def __init__(
self,
ssh_session,
local_root,
job_uuid=None,
):
self.local_root = os.path.abspath(local_root)
if job_uuid:
self.job_uuid = job_uuid
else:
self.job_uuid = str(uuid.uuid4())
self.remote_root = os.path.join(ssh_session.get_session_root(),
self.job_uuid)
dlog.info("local_root is %s" % local_root)
dlog.info("remote_root is %s" % self.remote_root)
self.ssh = ssh_session.get_ssh_client()
# keep ssh alive
transport = self.ssh.get_transport()
transport.set_keepalive(60)
try:
sftp = self.ssh.open_sftp()
sftp.mkdir(self.remote_root)
sftp.close()
except:
pass
def make_fp(iter_index, jdata, mdata):
iter_name = make_iter_name(iter_index)
work_path = os.path.join(iter_name, fp_name)
create_path(work_path)
picked_data_path = os.path.join(iter_name, model_devi_name,
picked_data_name)
if jdata.get("labeled", False):
dlog.info("already labeled, skip make_fp and link data directly")
os.symlink(os.path.abspath(picked_data_path),
os.path.abspath(os.path.join(work_path, "task.%03d" % 0)))
os.symlink(os.path.abspath(picked_data_path),
os.path.abspath(os.path.join(work_path, "data.%03d" % 0)))
return
systems = get_systems(picked_data_path, jdata)
fp_style = jdata['fp_style']
if 'user_fp_params' in jdata.keys():
fp_params = jdata['user_fp_params']
else:
fp_params = jdata['fp_params']
jj = 0
for system in systems:
for subsys in system:
sys_data = subsys.data
task_name = "task.%03d.%06d" % (0, jj)
task_path = os.path.join(work_path, task_name)
create_path(task_path)
if fp_style == "gaussian":
ret = make_gaussian_input(sys_data, fp_params)
with open(os.path.join(task_path, 'input'), 'w') as fp:
fp.write(ret)
else:
# TODO: support other formats
raise RuntimeError("unsupported fp style")
jj += 1
def make_fp_labeled(iter_index, jdata):
dlog.info("already labeled, skip make_fp and link data directly")
pick_data = jdata['pick_data']
use_clusters = jdata.get('use_clusters', False)
iter_name = make_iter_name(iter_index)
work_path = os.path.join(iter_name, fp_name)
create_path(work_path)
picked_data_path = os.path.join(iter_name, model_devi_name,
picked_data_name)
if use_clusters:
os.symlink(
os.path.abspath(picked_data_path),
os.path.abspath(
os.path.join(work_path, "task." + data_system_fmt % 0)))
os.symlink(
os.path.abspath(picked_data_path),
os.path.abspath(
os.path.join(work_path, "data." + data_system_fmt % 0)))
else:
picked_data_path = os.path.abspath(picked_data_path)
sys_path = glob.glob(os.path.join(picked_data_path, sys_name_pattern))
cwd = os.getcwd()
os.chdir(work_path)
for ii in sys_path:
sys_idx = os.path.basename(ii).split('.')[1]
data_dir = 'data.' + data_system_fmt % int(sys_idx)
task_dir = 'task.' + data_system_fmt % int(sys_idx)
os.symlink(os.path.relpath(ii), data_dir)
os.symlink(os.path.relpath(ii), task_dir)
os.chdir(cwd)
def make_dispatcher(mdata,
mdata_resource=None,
work_path=None,
run_tasks=None,
group_size=None):
if 'ali_auth' in mdata:
from dpgen.dispatcher.ALI import ALI
nchunks = len(_split_tasks(run_tasks, group_size))
dispatcher = ALI(mdata['ali_auth'], mdata_resource, mdata, nchunks)
dispatcher.init(work_path, run_tasks, group_size)
return dispatcher
else:
hostname = mdata.get('hostname', None)
#use_uuid = mdata.get('use_uuid', False)
if hostname:
context_type = 'ssh'
else:
context_type = 'local'
try:
batch_type = mdata['batch']
except:
dlog.info(
'cannot find key "batch" in machine file, try to use deprecated key "machine_type"'
)
batch_type = mdata['machine_type']
lazy_local = (mdata.get('lazy-local', False)) or (mdata.get(
'lazy_local', False))
if lazy_local and context_type == 'local':
dlog.info('Dispatcher switches to the lazy local mode')
context_type = 'lazy-local'
disp = Dispatcher(mdata,
context_type=context_type,
batch_type=batch_type)
return disp
def gen_simplify(args):
if args.PARAM and args.MACHINE:
if args.debug:
dlog.setLevel(logging.DEBUG)
