
group_model = model[:-4] + groupid
for i in xrange(1, runs + 1):
shutil.copyfile(os.path.join(inputdir, model), os.path.join(inputdir, group_model) + str(i) + ".cps")
replace_string_in_file(os.path.join(inputdir, group_model) + str(i) + ".cps",
model[:-4] + ".csv",
group_model + str(i) + ".csv")
# run copasi in parallel
# To make things simple, the last 10 character of groupid are extracted and reversed.
# This string will be likely different from groupid and is the string to replace with
# the iteration number.
str_to_replace = groupid[10::-1]
command = copasi + " " + os.path.join(inputdir, group_model + str_to_replace + ".cps")
parallel_computation(command, str_to_replace, cluster_type, runs, outputdir, pp_cpus)
# move the report files
report_files = [f for f in os.listdir(inputdir) if
re.match(group_model + '[0-9]+.*\.csv', f) or re.match(group_model + '[0-9]+.*\.txt', f)]
for file in report_files:
# Replace some string in the report file
replace_str_copasi_sim_report(os.path.join(inputdir, file))
# rename and move the output file
shutil.move(os.path.join(inputdir, file), os.path.join(outputdir, file.replace(groupid, "_")[:-4] + ".csv"))
# removed repeated copasi files
repeated_copasi_files = [f for f in os.listdir(inputdir) if re.match(group_model + '[0-9]+.*\.cps', f)]
for file in repeated_copasi_files:
os.remove(os.path.join(inputdir, file))
def generate_data(model, inputdir, cluster_type, pp_cpus, nfits, outputdir, sim_data_dir, updated_models_dir):
"""
The first pipeline step: data generation.
:param model: the model to process
:param inputdir: the directory containing the model
:param cluster_type: pp for parallel python, lsf for load sharing facility, sge for sun grid engine
:param pp_cpus: the number of cpu for parallel python
:param nfits: the number of fits to perform
:param outputdir: the directory to store the results
:param sim_data_dir: the directory containing the simulation data sets
:param updated_models_dir: the directory containing the Copasi models with updated parameters for
each estimation
:return: no output
"""
if int(nfits) < 1:
logger.error("variable " + nfits + " must be greater than 0. Please, check your configuration file.")
return
if not os.path.isfile(os.path.join(inputdir, model)):
logger.error(os.path.join(inputdir, model) + " does not exist.")
return
# folder preparation
refresh_directory(sim_data_dir, model[:-4])
refresh_directory(updated_models_dir, model[:-4])
copasi = get_copasi()
if copasi is None:
logger.error(
"CopasiSE not found! Please check that CopasiSE is installed and in the PATH environmental variable."
)
return
logger.info("Configure Copasi:")
logger.info(
"Replicate a Copasi file configured for parameter estimation and randomise the initial parameter values"
)
groupid = "_" + get_rand_alphanum_str(20) + "_"
group_model = model[:-4] + groupid
pre_param_estim = RandomiseParameters(inputdir, model)
pre_param_estim.print_parameters_to_estimate()
pre_param_estim.generate_instances_from_template(nfits, groupid)
logger.info("\n")
logger.info("Parallel parameter estimation:")
# To make things simple, the last 10 character of groupid are extracted and reversed.
# This string will be likely different from groupid and is the string to replace with
# the iteration number.
str_to_replace = groupid[10::-1]
command = (
copasi
+ " -s "
+ os.path.join(inputdir, group_model + str_to_replace + ".cps")
+ " "
+ os.path.join(inputdir, group_model + str_to_replace + ".cps")
)
parallel_computation(command, str_to_replace, cluster_type, nfits, outputdir, pp_cpus)
# Move the report files to the outputdir
report_files = [
f
for f in os.listdir(inputdir)
if re.match(group_model + "[0-9]+.*\.csv", f) or re.match(group_model + "[0-9]+.*\.txt", f)
]
for file in report_files:
# copy report and remove the groupid
shutil.move(os.path.join(inputdir, file), os.path.join(sim_data_dir, file.replace(groupid, "_")))
# removed repeated copasi files
repeated_copasi_files = [f for f in os.listdir(inputdir) if re.match(group_model + "[0-9]+.*\.cps", f)]
for file in repeated_copasi_files:
# os.remove(os.path.join(inputdir, file))
shutil.move(os.path.join(inputdir, file), os.path.join(updated_models_dir, file.replace(groupid, "_")))