dlog.info("start simplifying")
run_iter(args.PARAM, args.MACHINE)
dlog.info("finished")
def delete(self, ii):
'''delete one machine'''
request = DeleteInstancesRequest()
request.set_accept_format('json')
request.set_InstanceIds(
[self.dispatcher_list[ii]["entity"].instance_id])
request.set_Force(True)
count = 0
flag = 0
while count < 10:
try:
response = self.client.do_action_with_exception(request)
flag = 1
break
except ServerException as e:
time.sleep(10)
count += 1
if flag:
status_list = [
item["dispatcher_status"] for item in self.dispatcher_list
]
running_num = status_list.count("running")
running_num += status_list.count("unsubmitted")
self.change_apg_capasity(running_num)
else:
dlog.info("delete failed, exit")
sys.exit()
def run_jobs(self,
resources,
command,
work_path,
tasks,
group_size,
forward_common_files,
forward_task_files,
backward_task_files,
forward_task_deference=True,
mark_failure=False,
outlog='log',
errlog='err'):
ratio_failure = self.mdata_resources.get("ratio_failue", 0)
while True:
if self.check_all_dispatchers_finished(ratio_failure):
self.clean()
break
self.exception_handling(ratio_failure)
for ii in range(self.nchunks):
dispatcher_status = self.check_dispatcher_status(ii)
if dispatcher_status == "unsubmitted":
dlog.info(self.dispatcher_list[ii]["entity"].ip)
self.dispatcher_list[ii][
"entity"].job_handler = self.dispatcher_list[
ii]["dispatcher"].submit_jobs(
resources, command, work_path,
self.task_chunks[ii], group_size,
forward_common_files, forward_task_files,
backward_task_files, forward_task_deference,
outlog, errlog)
self.dispatcher_list[ii][
"entity"].job_record = self.dispatcher_list[ii][
"entity"].job_handler["job_record"]
self.dispatcher_list[ii]["dispatcher_status"] = "running"
elif dispatcher_status == "finished" and self.dispatcher_list[
ii]["entity"]:
# no jobs in queue, delete current machine
# else add current machine to server_pool
entity = self.dispatcher_list[ii]["entity"]
status_list = [
item["dispatcher_status"]
for item in self.dispatcher_list
]
flag = "unallocated" in status_list
if not flag: self.delete(ii)
else:
self.dispatcher_list[ii]["entity"] = None
self.server_pool.append(entity.instance_id)
self.ip_pool.append(entity.ip)
elif dispatcher_status == "running":
pass
elif dispatcher_status == "unallocated":
# if len(server_pool) > 0: make_dispatcher
# else: pass
self.create(ii)
elif dispatcher_status == "terminated":
pass
self.update()
time.sleep(10)
def create_apg(self):
request = CreateAutoProvisioningGroupRequest()
request.set_accept_format('json')
request.set_TotalTargetCapacity(str(self.nchunks_limit))
request.set_LaunchTemplateId(self.cloud_resources["template_id"])
request.set_AutoProvisioningGroupName(
self.cloud_resources["instance_name"] +
''.join(random.choice(string.ascii_uppercase) for _ in range(20)))
request.set_AutoProvisioningGroupType("maintain")
request.set_SpotAllocationStrategy("lowest-price")
request.set_SpotInstanceInterruptionBehavior("terminate")
request.set_SpotInstancePoolsToUseCount(1)
request.set_ExcessCapacityTerminationPolicy("termination")
request.set_TerminateInstances(True)
request.set_PayAsYouGoTargetCapacity("0")
request.set_SpotTargetCapacity(str(self.nchunks_limit))
config = self.generate_config()
request.set_LaunchTemplateConfigs(config)
try:
response = self.client.do_action_with_exception(request)
response = json.loads(response)
with open('apg_id.json', 'w') as fp:
json.dump({'apg_id': response["AutoProvisioningGroupId"]},
fp,
indent=4)
return response["AutoProvisioningGroupId"]
except ServerException as e:
dlog.info("create apg failed, err msg: %s" % e)
sys.exit()
except ClientException as e:
dlog.info("create apg failed, err msg: %s" % e)
sys.exit()
def describe_apg_instances(self):
request = DescribeAutoProvisioningGroupInstancesRequest()
request.set_accept_format('json')
request.set_AutoProvisioningGroupId(self.cloud_resources["apg_id"])
request.set_PageSize(100)
iteration = self.nchunks // 100
instance_list = []
for i in range(iteration + 1):
request.set_PageNumber(i + 1)
count = 0
flag = 0
err_msg = 0
while count < 10:
try:
response = self.client.do_action_with_exception(request)
response = json.loads(response)
for ins in response["Instances"]["Instance"]:
instance_list.append(ins["InstanceId"])
flag = 1
break
except ServerException as e:
# dlog.info(e)
err_msg = e
count += 1
except ClientException as e:
# dlog.info(e)
err_msg = e
count += 1
if not flag:
dlog.info("describe_apg_instances failed, err msg: %s" %
err_msg)
sys.exit()
return instance_list
def create_template(self, image_id, sg_id, vpc_id):
request = CreateLaunchTemplateRequest()
request.set_accept_format('json')
request.set_LaunchTemplateName(''.join(
random.choice(string.ascii_uppercase) for _ in range(20)))
request.set_ImageId(image_id)
request.set_ImageOwnerAlias("self")
request.set_PasswordInherit(True)
if "address" in self.cloud_resources and self.cloud_resources[
'address'] == "public":
request.set_InternetMaxBandwidthIn(100)
request.set_InternetMaxBandwidthOut(100)
request.set_InstanceType("ecs.c6.large")
request.set_InstanceName(self.cloud_resources["instance_name"])
request.set_SecurityGroupId(sg_id)
request.set_VpcId(vpc_id)
request.set_SystemDiskCategory("cloud_efficiency")
request.set_SystemDiskSize(70)
request.set_IoOptimized("optimized")
request.set_InstanceChargeType("PostPaid")
request.set_NetworkType("vpc")
request.set_SpotStrategy("SpotWithPriceLimit")
request.set_SpotPriceLimit(100)
try:
response = self.client.do_action_with_exception(request)
response = json.loads(response)
return response["LaunchTemplateId"]
except ServerException as e:
dlog.info(e)
sys.exit()
except ClientException as e:
dlog.info(e)
sys.exit()
def make_super_cell_poscar(jdata) :
out_dir = jdata['out_dir']
super_cell = jdata['super_cell']
path_sc = os.path.join(out_dir, global_dirname_02)
create_path(path_sc)
from_poscar_path = jdata['from_poscar_path']
assert(os.path.isfile(from_poscar_path)), "file %s should exists" % from_poscar_path
from_file = os.path.join(path_sc, 'POSCAR.copied')
shutil.copy2(from_poscar_path, from_file)
to_path = path_sc
to_file = os.path.join(to_path, 'POSCAR')
#minor bug for element symbol behind the coordinates
from_struct=Structure.from_file(from_file)
from_struct.make_supercell(super_cell)
from_struct.to('poscar',to_file)
# make system dir (copy)
lines = open(to_file, 'r').read().split('\n')
natoms_str = lines[6]
natoms_list = [int(ii) for ii in natoms_str.split()]
dlog.info(natoms_list)
comb_name = "sys-"
for idx,ii in enumerate(natoms_list) :
comb_name += "%04d" % ii
if idx != len(natoms_list)-1 :
comb_name += "-"
path_work = os.path.join(path_sc, comb_name)
create_path(path_work)
cwd = os.getcwd()
to_file = os.path.abspath(to_file)
os.chdir(path_work)
os.symlink(os.path.relpath(to_file), 'POSCAR')
os.chdir(cwd)
def get_image_id(self, img_name):
request = DescribeImagesRequest()
request.set_accept_format('json')
request.set_ImageOwnerAlias("self")
request.set_PageSize(20)
response = self.client.do_action_with_exception(request)
response = json.loads(response)
totalcount = response["TotalCount"]
iteration = totalcount // 20
if iteration * 20 < totalcount:
iteration += 1
for ii in range(1, iteration + 1):
count = 0
flag = 0
request.set_PageNumber(ii)
while count < 10:
try:
response = self.client.do_action_with_exception(request)
response = json.loads(response)
for img in response["Images"]["Image"]:
if img["ImageName"] == img_name:
return img["ImageId"]
flag = 1
break
except:
count += 1
time.sleep(10)
if not flag:
dlog.info("get image failed, exit")
sys.exit()
def all_finished(self, job_handler, mark_failure, clean=True):
task_chunks = job_handler['task_chunks']
task_chunks_str = ['+'.join(ii) for ii in task_chunks]
task_hashes = [
sha1(ii.encode('utf-8')).hexdigest() for ii in task_chunks_str
]
job_list = job_handler['job_list']
job_record = job_handler['job_record']
command = job_handler['command']
tag_failure_list = [
'tag_failure_%d' % ii for ii in range(len(command))
]
resources = job_handler['resources']
outlog = job_handler['outlog']
errlog = job_handler['errlog']
backward_task_files = job_handler['backward_task_files']
dlog.debug('checking jobs')
nchunks = len(task_chunks)
for idx in range(nchunks):
cur_hash = task_hashes[idx]
rjob = job_list[idx]
if not job_record.check_finished(cur_hash):
# chunk not finished according to record
status = rjob['batch'].check_status()
job_uuid = rjob['context'].job_uuid
dlog.debug('checked job %s' % job_uuid)
if status == JobStatus.terminated:
job_record.increase_nfail(cur_hash)
if job_record.check_nfail(cur_hash) > 3:
raise RuntimeError(
'Job %s failed for more than 3 times' % job_uuid)
dlog.info('job %s terminated, submit again' % job_uuid)
dlog.debug('try %s times for %s' %
(job_record.check_nfail(cur_hash), job_uuid))
rjob['batch'].submit(task_chunks[idx],
command,
res=resources,
outlog=outlog,
errlog=errlog,
restart=True)
elif status == JobStatus.finished:
dlog.info('job %s finished' % job_uuid)
if mark_failure:
rjob['context'].download(task_chunks[idx],
tag_failure_list,
check_exists=True,
mark_failure=False)
rjob['context'].download(task_chunks[idx],
backward_task_files,
check_exists=True)
else:
rjob['context'].download(task_chunks[idx],
backward_task_files)
if clean:
rjob['context'].clean()
job_record.record_finish(cur_hash)
job_record.dump()
job_record.dump()
return job_record.check_all_finished()
def convert_data(jdata):
s = dpdata.MultiSystems(*[
dpdata.LabeledSystem(x, fmt="gaussian/log")
for x in glob.glob(os.path.join(fp_path, "*", "output"))
],
type_map=jdata["type_map"])
s.to_deepmd_npy(data_path)
dlog.info("Initial data is avaiable in %s" % os.path.abspath(data_path))
def sepline(ch='-', sp='-', screen=False):
r'''
seperate the output by '-'
'''
if screen:
print(ch.center(MaxLength, sp))
else:
dlog.info(ch.center(MaxLength, sp))
def block_checkcall(self,
cmd) :
stdin, stdout, stderr = self.ssh.exec_command(('cd %s ;' % self.remote_root) + cmd)
exit_status = stdout.channel.recv_exit_status()
if exit_status != 0:
dlog.info("Error info: %s "%(stderr.readlines()[0]))
raise RuntimeError("Get error code %d in calling %s through ssh with job: %s "% (exit_status, cmd, self.job_uuid))
return stdin, stdout, stderr
def download(self,
job_dirs,
remote_down_files,
check_exists=False,
mark_failure=True,
back_error=False):
cwd = os.getcwd()
for ii in job_dirs:
local_job = os.path.join(self.local_root, ii)
remote_job = os.path.join(self.remote_root, ii)
flist = remote_down_files
if back_error:
os.chdir(remote_job)
flist += glob('error*')
os.chdir(cwd)
for jj in flist:
rfile = os.path.join(remote_job, jj)
lfile = os.path.join(local_job, jj)
if not os.path.realpath(rfile) == os.path.realpath(lfile):
if (not os.path.exists(rfile)) and (
not os.path.exists(lfile)):
if check_exists:
if mark_failure:
with open(
os.path.join(
self.local_root, ii,
'tag_failure_download_%s' % jj),
'w') as fp:
pass
else:
pass
else:
raise RuntimeError('do not find download file ' +
rfile)
elif (not os.path.exists(rfile)) and (
os.path.exists(lfile)):
# already downloaded
pass
elif (os.path.exists(rfile)) and (
not os.path.exists(lfile)):
# trivial case, download happily
shutil.move(rfile, lfile)
elif (os.path.exists(rfile)) and (os.path.exists(lfile)):
# both exists, replace!
dlog.info('find existing %s, replacing by %s' %
(lfile, rfile))
if os.path.isdir(lfile):
shutil.rmtree(lfile, ignore_errors=True)
elif os.path.isfile(lfile) or os.path.islink(lfile):
os.remove(lfile)
shutil.move(rfile, lfile)
else:
raise RuntimeError('should not reach here!')
else:
# no nothing in the case of linked files
pass
os.chdir(cwd)
def _main(path, calculator, output, id_prefix):
assert calculator.lower() in SUPPORTED_CACULATOR
dlog.info('data collection from: %s' % path)
if calculator == "vasp":
parsing_vasp(path, output, id_prefix)
elif calculator == 'gaussian':
parsing_gaussian(path, output)
else:
parsing_pwscf(path, output)
def delete_apg(self):
request = DeleteAutoProvisioningGroupRequest()
request.set_accept_format('json')
request.set_AutoProvisioningGroupId(self.apg_id)
request.set_TerminateInstances(True)
try:
response = self.client.do_action_with_exception(request)
except ServerException as e:
dlog.info(e)
except ClientException as e:
dlog.info(e)
def _rmtree(self, sftp, remotepath, level=0, verbose = False):
for f in sftp.listdir_attr(remotepath):
rpath = os.path.join(remotepath, f.filename)
if stat.S_ISDIR(f.st_mode):
self._rmtree(sftp, rpath, level=(level + 1))
else:
rpath = os.path.join(remotepath, f.filename)
if verbose: dlog.info('removing %s%s' % (' ' * level, rpath))
sftp.remove(rpath)
if verbose: dlog.info('removing %s%s' % (' ' * level, remotepath))
sftp.rmdir(remotepath)
def prepare(self):
restart = False
if os.path.exists('apg_id.json'):
with open('apg_id.json') as fp:
apg = json.load(fp)
self.cloud_resources["apg_id"] = apg["apg_id"]
task_chunks_str = ['+'.join(ii) for ii in self.task_chunks]
task_hashes = [
sha1(ii.encode('utf-8')).hexdigest() for ii in task_chunks_str
]
for ii in range(self.nchunks):
fn = 'jr.%.06d.json' % ii
if os.path.exists(
os.path.join(os.path.abspath(self.work_path), fn)):
cur_hash = task_hashes[ii]
job_record = JobRecord(self.work_path,
self.task_chunks[ii], fn)
if not job_record.check_finished(cur_hash):
if not self.check_spot_callback(
job_record.record[cur_hash]['context']
['instance_id']):
self.dispatcher_list[ii]["entity"] = Entity(
job_record.record[cur_hash]['context']['ip'],
job_record.record[cur_hash]['context']
['instance_id'], job_record)
self.make_dispatcher(ii)
self.dispatcher_list[ii][
"dispatcher_status"] = "unsubmitted"
else:
os.remove(
os.path.join(os.path.abspath(self.work_path),
fn))
else:
self.dispatcher_list[ii][
"dispatcher_status"] = "finished"
self.server_pool = self.get_server_pool()
self.ip_pool = self.get_ip(self.server_pool)
restart = True
img_id = self.get_image_id(self.cloud_resources["img_name"])
sg_id, vpc_id = self.get_sg_vpc_id()
self.cloud_resources["template_id"] = self.create_template(
img_id, sg_id, vpc_id)
self.cloud_resources["vsw_id"] = self.get_vsw_id(vpc_id)
if not restart:
dlog.info("begin to create apg")
self.cloud_resources["apg_id"] = self.create_apg()
time.sleep(120)
self.server_pool = self.get_server_pool()
self.ip_pool = self.get_ip(self.server_pool)
else:
dlog.info("restart dpgen")
def check_apikey():
try:
apikey=os.environ['MAPI_KEY']
except KeyError:
print("You have to get a MAPI_KEY from "+web)
print("and execute following command:")
print('echo "export MAPI_KEY=yourkey">> ~/.bashrc')
print("source ~/.bashrc")
os._exit(0)
try:
return MPRester(apikey)
except MPRestError:
dlog.info("MPRester Error, you need to prepare POSCAR manually")
os._exit(0)
def parsing_vasp(path,
config_info_dict,
skip_init,
output=OUTPUT,
id_prefix=None):
fp_iters = os.path.join(path, ITERS_PAT)
dlog.debug(fp_iters)
f_fp_iters = glob(fp_iters)
dlog.info("len iterations data: %s" % len(f_fp_iters))
fp_init = os.path.join(path, INIT_PAT)
dlog.debug(fp_init)
f_fp_init = glob(fp_init)
if skip_init:
entries = _parsing_vasp(f_fp_iters, config_info_dict, id_prefix)
dlog.info("len collected data: %s" % len(entries))
else:
dlog.info("len initialization data: %s" % len(f_fp_init))
entries = _parsing_vasp(f_fp_init,
config_info_dict,
id_prefix,
iters=False)
entries.extend(_parsing_vasp(f_fp_iters, config_info_dict, id_prefix))
dlog.info("len collected data: %s" % len(entries))
#print(output)
#print(entries)
dumpfn(entries, output, indent=4)
def __init__ (self,
remote_root,
work_path,
job_uuid=None,
) :
self.remote_root=os.path.join(remote_root,work_path)
self.local_root = os.path.abspath(work_path)
if job_uuid:
self.job_uuid=job_uuid
else:
self.job_uuid = str(uuid.uuid4())
dlog.info("local_root is %s"% self.local_root)
dlog.info("remote_root is %s"% self.remote_root)
def gen_init_surf(args):
try:
import ruamel
from monty.serialization import loadfn, dumpfn
warnings.simplefilter('ignore',
ruamel.yaml.error.MantissaNoDotYAML1_1Warning)
jdata = loadfn(args.PARAM)
if args.MACHINE is not None:
mdata = loadfn(args.MACHINE)
except:
with open(args.PARAM, 'r') as fp:
jdata = json.load(fp)
if args.MACHINE is not None:
with open(args.MACHINE, "r") as fp:
mdata = json.load(fp)
out_dir = out_dir_name(jdata)
jdata['out_dir'] = out_dir
dlog.info("# working dir %s" % out_dir)
if args.MACHINE is not None:
# Decide a proper machine
mdata = decide_fp_machine(mdata)
fp_machine = mdata['fp_machine']
fp_ssh_sess = SSHSession(fp_machine)
#stage = args.STAGE
stage_list = [int(i) for i in jdata['stages']]
for stage in stage_list:
if stage == 1:
create_path(out_dir)
make_super_cell_pymatgen(jdata)
place_element(jdata)
make_vasp_relax(jdata)
if args.MACHINE is not None:
run_vasp_relax(jdata, mdata, fp_ssh_sess)
# elif stage == 0 :
# # create_path(out_dir)
# # make_super_cell(jdata)
# # place_element(jdata)
# # make_vasp_relax(jdata)
# # make_scale(jdata)
# # pert_scaled(jdata)
# # poscar_elong('POSCAR', 'POSCAR.out', 3)
# pert_scaled(jdata)
elif stage == 2:
make_scale(jdata)
pert_scaled(jdata)
else:
raise RuntimeError("unknown stage %d" % stage)
def ensure_alive(self, max_check=10, sleep_time=10):
count = 1
while not self._check_alive():
if count == max_check:
raise RuntimeError(
'cannot connect ssh after %d failures at interval %d s' %
(max_check, sleep_time))
dlog.info('connection check failed, try to reconnect to ' +
self.remote_host)
self._setup_ssh(self.remote_host,
self.remote_port,
username=self.remote_uname,
password=self.remote_password)
count += 1
time.sleep(sleep)
def gen_init_reaction(args):
try:
import ruamel
from monty.serialization import loadfn, dumpfn
warnings.simplefilter('ignore',
ruamel.yaml.error.MantissaNoDotYAML1_1Warning)
jdata = loadfn(args.PARAM)
if args.MACHINE is not None:
mdata = loadfn(args.MACHINE)
except:
with open(args.PARAM, 'r') as fp:
jdata = json.load(fp)
if args.MACHINE is not None:
with open(args.MACHINE, "r") as fp:
mdata = json.load(fp)
record = "record.reaction"
iter_rec = -1
numb_task = 7
if os.path.isfile(record):
with open(record) as frec:
for line in frec:
iter_rec = int(line.strip())
dlog.info("continue from task %02d" % iter_rec)
for ii in range(numb_task):
sepline(str(ii), '-')
if ii <= iter_rec:
continue
elif ii == 0:
link_reaxff(jdata)
elif ii == 1:
dispatcher = make_dispatcher(mdata["reaxff_machine"])
run_reaxff(jdata, mdata, dispatcher)
elif ii == 2:
link_trj(jdata)
elif ii == 3:
dispatcher = make_dispatcher(mdata["build_machine"])
run_build_dataset(jdata, mdata, dispatcher)
elif ii == 4:
link_fp_input()
elif ii == 5:
dispatcher = make_dispatcher(mdata["fp_machine"])
run_fp(jdata, mdata, dispatcher)
elif ii == 6:
convert_data(jdata)
with open(record, "a") as frec:
frec.write(str(ii) + '\n')
def create_ess(self):
img_id = self.get_image_id(self.adata["img_name"])
sg_id, vpc_id = self.get_sg_vpc_id()
self.template_id = self.create_template(img_id, sg_id, vpc_id)
self.vsw_id = self.get_vsw_id(vpc_id)
self.apg_id = self.create_apg()
dlog.info("begin to create ess, please wait two minutes")
time.sleep(120)
new_server_list = self.describe_apg_instances()
new_ip_list = self.get_ip(new_server_list)
for ii in range(len(new_server_list)):
profile = self.mdata_machine.copy()
profile['hostname'] = new_ip_list[ii]
profile['instance_id'] = new_server_list[ii]
if self.check_server(profile):
self.instance_list[ii] = new_server_list[ii]
self.ip_list[ii] = new_ip_list[ii]
def _main():
parser = argparse.ArgumentParser(description="gen init confs")
parser.add_argument('PARAM', type=str, help="parameter file, json format")
parser.add_argument(
'STAGE',
type=int,
help="the stage of init, can be 1 or 2 "
"1: Setup vasp jobs for relaxation. "
"2: Collect vasp relaxed confs (if relax is not skiped). Perturb system. "
)
args = parser.parse_args()
try:
import ruamel
from monty.serialization import loadfn, dumpfn
warnings.simplefilter('ignore',
ruamel.yaml.error.MantissaNoDotYAML1_1Warning)
jdata = loadfn(args.PARAM)
except:
with open(args.PARAM, 'r') as fp:
jdata = json.load(fp)
out_dir = out_dir_name(jdata)
jdata['out_dir'] = out_dir
dlog.info("# working dir %s" % out_dir)
stage = args.STAGE
if stage == 1:
create_path(out_dir)
make_super_cell_pymatgen(jdata)
place_element(jdata)
make_vasp_relax(jdata)
# elif stage == 0 :
# # create_path(out_dir)
# # make_super_cell(jdata)
# # place_element(jdata)
# # make_vasp_relax(jdata)
# # make_scale(jdata)
# # pert_scaled(jdata)
# # poscar_elong('POSCAR', 'POSCAR.out', 3)
# pert_scaled(jdata)
elif stage == 2:
make_scale(jdata)
pert_scaled(jdata)
else:
raise RuntimeError("unknow stage %d" % stage)
def make_super_cell_pymatgen(jdata):
make_unit_cell(jdata)
out_dir = jdata['out_dir']
path_uc = os.path.join(out_dir, global_dirname_02)
from_path = path_uc
from_file = os.path.join(from_path, 'POSCAR.unit')
ss = Structure.from_file(from_file)
# ase only support X type element
for i in range(len(ss)):
ss[i] = 'X'
ss = AseAtomsAdaptor.get_atoms(ss)
all_millers = jdata['millers']
path_sc = os.path.join(out_dir, global_dirname_02)
#z_min = jdata['z_min']
layer_numb = jdata['layer_numb']
super_cell = jdata['super_cell']
cwd = os.getcwd()
path_work = (path_sc)
path_work = os.path.abspath(path_work)
os.chdir(path_work)
for miller in all_millers:
miller_str = ""
for ii in miller:
miller_str += str(ii)
path_cur_surf = create_path('surf-' + miller_str)
os.chdir(path_cur_surf)
#slabgen = SlabGenerator(ss, miller, z_min, 1e-3)
slab = general_surface.surface(ss,
indices=miller,
vacuum=1e-3,
layers=layer_numb)
#all_slabs = slabgen.get_slabs()
dlog.info(os.getcwd())
#dlog.info("Miller %s: The slab has %s termination, use the first one" %(str(miller), len(all_slabs)))
#all_slabs[0].to('POSCAR', 'POSCAR')
slab.write('POSCAR', vasp5=True)
if super_cell[0] > 1 or super_cell[1] > 1:
st = Structure.from_file('POSCAR')
st.make_supercell([super_cell[0], super_cell[1], 1])
st.to('POSCAR', 'POSCAR')
os.chdir(path_work)
os.chdir(cwd)